ARK's Big Ideas 2025: Unlocking Exponential Growth

    ARK's Big Ideas 2025: Unlocking Exponential Growth

    T
    @TrendSpotting
    6 Followers
    4 months ago 442

    AIAI Summary

    toggle
    Bulleted
    toggle
    Text

    Key Insights

    FEBRUARY 04, 2025 RESEARCH REPORT FOR INFORMATIONAL PURPOSES ONLY WWW.ARK-INVEST.COM
ARK Investment Management LLC. This is not a recommendation in relation to any named particular securities/cryptocurrencies and no warranty or guarantee is provided. Any references to particular 
securities/cryptocurrencies are for illustrative purposes only. There is no assurance that the Adviser will make any investments with the same or similar characteristics as any investment presented. The 
reader should not assume that an investment identified was or will be profitable. PAST PERFORMANCE IS NOT INDICATIVE OF FUTURE PERFORMANCE, FUTURE RETURNS ARE NOT GUARANTEED.
BIG IDEAS 2025
    1/148
    Risks Of Investing In Innovation
Please note: Companies that ARK believes are capitalizing on disruptive innovation and developing technologies to displace older
technologies or create new markets may not in fact do so. ARK aims to educate investors and seeks to size the potential investment 
opportunity, noting that risks and uncertainties may impact our projections and research models. Investors should use the content 
presented for informational purposes only, and be aware of market risk, disruptive innovation risk, regulatory risk, and risks related to 
certain innovation areas. 
Please read risk disclosure carefully.
DISRUPTIVE 
INNOVATION
RAPID PACE OF CHANGE
UNCERTAINTY AND UNKNOWNS 
EXPOSURE ACROSS SECTORS AND MARKET CAP
RISK OF INVESTING IN INNOVATION
REGULATORY HURDLES
COMPETITIVE LANDSCAPE
POLITICAL OR LEGAL PRESSURE
Sources: ARK Investment Management LLC, 2023.
à Aim to understand the regulatory, market, sector, 
 and company risks. (See Disclosure Page)
à Aim for a cross-sector understanding of technology
 and combine top-down and bottom-up research.
ARK’s Big Ideas 2025
    2/148
    ARK’s Big Ideas 2025
Table Of Contents
Introduction Convergence AI Agents Bitcoin Stablecoins
Scaling
Blockchains
Robotaxis
Autonomous
Logistics Energy Robotics
Reusable
Rockets Multiomics
04 05 15 29 41 54
68 84 93 103 114 125
    3/148
    ARK’s Big Ideas 2025
BIG IDEAS 2025
4
As we stand on the cusp of a new era of unprecedented growth, 
ARK Invest’s “Big Ideas 2025” illuminates the complex 
convergences between and among the five technologically 
enabled innovation platforms evolving today: Artificial 
Intelligence, Robotics, Energy Storage, Public Blockchains, and 
Multiomic Sequencing. These platforms should drive exponential 
advances across industries and catalyze a step change in global 
economic growth.
Presented in this year’s report are 11 Big Ideas illustrating the 
massive transformations occurring today. Our research suggests 
that these Big Ideas are poised to boost productivity dramatically, 
to revolutionize industries, and to create long-term investment 
opportunities.
Join our journey into the future of innovation and its profound 
implications for investors, businesses, and society at large.
Welcome to ARK’s Big Ideas 2025.
Unlocking Exponential 
Growth Through 
Disruptive Innovation
    4/148
    The Second 
Half Of The 
Chessboard
Convergence Among 
Technology Platforms Leading 
To Significant Acceleration In 
Macroeconomic Growth
Brett Winton
CHIEF FUTURIST
ARK’s Big Ideas 2025
5
    5/148
    Five Innovation Platforms Are Catalyzing Accelerated Growth
Public Blockchains
Upon large-scale adoption, all money and contracts 
will likely migrate onto Public Blockchains that 
enable and verify digital scarcity and proof of 
ownership. The financial ecosystem is likely to 
reconfigure to accommodate the rise of 
Cryptocurrencies and Smart Contracts. These 
technologies increase transparency, reduce the 
influence of capital and regulatory controls, and 
collapse the costs of contract execution. In such a 
world, Digital Wallets will become increasingly 
necessary as more assets become money-like and 
corporations and consumers adapt to the new 
financial infrastructure. Corporate structures may 
be called into question. 
Multiomics
The cost to gather, sequence, and understand 
digital biological data is falling precipitously. 
Multiomic Technologies provide research scientists, 
therapeutic organizations, and health platforms 
with unprecedented access to DNA, RNA, protein, 
and digital health data. Cancer care should 
transform with pan-cancer blood tests. Fed by rich 
multiomic data and powered by Programmable 
Biology, AI systems running autonomous labs 
could collapse the cost of drug discovery, 
development, and trial, transforming returns in a 
sector that has stagnated. Biological discoveries 
should power novel Precision Therapies that target 
and cure rare diseases and chronic conditions, 
unlocking profound economics. Over time, the 
design and synthesis of novel biological constructs 
will yield advances in agriculture, material science, 
and even computation. 
AI
Computational systems and software that evolve with 
data can solve intractable problems, automate knowledge 
work, and accelerate technology’s integration into every 
economic sector. The adoption of Neural Networks should 
prove more momentous than electrification and 
potentially create tens of trillion dollars of value. At scale, 
these systems will require unprecedented computational 
resources, and AI-specific compute hardware should 
dominate the Next Gen Cloud datacenters that train and 
operate AI models. The potential for end-users is clear: a 
constellation of AI-driven Intelligent Devices that 
permeate people's lives, changing the way that they spend, 
work, and play. The adoption of artificial intelligence 
should transform every sector, impact every business, and 
catalyze every innovation platform.
Robotics
Catalyzed by artificial intelligence, Humanoid Robots should operate alongside 
humans and navigate legacy infrastructure, changing the way products are 
made and sold, and eventually the way we live our lives at home. 3D Printing 
should contribute to the digitization of manufacturing, increasing not only the 
performance and precision of end-use parts, but also the resilience of supply 
chains. Meanwhile, the world’s fastest robots, Reusable Rockets, should 
continue to reduce the cost of launching satellite constellations and enable 
uninterruptible connectivity and earth observation. A nascent innovation 
platform, robotics could collapse the cost of transporting across distance, with 
hypersonic travel, the cost of manufacturing complexity with 3D printers, and 
the cost of physical work with AI-guided robots.
Energy Storage 
The declining costs of Advanced Battery Technology should cause an 
explosion in form factors, enabling Autonomous Mobility systems that 
collapse the cost of transportation. Electric drivetrain cost declines 
should unlock micro-mobility and aerial systems, including flying taxis, 
enabling business models that transform cities. Autonomy should reduce 
the cost of taxi, delivery, and surveillance by an order of magnitude, 
enabling frictionless transport that will increase the velocity of ecommerce and make individual car ownership the exception rather than 
the rule. These innovations combined with large-scale stationary batteries 
and Distributed Energy Generation, notably solar and small-scale fission, 
should cause a transformation in energy, substituting electricity for liquid 
fuel and increasing system-wide resilience, reliability, and production.
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not 
be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Convergence
6
    6/148
    Row 2X Wheat 
Grains Value Year Computers Crossed The 
Same Compounding Threshold
8 128 2 teaspoons of 
wheat 1975
16 65 thousand 9 loaves of 
bread 1983
24 17 million 4 year’s 
sustenance 1998
32 4 billion 7 pounds of 
gold 2008
40 1 trillion 1 ton of 
gold 2018
48 300 trillion 20% of India 
GDP in 600 AD 2023
56 70 
quadrillion
50x Global 
GDP in 600 AD 2027e
64 18 
quintillion
9x Global 
GDP in 2024 2030e
By lore, when he agreed to the 
reward, the Indian ruler thought it 
quite reasonable. By the 6th row of 
the chessboard, however, he had 
exhausted India’s treasury, having 
paid just ~.001% of the bill.
Today, computational advance has 
completed the equivalent of the 6th
row on the chessboard. Entering the 
AI cycle in 2018, computers had 
crossed 40 performance doublings 
and in 2023 surpassed 48. By the end 
of this decade, thanks to the 
acceleration in AI, computation 
could reach the end of the 
chessboard.
In Compounding, The Real Action Takes Place 
On The Second Half Of The Chessboard
I don’t require much, 
your highness, just a 
grain of wheat on the 
first square of the 
board, doubling 
consecutively on each 
square thereafter.
Legendarily requesting his reward 
for inventing chess, the 6th century 
advisor to the Indian ruler:
Note: “GDP”: Gross Domestic Product, aggregate global production for the global datapoints. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, Kurzweil 2006 as of 
December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is 
not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Convergence
7
    7/148
    1975
8 performance 
doublings
1983
16 performance 
doublings
1998
24 performance 
doublings
2008
Halfway through 
the chessboard
Google announces 
the TPU
2023
48 performance 
doublings
2027
56 performance 
doublings
2030
End of the 
chessboard
2018
40 performance 
doublings
2028
48 performance 
doublings
1E-02
1E+00
1E+02
1E+04
1E+06
1E+08
1E+10
1E+12
1E+14
1E+16
1E+18
Compute Capability (Per 2024 $)
Cumulative Compute Purchased
(Unit Of xAI Colossus-Sized Datacenters)
Integrated Circuit Performance Improvement Curve
100 
Billionth
1 
Billionth
100 
Millionth
1 
Millionth
10 
Thousandth
1 
Hundredth
1 100 10,000 1,000,000
AI Is Accelerating The Performance 
Doubling Rate, Closing In On The 
End Of The Chessboard
Enabled primarily by architectural improvements in AI systems, 
performance per dollar of AI compute is expected to improve >1000x 
by 2030. At that time, we expect that compute performance will have 
doubled 64 times since the advent of the integrated circuit.
Improvements in 
AI Architecture 
Driving Compute 
Performance
Compute 
Hardware 
Driving 
Performance 
Improvements
Each Circle Represents 1 Performance Doubling
Hollow Circles Are Forecasts
ARK’s Big Ideas 2025: Convergence
8
Note: The “Compute Capability per $2024 Dollar” displayed on the Y axis above is a single metric that measures 1000s of computations per second purchasable for $1. For AI compute performance, computational impact is adjusted 
upwards by the rate of AI architectural performance per dollar improvements according to ARK’s research. It's a very large number, so we use scientific notation, e.g., "1E+21," meaning 1 followed by 21 zeros. Source: ARK Investment 
Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Kurzweil 2005 and Jurvetson 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    8/148
    2023
2024
0
50
100
150
200
250
0 500 1000 1500
# Of Hypotheses Tested Per FTE Researcher
H100 Equivalents Compute 
In Recursion’s Datacenter
Recursion Drug Development Efficiency
10/1/23 4/18/24 11/4/24 5/23/25
Log Error Rate Scale
SWE-Bench Verified
Best In Class AI Software Engineering
AI Advances Should Unlock Massive Market Opportunities
72%
54%
As AI continues to accelerate, robotaxis should proliferate, drug development timelines and costs should collapse, and AI 
agents should solve software engineering challenges autonomously, monitoring and modifying systems around the clock.
4%
22%
38%
45%
FSD 11.5
12/31/23
FSD 12.5
10/24/24
FSD 13.4
1Q25*
FSD 13.6
2H25*
1
10
100
1000
10000
100000
1000000
 - 100,000 200,000
Miles Per Intervention
H100 Equivalent Compute 
In Tesla’s Datacenter
Tesla FSD Performance
Waymo intervention rate 
post commercialization
Human driver accident rate
ARK’s Big Ideas 2025: Convergence
9
*Estimates. Note: “FSD”: Full Self-Driving. SWE-bench is a dataset that tests systems' ability to solve GitHub issues automatically. The dataset collects 2,294 Issue-Pull Request pairs from 12 popular Python repositories. 
Evaluation is performed by unit test verification using post-PR behavior as the reference solution. See Jiminez et al. 2024. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external 
data sources, including Tesla, Recursion, and SWEbench.com, and OpenAI as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment 
advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    9/148
    Convergence Is Accelerating 
The Technology Revolution
ARK measures the degree to which 
technologies serve as catalysts between and 
among innovation platforms. The 
convergences among them are increasing, with 
network density up 30% in the past year.
AI is proving more critical to unlocking the 
value of Precision Therapies and Multiomic 
Technologies. Smart Contract ecosystems are 
serving as test beds in which autonomous AI
agents can be renumerated for sharpening 
their capabilities. Next Gen Cloud energy 
demand is pulling forward timelines for 
Distributed Energy Generation. 
Catalyzed 
Increase In 
Market 
Opportunity
Order Of 
Magnitude+
Multiples
Very Meaningful
Meaningful
Potential
ARK’s Big Ideas 2025: Convergence
10
Note: “Network density” measures the degree of interconnectedness between nodes relative to the maximum potential interconnectedness. In our research, if every technology were expected to catalyze 
another technology to increase in value by an order of magnitude or more, that would equate to a fully interconnected network. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a 
range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, 
sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    10/148
    The Acceleration In Neural Networks Is 
Accelerating Every Other Disruptive Technology
Cryptocurrencies
Smart
Contracts
Digital
Wallets
Precision
Therapies
Multiomic
Technology
Programmable
Biology
Neural
Networks
Next Gen
Cloud
Intelligent
Devices
Autonomous
Mobility
Advanced Battery
Technology
Reusable
Rockets
Humanoid
Robotics
3D
Printing
Distributed Energy 
Generation
CryptocurrenciesSmart 
Contracts
Digital
Wallets
Precision 
Therapies
Multiomic
Technology
Neural
Networks
Next Gen
Cloud
Intelligent
Devices
Autonomous
Mobility
Advanced
Battery
Technology
Reusable
Rockets
Humanoid
Robotics
3D
Printing
Programmable
Biology
Distributed
Energy 
Generation
Technology
Catalyzing Technology
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Convergence
11
    11/148
    Advances In AI, Energy Storage, And Public Blockchains Are Critical 
To The Pace Of Technological Advance
Technology Importance As A Catalyst
Advances 
another 
technology 
by 1 order of 
magnitude
Neural
Networks
Intelligent 
Devices
Next Gen 
Cloud
Advanced Battery 
Technology
Digital Wallets Reusable Rockets
3D
Printing
Cryptocurrencies Smart Contracts
Autonomous 
Mobility Distributed Energy 
Generation
Multiomic 
Technologies
Precision
Therapies
Programmable 
Biology
Humanoid
Robots
Among disruptive technologies, Neural Networks
are the most important catalyst. 
According to our research, advances in neural 
networks will increase the value of six of the other 
14 technologies by at least an order of magnitude, 
creating massive market expansions for Next Gen 
Cloud, Intelligent Devices, Autonomous Mobility, 
Humanoid Robots, Precision Medicine, and 
Multiomic Technology.
The AI-led revolution in technology is likely to 
lead to dramatic productivity gains and a stepfunction increase in economic growth.
Technology 
Platform
AI
Energy 
Storage
Multiomics
Robotics
Public 
Blockchains
SCALE
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Convergence
12
    12/148
    Technology Inflection Should Lead To GDP Inflection
Structural changes in the underlying rate of macroeconomic 
growth are the historical rule, not the exception. 
After 100,000 years of economic stagnation leading to year 1, 
innovations, notably writing, allowed empires to stitch 
together continents, which quadrupled real growth through 
1000 AD. 
Agricultural innovations enabled denser populations and 
labor specialization, leading to a doubling of growth through 
1500 to 0.3% per annum.
During the 400 years through 1900, annual GDP growth 
doubled again to 0.6% as the Enlightenment and Industrial 
Revolution swept the globe. 
The Second Industrial Revolution—marked by electrification, 
the automobile, and telephony—ushered in the modern age 
and quintupled growth to 3% on average over the last 125 
years.
Technological breakthroughs in AI, Autonomous Mobility, 
and Humanoid Robots are likely to turbocharge productivity 
again, causing another step function in growth over the next 
5 to 10 years.
0.0%
0.1%
1.0%
10.0%
100.0%
Compounded Annual Growth
(Log Scale)
Global Real GDP Growth
Log Years Until 2050
100,000 BC 1 
1000
1500
1900
2024
.037%
0.14%
0.3%
0.6%
3%
IMF Forecast
3.1%
ARK Forecast
7.3%
2030
ARK’s Big Ideas 2025: Convergence
13
Note: “Real GDP”: Aggregate Global Production adjusted for changes in price levels. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources including DeLong 1998 and Open 
Philanthropy as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past 
performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    13/148
     -
 50
 100
 150
 200
 250
 300
$ Trillions
(Nominal)
Innovation Should Drive A Momentous Business Cycle
Disruptive Innovation Could Grow To Dominate Markets
2014 2019 2024 2030e
Disruptive innovation could 
command more than two-thirds 
of the global equity market, 
compounding at a 38% rate 
through 2030.
Within innovation, market value 
could broaden out beyond the 
so-called Mag 6.
Non-innovation businesses could 
begin to shrink as technological 
deflation associated with 
innovation threatens margins and 
competitiveness, even amid rapid 
macroeconomic growth.
GDP Non-disruptive
Mag 6**
Other Disruptive 
Innovation*
45
1.7 
4.7
73
92
127
220
38%
CAGR
ARK’s Big Ideas 2025: Convergence
65
5 
9
103
16 
12
89
-2%
CAGR
60
24%
CAGR
140
50%
CAGR
14
Note: “CAGR”: Compounded Annual Growth Rate. *We include cryptocurrencies and crypto assets in disruptive innovation capitalization. **Microsoft, Nvidia, Meta, Apple, Amazon, and Alphabet. Source: ARK Investment 
Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered 
investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
Global Market 
Capitalization
    14/148
    Redefining Consumer 
Interactions And 
Business Workflows
ARK’s Big Ideas 2025
AI Agents
Nicholas Grous Frank Downing
ASSOCIATE PORTFOLIO 
MANAGER
DIRECTOR OF RESEARCH, 
NEXT GENERATION INTERNET
Varshika Prasanna Jozef Soja
RESEARCH ASSOCIATE, 
NEXT GENERATION INTERNET
RESEARCH ANALYST, 
NEXT GENERATION INTERNET
15
    15/148
    AI Agents:
Understand intent through 
natural language
Plan using reasoning and 
appropriate context
Take action using tools to 
accomplish the intent
Improve through iteration 
and continuous learning
What Are AI Agents?
Platform-Level
Agents
Siloed,
Single-Purpose 
Agents
Usefulness
Level Of Intelligence
Universal Agents
AI agents are poised to accelerate the 
adoption of digital applications and create an 
epochal shift in human-computer interaction.
ARK’s Big Ideas 2025: AI Agents
Smarter models 
using more tools to 
accomplish highervalue tasks
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 16
    16/148
    0
1
2
3
4
5
6
7
8
9
10
 - 2 4 6 8 10 12 14
Revenue ($B)
Years Since Product Launch
AI Monetization is Outpacing Social Media 
OpenAI Instagram Facebook
TikTok YouTube WeChat
AI Is Accelerating The 
Adoption of Hardware 
And Software
OpenAI could surpass $10 billion in revenue in 2025, monetizing at a faster rate 
than social media companies over the past decade. If the adoption of ChatGPT is 
an indicator, AI should drive rapid demand for a range of new technologies. 
ARK’s Big Ideas 2025: AI Agents
$0
$1
$2
$3
$4
$5
$6
$7
$8
$9
$10
 - 5 10 15
Revenue ($ Billions)
Years Since Product Launch
AI Monetization Is Outpacing Social Media 
OpenAI Instagram Facebook
TikTok YouTube WeChat
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1974 1983 1992 2001 2010 2019 2028E 2037E
Percent Penetration
US Consumer Hardware Adoption 
PC Cellphone Smartphone Smart Home
Smartwatch VR Headset AI Hardware
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Hartig 1998, Dediu 2017, and Sidoti et al. 2024, as of December 31, 2024, which may be provided upon 
request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are 
inherently limited and cannot be relied upon. 17
Starting in 2025, we believe 
most hardware sales will be 
AI-enabled, leading to fast 
mass adoption.
    17/148
    Embedded in the operating systems of consumer hardware, AI agents 
empower consumers to delegate all discovery and research to AI—a massive 
time-saver. Curated AI results will contextualize digital ad impressions. 
AI Agents Will Transform 
Consumer Search And 
Discovery
TEXT / 
VOICE 
COMMAN
D
CHECKOU
T
AWARENESS AUTONOMOUS 
DISCOVERY + RESEARCH CONVERSION
QUERY
PERSONAL
IZED AI 
AGENT
CURATED 
RESULTS
SEARCH 
ENGINE 1
SEARCH 
ENGINE 2
SEARCH 
ENGINE 3
MARKETPLACE 1
MARKETPLACE 2
MARKETPLACE 3 = Ad Opportunities
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 18
ARK’s Big Ideas 2025: AI Agents
    18/148
    AI Agent Revenue Opportunities AI-Mediated Ads Should Take 
The Lion’s Share Of Digital Ad 
Revenue by 2030
If search shifts to personal AI agents, AI-mediated ad revenue 
could surge. By 2030, we believe AI ad revenue could account 
for more than 54% of the $1.1 trillion digital ad market. 
ARK’s Big Ideas 2025: AI Agents
0%
1%
2%
3%
4%
5%
6%
7%
8%
August-21 August-22 August-23 August-24
Share Of Page Views
ChatGPT, Claude, Perplexity, And Bing Share Of Search 
VS. Google
(Based On 7-Day Rolling Average Of Page Views)
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 19
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
 $-
 $200
 $400
 $600
 $800
 $1,000
 $1,200
2024 2025E 2026E 2027E 2028E 2029E 2030E
Gross Spend ($ Billions)
Global Advertising Forecast
AI Ad Revenue Digital Ad Spend AI Ad Revenue % of Total Digital Spend
    19/148
    AI-Mediated Shopping Is Likely 
To Approach 25% Of Addressable 
Online Sales Globally By 2030
The growing use of AI agents in consumer shopping should 
streamline product discovery, personalization, and purchasing. 
ARK research suggests that AI agents could facilitate nearly $9 
trillion in global gross online consumption by 2030.
Note: ARK measures “total addressable online sales” using internal and external estimates for global e-commerce and addressable services spending out to 2030. Source: ARK Investment Management LLC, 2025. 
This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 20
0%
5%
10%
15%
20%
25%
30%
 $-
 $1
 $2
 $3
 $4
 $5
 $6
 $7
 $8
 $9
 $10
2023 2024 2025E 2026E 2027E 2028E 2029E 2030E
Percent
$ Trillions
AI-Facilitated Online Spend
AI-Facilitated Spend (LHS) % of Total Addressable Online Sales
ARK’s Big Ideas 2025: AI Agents
    20/148
    Digital Wallets Are Positioned 
For Continued Share Gains In 
E-Commerce
ARK’s research suggests that digital wallets empowered by 
AI purchasing agents—taking share from payment methods 
like credit and debit cards—could account for 72% of all ecommerce transactions by 2030.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
2018 2019 2020 2021 2022 2023 2024 2025E 2026E 2027E 2028E 2029E 2030E
Percent
Share Of E-Commerce Purchase Volume
Digital Wallet Other Payment Methods
21
Note: Other Payment Methods for E-commerce purchase volume include Credit Card, Debit Card, Bank Transfer, Cash on Delivery, and Buy Now Pay Later (BNPL). “Credit Card” and “Debit Card” refer to stored card 
credentials, while “Digital Wallet” includes credit/debit cards stored within the digital wallet. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including 
WorldPay as of December 31, 2024, which may be provided upon request. 2018-2023 figures are documented in the Worldpay Global Payments Reports. 2024-2030 figures are ARK's adoption estimates for e-commerce 
payment methods. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future 
results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: AI Agents
    21/148
    Digital Wallets Are Consolidating 
Financial Services And E-Commerce
Based on their consumer-facing operations, the 
market is valuing leading digital wallet platforms like 
Block, Robinhood, and SoFi at $1,800 per user today.
$1,800 
$250 
$1,000 
$300 
$150 
$100 
 $-
 $200
 $400
 $600
 $800
 $1,000
 $1,200
 $1,400
 $1,600
 $1,800
 $2,000
Checking & Savings Lending Insurance Investing E-Commerce Total
$
Digital Wallet Enterprise Value Per User
(US)
22
Note: “Checking & Savings” data include Checking & Savings, Debit Card, and Peer to Peer Payments. “Lending” data include Loans & Mortgages, Credit Card, and Buy Now Pay Later (BNPL). Source: ARK Investment Management 
LLC, 2025. This ARK analysis draws on a range of external data sources representing the revenues, margins, and multiples of leading digital wallets and established incumbents. as of December 31, 2024, which may be provided 
upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: AI Agents
    22/148
    Agentic lead generation should move digital wallets 
upstream to capture market share in e-commerce and digital 
spending globally. “One-click checkout” should give way to 
“one-query purchasing.”
Digital Wallet Purchasing 
Agents Could Become Central 
To The Shopping Journey
Interaction With 
Digital Wallet
Existing System
System With Purchasing Agent
Consumer Search Engine /
3rd-Party Aggregator
Untailored
Results
Shopping
Cart / Bag
Checkout Pay With
Digital Wallet
SEARCH LAYER AGGREGATION LAYER CHECKOUT LAYER
SEARCH LAYER AGGREGATION LAYER CHECKOUT LAYER
Consumer Marketplaces Personalized Results Automatic Checkout
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 23
ARK’s Big Ideas 2025: AI Agents
    23/148
    In 2030, AI-powered purchasing agents could add between 
$50 and $200 per user to the enterprise value (EV) of digital 
wallets in the US.
Purchasing Agents Should Increase The Enterprise Value 
Of Digital Wallets, Notably In E-Commerce 
2030 Forecast
Global Gross Spend 
Powered By AI Agents ~ $9 Trillion
Percent Of AI Agent Spend In Digital Wallets
5% 10% 20%
Lead 
Generation 
Take-Rate 
For Digital 
Wallets
2.5% $10B $20B $40B
5% $20B $40B $90B
10% $40B $90B $200B
Based on lead-generation take rates, AI purchasing agents 
could generate global revenue between $40 billion and $200 
billion—ARK’s base and bull cases, respectively—for digital 
wallet platforms in 2030.
$100 
$50 
$200 
 $-
 $50
 $100
 $150
 $200
 $250
 $300
 $350
2023 2030E - Base Case 2030E - Bull Case
$
AI-Driven Digital Wallet EV Per User 
In E-Commerce (US)
E-Commerce Forecast - Additional Purchasing Agent Opportunity
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 24
ARK’s Big Ideas 2025: AI Agents
    24/148
    In The Enterprise, Agents Will 
Increase Productivity Via Software
INBOUND
REQUESTS
AI Agent Platform
SOFTWARE
INVESTMENT
LABOR
SAVINGS
Traditional Customer Support AI-Enabled Customer Support
Companies that deploy agents should be able to increase unit 
volume with the same workforce and/or optimize their workforce 
toward higher-value activities. As AI evolves, agents are likely to 
handle a higher percentage of workloads and to complete highervalue tasks independently.
HUMAN AGENTS INBOUND
REQUESTS CUSTOMERS CUSTOMERS HUMAN AGENTS
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 25
ARK’s Big Ideas 2025: AI Agents
    25/148
    AI Cost Declines Should Impact Agent Economics Significantly
New products from OpenAI and Salesforce are supplementing human customer service representatives cost-effectively. Even at a fixed 
cost of $1 per conversation, AI agents could save enterprises significant sums once they can handle 35% of customer service inquiries. AI 
agents also should lower onboarding and hiring costs, as well as seat-based software costs, while scaling more easily than human labor. 
Potential Enterprise Customer Service Savings** 
($ Billions)
Dollar Cost Per Conversation
8x $290 $430 $500 $540
4x $140 $220 $250 $270
2x $70 $110 $130 $130
1x $40 $50 $60 $70
$1.00 $0.50 $0.25 $0.125 $0
$20
$40
$60
$80
$100
$120
$140
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
$ Billions
Percent Of Customer Interactions Handled Successfully By AI
AI Improvements In Customer Service Should Shift 
Labor Costs To Software Costs*
Human Labor Costs AI Software Costs Savings
Current U.S. Human Labor Spend
Demand For Customer Service 
Relative To Today
*Assumes fixed cost of $1 per customer interaction. **Assuming 70% of interactions can be handled by AI in all scenarios, corresponding to the 70% bar in the left-hand chart. Source: ARK Investment Management LLC, 2025. 
This ARK analysis draws on a range of external data sources, including ZipRecruiter and OpenAI, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered 
investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 26
ARK’s Big Ideas 2025: AI Agents
    26/148
    0
-1
7
11
-7 -10 -10
-5
0
5
10
15
Share Shift 
(Percentage Points)
PP
19%
31%
23%
39%
15%
27%
0%
20%
40%
60%
Spend CAGRs
Estimated Enterprise Software Stack Evolution
Infrastructure Layer Platform Layer Application Layer
o3 (OpenAI)
Q Developer Agent (Amazon)
Claude 3.5 Sonnet + Tools (Anthropic)
Genie (Cosine)
SWE-agent + GPT 4 (Princeton, OpenAI)
Claude 2 + RAG (Anthropic)
0%
10%
20%
30%
40%
50%
60%
70%
80%
Sep-23 Dec-23 Mar-24 Jun-24 Sep-24 Dec-24
SWE-Bench Verified Score*
Percentage Of Real World Coding Tasks Solved 
By Coding Agents
The coding skills of AI agents are improving rapidly, accelerating the software development lifecycle. As the cost to create software 
declines, software production should accelerate and sway enterprise “build vs. buy” decisions, displacing traditional software 
incumbents that are slow to adapt. As customer software proliferates, growth in all layers of the software stack should accelerate, 
even as share shifts toward the platform layer. 
AI Is Reshaping The Software Value Chain
2019 To 2024
(Actual)
2024 To 2030
(Forecast)
Note: “CAGR”: Compound Annual Growth Rate. *SWE-Bench is a benchmark that measures AI agents’ ability to write code autonomously. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of 
external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular 
security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 27
ARK’s Big Ideas 2025: AI Agents
    27/148
    AI is driving a Cambrian explosion in software. Through 2030, we expect the amount of 
software deployed per knowledge worker to grow considerably as businesses invest in 
productivity solutions. Depending on adoption rates, global spend on software could 
accelerate from an annual rate of 14% over the last ten years to annual rates of 18% to 48%.
AI Will Supercharge
Knowledge Work
Modest
Investment
Accelerated
Investment
Rapid
Mass Adoption
Annual knowledge worker 
employment growth 6.3% 3.2% 1.3%
Percentage of current working 
time automated by 2030 31% 61% 81%
Reduction in productive 
working hours 0% 8% 20%
Productivity surplus created* $22 trillion $57 trillion $117 trillion
Value capture of productivity 
solutions** 10% 10% 10%
New software revenue $2.2 trillion $5.7 trillion $11.7 trillion
2030 Software Market Estimate
(current size + AI revenue)
$3.5 trillion
18% CAGR
$7 trillion
33% CAGR
$13 trillion
48% CAGR
Adoption Scenarios
$0
$10
$20
$30
$40
$50
$60
Current Modest
Investment
Scenario
Accelerated
Investment
Scenario
Rapid Mass
Adoption
Scenario
2030 Enterprise Spending Forecast
Knowledge Work Labor Software
28
Note: “CAGR”: Compound Annual Growth Rate. *Traditional production statistics are unlikely to adequately capture the surplus created by AI software. **Value capture rates are likely to vary across scenarios 
based on competition in the market and other factors. The rate is held constant in these scenarios for simplicity. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external 
data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any 
particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: AI Agents
    28/148
    A Maturing Global Monetary 
System With Sound Network 
Fundamentals And Growing 
Institutional Adoption
ARK’s Big Ideas 2025
Bitcoin
Yassine Elmandjra David Puell
FORMER DIRECTOR OF DIGITAL ASSETS RESEARCH TRADING ANALYST &
ASSOCIATE PORTFOLIO MANAGER,
DIGITAL ASSETS
Lorenzo Valente
DIRECTOR OF DIGITAL ASSETS
29
    29/148
    $30,000
$40,000
$50,000
$60,000
$70,000
$80,000
$90,000
$100,000
$110,000
Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24
($)
Bitcoin Price
Bitcoin Hit An All-Time High In 2024
Bitcoin Undergoes 
Its Fourth Halving
US Launches Spot 
Bitcoin ETFs 
Ethereum’s Dencun 
Upgrade Activates 
On Mainnet
Sam Bankman Fried 
Sentenced to 25 Years 
In Prison US Launches Spot Ether 
ETFs 
German 
Government Sells 
Out Of 50,000 
Seized Bitcoin
Bitcoin Hash Rate Hits An 
All-Time High
Genesis Settles SEC 
Lawsuit Over Gemini 
Earn
Hong Kong Debuts 
Asia’s First Spot 
Bitcoin And Ether 
ETFs
SEC Issues 
Robinhood Crypto 
With Wells Notice
Bitcoin Market Cap Dominance 
Surges Past 55% For The First 
Time in Three Years
Bitcoin Hits A Technical 
All-Time High
ARK’s Big Ideas 2025: Bitcoin
As Donald Trump Wins 
US Presidential 
Election, Bitcoin 
Breaks All-Time High
SEC Grants Trading 
And Listing Bitcoin 
ETF Options 
Pennsylvania Is First US 
State To Propose Bitcoin 
Strategic Reserve
David Sacks Appointed 
US AI And Crypto Czar
Source: ARK Investment Management LLC, 2025, based on data from Glassnode and TradingView as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation 
to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
Mt. Gox Begins $9 Billion 
Payout To Creditors
MicroStrategy Announces 
$42-Billion Raise Over 
The Next 3 Years
30
    30/148
    The Spot Bitcoin ETF Complex Was The Most Successful ETF Launch In History
On their first day of trading, the spot bitcoin ETFs attracted over $4 billion of inflows, a record high for ETF launches, surpassing the 
$1.2 billion that flowed into the gold ETF in its first month in November 2004. Inflows into the spot bitcoin ETFs have dwarfed the firstmonth inflows into each of the ~6,000 ETFs launched over the past 30 years. 
$0
$20
$40
$60
$80
$100
$120
Jan-24
Feb-24Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Dec-24
$ Billions
Spot Bitcoin ETFs1 AUM2
IBIT FBTC ARKB BITB BTC HODL BRRR BTCO EZBC BTCW
0
10
20
30
40
50
60
70
80
90
100
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86
$ Billions
Number Of Months From Launch
Spot Bitcoin ETFs And Gold ETF AUM 
In Year One and Beyond
Bitcoin ETF Cumulative Flows Gold ETF Cumulative Flows
[1] GBTC and DEFI are not included. [2] “AUM”: assets under management. Source: ARK Investment Management LLC, 2025, based on data from Glassnode and World Gold Council as of December 31, 2024. For informational 
purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 31
ARK’s Big Ideas 2025: Bitcoin
    31/148
    Upon Its Fourth Halving, Bitcoin’s 
Inflation Rate Dropped Below Gold’s 
Long-Term Supply Growth
Growth in the supply of bitcoin “halved” for the fourth time in 
history, dropping from ~1.8% at an annual rate to ~0.9%. Critical 
to its design, the halving highlights bitcoin’s predictable monetary 
policy and its role as a scarce asset.
Seven Lines Of Code Created Bitcoin’s Mathematically-Metered Monetary Policy1
0%
20%
40%
60%
80%
100%
0
5
10
15
20
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Percent
BTC, Millions
Bitcoin Supply
Mined Supply (lhs) Percent to Be Mined (rhs)
0%
1%
10%
100%
1000%
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
Annualized Percent
Inflation Rate: Bitcoin VS. Gold
BTC Supply Growth Gold Supply Growth
[1] Screenshot as of December 31, 2024, sourced from (github.com/bitcoin/bitcoin/blob/master/src/validation.cpp) lines 
1939-1949. Source: ARK Investment Management LLC, 2025, based on data from Glassnode and World Gold Council as of 
December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation 
to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 32
ARK’s Big Ideas 2025: Bitcoin
    32/148
    Bitcoin’s Annual Volatility Fell To An All-Time Low, While Its Risk-Adjusted 
Returns Remained Superior To Most Major Asset Classes
Asset Class1 2024 
Return
Sharpe 
Ratio2
Sortino 
Ratio3
5-Year 
CAGR4
Bitcoin 122.2% 1.4 4.1 67.2%
Gold 26.6% 1.7 4.8 10.4%
Equities 19.2% 1.3 2.3 11.7%
Emerging 
Markets 8% 0.2 0.4 2.1%
Real Estate 3.9% -0.01 -0.01 1.5%
0% Bonds -3.1% -1.1 -1.2 -2.7%
50%
100%
150%
200%
250%
2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
Annualized Percent, 1-Year, Rolling
Bitcoin Volatility
[]1 Data used for calculations are as follows: S&P Bitcoin PR USD for bitcoin; S&P GSCI Gold TR for gold; MSCI World GR USD for equities; MSCI EM GR USD for emerging markets; S&P Global REIT TR USD for real estate; and 
Bloomberg Gbl Agg Govt TR USD for bonds. The returns shown are of unmanaged indices and therefore do not reflect the deduction of any fees or expenses that would be applicable to investors in these asset classes, which would 
lower the returns shown. [2] Sharpe ratio is calculated using average returns over 2024 adjusted for the risk-free rate, divided by the standard deviation. [3] Sortino ratio is calculated using average returns over 2024 adjusted for the 
risk-free rate, divided by the downside deviation. [4] “CAGR”: Compound Annual Growth Rate. Source: ARK Investment Management LLC, 2025, based on data from Glassnode and Morningstar Direct as of December 31, 2024. For 
informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 33
ARK’s Big Ideas 2025: Bitcoin
    33/148
    0
200
400
600
800
1,000
1,200
1,400
1,600
1,800
2,000
$0
$10
$20
$30
$40
$50
$60
$70
$80
Jan-22 Jul-22 Jan-23 Jul-23 Jan-24 Jul-24
BTC
$ Millions
Miner Revenue
Miner Revenue ($) Miner Revenue (BTC)
Despite A Steep Decline In Miner 
Revenue After The Halving, Bitcoin’s 
Hash Rate Hit An All-Time High
Even though the halving cut bitcoin miner revenue by 50%, its hash 
rate—a proxy for network security—hit a record high. In other 
words, miners’ long-term conviction in bitcoin remains strong.
Miner revenue 
cut in half
0
100,000,000
200,000,000
300,000,000
400,000,000
500,000,000
600,000,000
700,000,000
800,000,000
900,000,000
2017 2018 2019 2020 2021 2022 2023 2024
TH/s
Hash Rate1
[1] In terahashes per second, a unit of measurement used to determine the estimated rate of processing power securing the Bitcoin network. It equals 1 trillion hashes per second. Source: ARK Investment 
Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any 
particular security or cryptocurrency. Past performance is not indicative of future results. 34
ARK’s Big Ideas 2025: Bitcoin
    34/148
    Transaction Count Soared, 
Thanks To Runes
Bitcoin’s daily transaction count hit a record high, thanks to the launch of the 
Runes protocol. Runes facilitates the creation of fungible tokens directly on the 
Bitcoin blockchain. 
0
100,000
200,000
300,000
400,000
500,000
600,000
700,000
800,000
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Change-Adjusted, 14-Day Average
Transaction Count
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
May-23 Aug-23 Nov-23 Feb-24 May-24 Aug-24 Nov-24
Distribution (%)
Transaction Count Share1
Runes BRC20 Inscriptions Ex BRC20 Regular Transactions
[1] “BRC20”: Token standard that enables the minting and transaction of fungible tokens via the Ordinals protocol on the Bitcoin network. Source: ARK Investment Management LLC, 2025, based on data from 
Dune Analytics and Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or 
cryptocurrency. Past performance is not indicative of future results. 35
ARK’s Big Ideas 2025: Bitcoin
    35/148
    Bitcoin Absorbed Major Selling Pressure In 2024
In January 2024, the German government 
seized 50,000 bitcoin linked to an online 
piracy group. Six months later, it sold the 
bitcoin. The market absorbed the supply 
successfully, and then the price of bitcoin 
rallied from $53,000 to $68,000.
Also, around mid-year, the long-awaited 
Mt. Gox1 creditor repayment process 
distributed more than 109,000 bitcoin, 
removing its largest overhang.
$0
$20,000
$40,000
$60,000
$80,000
$100,000
$120,000
0
50,000
100,000
150,000
200,000
250,000
Jan-24 Apr-24 Jul-24 Oct-24
Bitcoin Balance (BTC)
Bitcoin Price ($)
Major Bitcoin Distributions VS. Bitcoin Price
US Government Balance (BTC) Mt. Gox Balance (BTC)
German Government Balance (BTC) Bitcoin Price ($)
[1] Mt. Gox was a bitcoin exchange that once handled up to 70% of bitcoin’s total trading volume. It declared bankruptcy in 2014 after it was found to have run a fraudulent fractional reserve, incurring the loss of 
850k bitcoin from users’ funds at the time. During 2024, a large portion of reclaimed bitcoin was repaid to creditors, from an estimated 141,686 bitcoin at the start of the year to 27,877 by November 13. Data 
estimates sourced from Glassnode. Source: ARK Investment Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered 
investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 36
ARK’s Big Ideas 2025: Bitcoin
    36/148
    An Increasing Number Of Public 
Companies Now Hold Bitcoin
Public Company
(Top 10)
Bitcoin 
Holdings
USD Value
(Millions)
BTC Holdings 
As Percentage 
Of Market Cap
Microstrategy
(MSTR) 446,400 $41,780 58.7%
Marathon Digital
(MARA) 44,893 $4,201 73.8%
Riot Platforms
(RIOT) 17,429 $1,631 46.4%
Hut 8 Mining
(HUT) 10,096 $944 49.4%
Tesla
(TSLA) 9,720 $909 0.07%
Coinbase
(COIN) 9,480 $887 1.4%
CleanSpark
(CLSK) 9,297 $870 32.3%
Block
(SQ) 8,363 $782 1.4%
Bitcoin Group SE
(ADE) 3,589 $335 133.5%
Boyaa Interactive
(BOYAA/0434.HK) 3,183 $297 10.4%
$0
$10
$20
$30
$40
$50
$60
0
100
200
300
400
500
600
700
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
$ Billions
BTC, Thousands
Bitcoin On Public Company Balance Sheets
Bitcoin Balance USD Value
Source: ARK Investment Management LLC, 2025, based on data from Bitcoin Treasuries and TradingView as of December 31, 2024. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
Seventy-four public companies now hold bitcoin on their balance sheets. 
The value on corporate balance sheets quintupled during the past year, 
from $11 billion in 2023 to $55 billion.
37
ARK’s Big Ideas 2025: Bitcoin
    37/148
    Bitcoin’s Aggregate Cost Basis Hit A New High In 2024
In 2024, Bitcoin’s realized capitalization—or cost basis—grew by 86%, as its average acquisition cost hit a record high at $40,980 per 
bitcoin, or $811.7 billion in total.
$0
$500
$1,000
$1,500
$2,000
$2,500
2019 2020 2021 2022 2023 2024
$ Billions
Bitcoin Cost Basis
Market Cap ($) Realized Cap ($)
Source: ARK Investment Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or 
hold any particular security or cryptocurrency. Past performance is not indicative of future results. 38
ARK’s Big Ideas 2025: Bitcoin
    38/148
    Bitcoin’s Transaction Velocity1 And 
Holding Behavior Underscore Its 
Role As A Store Of Value
Bitcoin’s velocity dropped to a 14-year low in 2024, as supply 
held for three years or more reached an all-time high. 
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Percent Bitcoin Supply (%)
Supply Last Active 3+ Years Ago
1
10
100
1,000
2011
2012
2013
2014
2015
2016
2017
20182019
2020
2021
2022
2023
2024
Change-Adjusted, Annualized, 30-Day Average
Bitcoin's Velocity
[1] “Velocity”: a measure of how quickly bitcoin units circulate in the network, calculated by dividing the on-chain bitcoin transaction volume by the outstanding supply (annualized). Source: ARK Investment 
Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any 
particular security or cryptocurrency. Past performance is not indicative of future results. 39
ARK’s Big Ideas 2025: Bitcoin
    39/148
    $0
$200,000
$400,000
$600,000
$800,000
$1,000,000
$1,200,000
$1,400,000
$1,600,000
2030 Price Target
Bear Case
2030 Price Target
Base Case
2030 Price Target
Bull Case
($)
2030 Bitcoin Price Targets
(12/31/24 to 12/31/30 CAGR)
Institutional Investment Digital Gold
Emerging Market Safe Haven Nation-State Treasury
Corporate Treasury Bitcoin On-Chain Financial Services
Bitcoin Is On Pace To Meet Our 2030 Price Targets
Bear
Case
Base
Case
Bull
Case
Institutional Investment
TAM: Global Market Portfolio Ex Gold 1% 2.5% 6.5%
Digital Gold
TAM: Gold Market Cap 20% 40% 60%
Emerging Market Safe Haven
TAM: EM M2 Monetary Base 0.5% 2.5% 6%
Nation-State Treasury
TAM: Global Treasury Reserves Ex Gold 0.5% 2.5% 7%
Corporate Treasury
TAM: Global Cash & Cash Equivalents 1% 2.5% 10%
Bitcoin On-Chain Financial Services
TAM: L2s, LN, Sidechains, 
Restaking & WBTC
20%
CAGR
40% 
CAGR
60% 
CAGR
$300,000
CAGR ~21%
$710,000
CAGR ~40%
$1.5 million
CAGR ~58%
Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external sources as of December 31, 2024, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Forecasts are inherently 
limited and cannot be relied upon. 40
Price Target Assumptions
ARK’s Big Ideas 2025: Bitcoin
    40/148
    Reshaping The 
Digital Asset Space
Stablecoins
Yassine Elmandjra Lorenzo Valente
FORMER DIRECTOR OF 
DIGITAL ASSETS
David Puell
RESEARCH TRADING ANALYST &
ASSOCIATE PORTFOLIO MANAGER,
DIGITAL ASSETS
DIRECTOR OF 
DIGITAL ASSETS
Frank Downing
DIRECTOR OF RESEARCH, 
NEXT GENERATION INTERNET
41
ARK’s Big Ideas 2025
    41/148
    One Of The Fastest-Growing Segments In Digital Assets, Stablecoins 
Overtook Mastercard And Visa In Transaction Value In 2024
Despite a two-year bear market and a 70%+ decline in market capitalization, stablecoin growth has remained uninterrupted. In 2024, 
the annualized transaction value of stablecoins hit $15.6 trillion—roughly 119% and 200% that of Visa and Mastercard, respectively. The 
number of transactions hit 110 million monthly, roughly 0.41% and 0.72% of those processed by Visa and Mastercard, respectively. In 
other words, the stablecoin value per transaction is much higher than that for Visa and Mastercard.
-0.50
0.50
1.50
2.50
3.50
4.50
0
20
40
60
80
100
120
140
Jan-20
May-20
Sep-20
Jan-21
May-21
Sep-21
Jan-22
May-22
Sep-22
Jan-23
May-23
Sep-23
Jan-24
May-24
Sep-24
$ Trillions
Millions
Stablecoin Transaction Count VS. 
Total Digital Asset Market Capitalization
Stablecoin Transaction Count (lhs)
Digital Asset Market Capitalization (rhs)
$0
$2
$4
$6
$8
$10
$12
$14
$16
$18
2019 2020 2021 2022 2023 2024
$ Trillions 
Transaction Value: 
Stablecoins VS. Traditional Payment Processors
Visa Mastercard American Express Stablecoins
42
Note: In the lefthand chart, stablecoins data are for calendar year, while credit card data are for fiscal year. Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external 
data sources, including Wall Street Zen, CoinGecko, and Visa OnChain Analytics, as of December 31, 2024, which may be provided upon request, For informational purposes only and should not be considered 
investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: Stablecoins
    42/148
    0
0.5
1
1.5
2
2.5
3
Dec-23
Jan-24
Feb-24
Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Dec-24
$ Trillions 
Stablecoin Volume By Blockchain
Solana Tron Ethereum Base Arbitrum BNB Chain Avalanche Optimism Polygon POS TON Celo
Stablecoin Volumes Hit Record Highs In December 2024
Solana, Tron, Ethereum, 
and Base are the 
leading blockchains 
that drove stablecoin 
volume in 2024.
December 2024 set new 
records with a daily 
volume of $270 billion 
and a monthly volume 
of $2.7 trillion, 
highlighting the sector's 
rapid growth.
3.8 x
Source: ARK Investment Management LLC, 2025., based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, 
sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Stablecoins
43
    43/148
    Ethena Labs Tokenized A Popular 
Basis Trade And Amassed $6 Billion 
In 12 Months
Innovation in the stablecoin space is thriving, with Ethena Labs as 
one of the fastest-growing projects. Despite criticism of its novel 
design, the project has captured significant share of the non-fiatbacked stablecoin market and has become a key competitor in this 
$200 billion space.
Ethena Labs reached $6 billion in total value 
locked during its first twelve months.
By tokenizing a delta-neutral position, Ethena 
Labs can offer yields as high as 20%-30%, 
depending on market conditions. The yield 
could turn negative in a bear market.
Having captured 10% of total Ether open 
interest, Ethena Labs is expanding into Solana 
and Bitcoin delta-neutral trades.
User provides 
$100 USD worth 
of ETH 
Ethena swaps the collateral for 
$100 worth of yield bearing ETH 
such as stETH which is deposited 
at a custodian
An Exchange opens a 
short position worth 
$100 with the deposit 
at the custodian as 
collateral
Delta neutral position in 
which the short position 
offsets the spot collateral
$100 USDE are 
minted to the user
User
Note: A delta neutral position in finance is a strategy designed to hedge against price movements in the underlying asset. Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
1
2
3
4
5
44
ARK’s Big Ideas 2025: Stablecoins
    44/148
    UDST (Tether) continues to dominate the stablecoin landscape, followed by USDC (Circle). Together, they account for 90% of the total 
supply. Stablecoins are “multichain” and have penetrated almost every major Layer 1 blockchain. At $203 billion, the stablecoin supply 
represents ~0.97% of M2* money supply in the US. During December, active stablecoin addresses hit 23 million, an all-time high. As 
measured by monthly active addresses, Tron is the leading network, favored in emerging markets for its low transaction fees.
After A Drawdown In 2023, The Supply Of Stablecoins And The Number Of 
Active Stablecoin Addresses Hit All-Time Highs In 2024
ARK’s Big Ideas 2025: Stablecoins
*M2 is a measure of the US money stock that includes M1 (currency and coins held by the non-bank public, checkable deposits, and travelers’ checks) plus savings deposits (including money market deposit accounts), small 
time deposits under $100,000, and shares in retail money market mutual funds. Source: ARK Investment Management LLC, 2025, based on data from rwa.xyz (“Stablecoins”) as of December 31, 2024. For informational purposes 
only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
0
5
10
15
20
25
30
Jan-18
Apr-18 Jul-18
Oct-18 Jan-19Apr-19 Jul-19Oct-19
Jan-20
Apr-20 Jul-20
Oct-20 Jan-21
Apr-21 Jul-21Oct-21Jan-22
Apr-22 Jul-22
Oct-22 Jan-23
Apr-23 Jul-23
Oct-23 Jan-24
Apr-24 Jul-24Oct-24
$ Millions
Monthly Active Stablecoin Addresses By Network
TRON Binance Smart Chain Solana
Ethereum Arbitrum Celo
Polygon Optimism Aptos
Avalanche NEAR Mantle
XDC Base
$0
$50
$100
$150
$200
$250
Jan-20
Jan-21
Jan-22
Jan-23
Jan-24
$ Billions
Stablecoin Supply
USDT USDC USDS USDe FDUSD
45
    45/148
    Digital Assets are undergoing “stablecoinization” and “dollarization,” while China and Japan have sold record volumes of US Treasuries, Saudi Arabia has ended 
its 45-year petrodollar agreement, and BRICS* nations have reduced their reliance on US dollar payments by circumventing the SWIFT network. Historically, 
bitcoin and ether were the bridges into the digital asset ecosystem. During the past two years, however, stablecoins have dethroned them and now account for 
35% to 50% of on-chain transaction volumes.
Dollar-pegged stablecoins dominate, accounting for 98%+ of the supply, followed by Gold-backed and Euro-backed stablecoins, with ~1% and ~0.5%, 
respectively. ARK’s research suggests that the market will expand and include Asian currency-backed stablecoins.
As Some Countries Shift Away From The Dollar, 
Digital Assets Are Moving Toward It 
0%
20%
40%
60%
80%
100%
Sep-19
Dec-19
Mar-20
Jun-20
Sep-20
Dec-20
Mar-21
Jun-21
Sep-21
Dec-21
Mar-22
Jun-22
Sep-22
Dec-22
Mar-23
Jun-23
Sep-23
Dec-23
Mar-24
Jun-24
Sep-24
Dec-24
Share
Settlement Volume Across Different Blockchains 
Stablecoins L1 L2 Bitcoin Ethereum
*BRICS refers to five major emerging economies—Brazil, Russia, India, China, and South Africa. Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational 
purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 46
ARK’s Big Ideas 2025: Stablecoins
    46/148
    Stablecoins Are Attracting Retail Interest, Thanks To The Lower Cost 
And Higher Efficiency Of Layer 2s
Retail investors are flocking to Layer 2s for less expensive and more convenient stablecoin transactions, increasing the market share of 
Blockchains like Arbitrum, Base, and Optimism. Meanwhile, whales* and institutions continue to operate on Ethereum's base layer.
Transactions below $100 dominate activity on Base and Optimism, while those above $100 dominate Ethereum’s base layer.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Jan-22
Mar-22
May-22
Jul-22
Sep-22
Nov-22
Jan-23
Mar-23
May-23
Jul-23
Sep-23
Nov-23
Jan-24
Mar-24
May-24
Jul-24
Sep-24
Nov-24
Stablecoin Transaction Count Share By Blockchain
Ethereum Arbitrum Base Optimism Celo
0% 20% 40% 60% 80% 100%
Celo
Optimism
Base
Arbitrum
Ethereum
Transaction Size Share By Blockchain In 2024 
Less than $100 Between $100 and $1K
Between $1K and $10K Between $10k and $100K
Between $100k and $1M Between $1M and $10M
Over $10M
*“Whale”: an individual or entity that holds a large amount of a specific cryptocurrency, usually north of 5-10 million dollars depending on the asset. Source: ARK Investment Management LLC, 2025, based on data from 
Visa OnChain Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. 
Past performance is not indicative of future results. 47
ARK’s Big Ideas 2025: Stablecoins
    47/148
    Externally Owned Accounts (EOAs)—
standard Ethereum addresses used for 
peer-to-peer (P2P) transactions and 
storage—account for 60% of USDC’s 
usage, while centralized exchanges 
account for 11%, bridges across Layer 2 
solutions for 7%, and decentralized 
exchanges (DEXs) and money markets
each for 1.7%.
As DeFi usage surges in the next few 
years, DEXs, Bridges, and Money 
Markets are likely to retake market 
share from P2P.
While usage in lending/borrowing 
markets, DEXs, and Bridges fluctuates 
with market cycles, P2P transactions 
and storage are more resilient because 
of product-market fit beyond trading.
Peer-to-Peer Transactions And Personal Wallet 
Storage Dominate Stablecoin Use Cases
Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, 
or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 48
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Jan-21
May-21
Sep-21
Jan-22
May-22
Sep-22
Jan-23
May-23
Sep-23
Jan-24
May-24
Sep-24
Share 
Top Addresses Storing USDC
EOA CeFi Other Bridge Treasury DEX Lending Yield Farm
ARK’s Big Ideas 2025: Stablecoins
    48/148
    Four Stablecoin Issuers Dominate The Revenue Generated In Digital Assets
With fewer than 200 employees, Tether 
reported $5.2 billion in profits during 
the first half of 2024, including USDT, 
the rest of its products and services, 
and unrealized gains on its Digital 
Assets—clearly one of the most capitalefficient businesses in history.
Tether (USDT) and Circle (USDC) 
account for 60% of the revenue 
generated by the top five networks and 
applications.
Collectively, stablecoins USDT, USDC, 
DAI/USDS, and USDE generated 
revenues of $3.35 billion, or $6.7 billion 
at an annual rate, during the second 
half of 2024.
0
500
1000
1500
2000
2500
3000
Tether (usdt)
Tron
Circle (usdc)
Ethereum
Solana
Aerodrome
Sky (usds/dai)
Ethena labs (usde)
Lido Finance
Aave
Uniswap
Pancakeswap
$ Millions
Top 12 Blockchain Networks And Applications By Revenue*
Stablecoin Issuers Others
*Second Half of 2024. Source: ARK Investment Management LLC, 2025, based on data from Token Terminal of December 31, 2024. For informational purposes only and should not be considered investment advice or 
a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
Tether (USDT)
Circle (USDC)
Sky (USDS/DAI)
Ethena Labs (USDE)
49
ARK’s Big Ideas 2025: Stablecoins
    49/148
    Tether’s Financial Performance Is Stunning, Both Absolutely And Relatively
Tether employs fewer than 200 people compared to 300,000+ each at JP Morgan and Berkshire Hathaway. The only institutions 
in the S&P Financial Select Sector Index generating more net income than Tether in the first half of 2024 were Berkshire 
Hathaway, JP Morgan, Bank of America, and Wells Fargo.
1
10
100
1000
10000
100000
1000000
10000000
100000000
0
5
10
15
20
25
30
35
Tether Visa Mastercard Berkshire
Hathaway
Goldman
Sachs
JP Morgan Amercan
Express
Bank of
America
Progressive S&P Global Wells Fargo
$, Log Scale
$ Billions
Tether VS. Leading Financial Firms*
Net Income (lhs) Net Income Per Employee (rhs)
*First Half of 2024. Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external data, sources as of December 31, 2024, which may be provided upon request. For informational purposes 
only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 50
ARK’s Big Ideas 2025: Stablecoins
    50/148
    In a world trending toward deglobalization and de-dollarization, 
stablecoins could drive steady 
demand for US Treasury securities. 
Combined, Tether and Circle already 
ranked as the 20th largest holder of US 
Treasuries as of December 2024.
In populous emerging markets like 
Brazil, Nigeria, Turkey, Indonesia, and 
India, individuals and companies are 
adopting stablecoins as a store of 
value, means of payment, and crossborder currency. Stablecoins could 
become one of the most effective 
ways to export US dollars.
Counterbalancing ”De-Dollarization,” Stablecoins Are Increasing 
The Demand For US Government Debt As Collateral
0
200
400
600
800
1000
1200
Japan
China
United Kingdom
Cayman Islands
Luxembourg
Canada
Belgium
France
Ireland
Switzerland
Taiwan
Singapore
India
Brazil
Hong Kong
Norway
Saudi Arabia
Korea, South
Germany
Tether & Circle
Mexico
$ Billions
Top Foreign Holders Of US Treasury Securities As Of September 2024
Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external data sources, including Tether, Circle, and TicData Treasury, as of December 31, 2024, which may be provided upon request. 
For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 51
ARK’s Big Ideas 2025: Stablecoins
    51/148
    Circle and Tether have been generating billions in revenue from 
Treasury bills and other securities serving as collateral for their 
stablecoins. In 2024, however, in response to competition and 
demand, stablecoins operating outside of the US began to pass a 
significant portion of their interest income to users.
Circle and Tether are unlikely to follow the trend until absolutely 
necessary. Though still small, yield-bearing stablecoins are the 
fastest-growing category in the stablecoin market.
Stablecoin Issuers Are Transferring Risk-Free Rates To Users
Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 52
0
1
2
3
4
5
6
7
8
Mar-23
May-23
Jul-23
Sep-23
Nov-23
Jan-24
Mar-24
May-24
Jul-24
Sep-24
Nov-24
$ Billions
Size Of The Yield Bearing Stablecoin Market
sUSDe USD0 sDAI USDY deUSD
USDB sFRAX USDM USDL rUSD
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Jan-23
Apr-23
Jul-23
Oct-23
Jan-24
Apr-24
Jul-24
Oct-24
Share Of Interest-Bearing VS. Non Interest Bearing Stablecoins
Yield Bearing Non Yield-Bearing
ARK’s Big Ideas 2025: Stablecoins
    52/148
    203
1409
0.00%
0.10%
0.20%
0.30%
0.40%
0.50%
0.60%
0.70%
0.80%
0.90%
1.00%
$0
$200
$400
$600
$800
$1,000
$1,200
$1,400
$1,600
2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
$ Billions
Stablecoin Supply VS. Fiat Global M2
2019 To 2030 (2025-2030 Forecast) 
Stablecoin Supply (lhs) Stablecoins as Percentage of Global M2 Supply (rhs)
Stablecoins Could Increase From 0.17% to 0.9% Of Fiat* 
Global M2 Supply By 2030
Currently $203 billion, or 
0.17% of global M2** supply, 
stablecoins could grow to 
$1.4 trillion and 0.9%, 
respectively, by 2030.
If so, stablecoins would be 
the 13th largest currency in 
circulation, behind Spain 
and ahead of Netherlands.
38% CAGR
53
Note: “CAGR”: Compound Annual Growth Rate. *Fiat money is a type of currency that is issued by a government and has value because the government declares it to be legal tender. **M2 is a measure of the US 
money stock that includes M1 (currency and coins held by the non-bank public, checkable deposits, and travelers’ checks) plus savings deposits (including money market deposit accounts), small time deposits 
under $100,000, and shares in retail money market mutual funds. Source: ARK Investment Management LLC, 2025, based on data from rwa.xyz (“Stablecoins”) as of December 31, 2024. For informational purposes 
only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: Stablecoins
    53/148
    Enabling Steep Cost 
Declines And New Use-Cases 
At The Application Layer
Scaling 
Blockchains 
Yassine Elmandjra Lorenzo Valente
FORMER DIRECTOR OF 
DIGITAL ASSETS
David Puell
RESEARCH TRADING ANALYST &
ASSOCIATE PORTFOLIO MANAGER,
DIGITAL ASSETS
DIRECTOR OF 
DIGITAL ASSETS
54
Frank Downing
DIRECTOR OF RESEARCH, 
NEXT GENERATION INTERNET
ARK’s Big Ideas 2025
    54/148
    Navigating The Smart 
Contract Landscape
As the digital asset space becomes increasingly complex, smart contracts are driving innovation 
across a growing number of sectors. The ecosystem is evolving rapidly to meet diverse and 
dynamic needs—from user-focused applications like gaming and SocialFi, to advanced financial 
tools like derivatives and structured products, to decentralized infrastructure networks powering 
wireless connectivity and energy storage.
Consumer Applications
Gaming Ticketing Smart 
Devices 
Trading 
Apps
Decision 
Markets SocialFi Music Messaging Prediction Markets Metaverse
Community And Culture
Memecoins NFTs Media And 
Entertainment
Community 
Accelerators
Conference 
and Events DAOs Generative Art
Clubs And 
Collectibles
DeFi
RWA
Derivatives 
And 
Perpetuals
NFT 
Marketplaces 
and Tooling
Trading 
Bots Launchpads DEX/CEX/HEX Restaking Liquid Staking
Asset 
Creators/ 
Issuers
NFT Lending And 
Fractionalization Stablecoins Money 
Markets
OTC/
RFQ
Derivatives 
And Spot 
Aggregators
YieldStructured 
Products
Indexes Insurance Risk 
Management
DePIN
Mapping Wireless Storage Environment Compute Data 
Management Energy Services
Crypto x AI
Compute 
Marketplaces
AI 
Agents
Data 
Availability Gaming Privacy–ZK-FHE
Coordination 
Networks
CoProcessors
Model 
TrainingInference
Model 
Creation
Infrastructure And Developer Tooling
Wallets DAO 
Tooling Oracles
Developer 
Tooling 
And 
Standards
Clients 
And Nodes
Data And 
Analytics
Bridges 
And CrossChain 
Messaging
Layer 1s RaaS-Shared 
Sequencers
Web 3 
Identity
Payment 
And Blink 
Tools
MEV 
Clients 
And 
Tooling
Privacy Testing and 
Frameworks
Custodial 
Services
Layer 2s, 
Layer 3s, 
and RaaS 
Onramp 
Tools
Auditing 
and Bug 
Bounties
Roll-Up 
Frameworks
PoW And 
PoS 
Services
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency.
ARK’s Big Ideas 2025: Scaling Blockchains
55
    55/148
    Applications Seeking Lower Fees And Greater Efficiency Deploy Either On 
Solana’s High Throughput—Layer 1—Or On Ethereum’s Layer 2s 
Layer 1
Blockchain
Layer 2
Blockchain
Application
Layer
User Aggregation 
Layer
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 
Past performance is not indicative of future results. 56
Ledger Metamask Coinbase Wallet
Arbitrum Base Polygon
Solana Ethereum
Pendle
Interest rate 
Marketplace
GMX
Perpetual 
Exchange
Polymarket
Prediction
Market
Aave
Money
Market
Morpho
Borrow/Lend 
Infrastructure
Jupiter
Derivatives 
Exchange
Jito
Payment For 
Order Flow 
Orca
Spot 
Exchange
ARK’s Big Ideas 2025: Scaling Blockchains
    56/148
    0
0.5
1
1.5
2
2.5
Dec-23
Jan-24
Feb-24
Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Dec-24
$
Ethereum Layer 2s Median Fee
Arbitrum Optimism Base zkSync Scroll Blast Linea
One of Ethereum's most significant technical upgrades to date, EIP 4844, 
impacted Layer 2 networks by enabling faster, less-expensive transactions.
Ethereum’s EIP 4844 Technical 
Upgrade Cut Transaction Costs 
By 10x, Spurring Use
Prior to the EIP 4844 upgrade, 
the average transaction fee on 
Ethereum Layer 2s was ~$0.50. 
Now, users pay ~$0.05.
EIP 4844 was the first 
milestone in a roadmap that 
should enable a 250- to 1250-
fold increase in transactions 
per second (TPS) from ~400 
today to 100,000-500,000.
More than 200 Layer 2 projects 
have launched, with others on 
the way.
EIP 4844 Upgrade
Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, 
or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 57
ARK’s Big Ideas 2025: Scaling Blockchains
    57/148
    Plummeting Transaction Costs Have Led To A Boom In Layer 2 Activity, 
Pulling Users Away From Ethereum’s Base Layer
Of the Daily Active Addresses transacting on the Ethereum ecosystem, Layer 2s have captured 85% share. The activity on Layer 2s 
enabled Ethereum to scale daily transactions by 400% from 3 million to 15 million in 2024.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dec-21
Mar-22
Jun-22
Sep-22
Dec-22
Mar-23
Jun-23
Sep-23
Dec-23
Mar-24
Jun-24
Sep-24
Dec-24
Share
Daily Active Addresses: Ethereum VS. Layer 2s
Ethereum Top 10 Ethereum L2s
0
5
10
15
20
25
30
Dec-21
Feb-22
Apr-22
Jun-22
Aug-22
Oct-22
Dec-22
Feb-23
Apr-23
Jun-23
Aug-23
Oct-23
Dec-23
Feb-24
Apr-24
Jun-24
Aug-24
Oct-24
Number (Millions)
Daily Transactions: Ethereum VS. Layer 2s
Ethereum Top 10 Ethereum L2s Combined
Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, 
sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 58
ARK’s Big Ideas 2025: Scaling Blockchains
    58/148
    Within one year of launch, Base outpaced all other Ethereum Layer 2 solutions in growth and market share. In 2024, Base captured
46% of active users and generated 63% of fees among Ethereum Layer 2s. With $15 billion in total value locked and more than 300 
applications deployed, Base has been contributing significantly to Coinbase’s cash-flows.
Base Is The Fastest Growing Ethereum Layer 2 Blockchain 
Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 59
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dec-23
Jan-24
Feb-24
Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Share
Layer 2 Daily Active Addresses
Base Celo Arbitrum Mantle OP Mainnet
Linea zkSync Era Blast Scroll StarkNet
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Dec-23
Jan-24
Feb-24
Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Share
Layer 2 Fees
Base Arbitrum OP Mainnet Linea Mantle
StarkNet zkSync Era Blast Scroll Celo
ARK’s Big Ideas 2025: Scaling Blockchains
    59/148
    Institutions, high-value users, and whales are settling their 
transactions primarily on the Ethereum base layer. The unit 
economics of Ethereum’s base layer, as measured by Total Value 
Locked (TVL) and Decentralized Exchange (DEX) volume per 
user, are unparalleled.
Despite Migration To Layer 2s, 
Ethereum's Base Layer Still Dominates 
High-Value Storage And Settlement
Note: “Whale”: an individual or entity that holds a large amount of a specific cryptocurrency, usually north of 10 million dollars depending on the asset. Source: ARK Investment Management LLC, 2025, based on 
data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or 
cryptocurrency. Past performance is not indicative of future results. 60
$0
$20
$40
$60
$80
$100
$120
$140
$160
$180
$200
Dec-23
Jan-24
Feb-24Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
$ Thousands 
Total Value Locked Per Daily Active Address
Ethereum StarkNet Blast Scroll Celo Linea Mantle OP Mainnet zkSync Era
ARK’s Big Ideas 2025: Scaling Blockchains
    60/148
    When Binance settled with the U.S. Securities and Exchange Commission (SEC) for more than $4 billion and its CEO stepped down in 
2024, centralized exchanges (CEXs) lost some share to decentralized exchanges (DEXs). From January to their peaks, spot and 
derivatives DeFi volume nearly doubled their market share from 8% and 3%, respectively, to 14% and 8%, reaching all-time highs. 
During the same period, Binance’s market share among centralized exchanges dropped from 62% to 35%.
As Decentralized Finance (DeFi) Surged To All-Time Highs, Decentralized 
Exchanges Challenged Centralized Venues In Both Spot And Derivatives Trading 
0%
2%
4%
6%
8%
10%
12%
14%
16%
Jan-20
Apr-20
Jul-20
Oct-20
Jan-21
Apr-21
Jul-21
Oct-21Jan-22
Apr-22
Jul-22
Oct-22
Jan-23
Apr-23
Jul-23
Oct-23
Jan-24
Apr-24
Jul-24
Oct-24
DEX Market Share By Trading Volume
Spot Trading Volume Derivatives Trading Volume
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Jan-23
Mar-23
May-23
Jul-23
Sep-23
Nov-23
Jan-24
Mar-24
May-24
Jul-24
Sep-24
Nov-24
Share
Centralized Exchange Market Share By Volume
Binance Others
Source: ARK Investment Management LLC, 2025, based on data from The Block as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation 
to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 61
ARK’s Big Ideas 2025: Scaling Blockchains
    61/148
    Decentralized exchanges (DEXs) like Uniswap, Aerodrome, and Raydium leverage the efficiency of small, agile teams to develop and
maintain core protocol infrastructure. With massive efficiency advantages over centralized venues, DEXs have one-tenth the headcount. 
With ~9,000 employees, Binance leads the centralized exchange space headcount.
Smart Contract-Powered Exchanges Are 5-10 Times More Efficient Than Their 
Centralized Counterparts
*The Spot Volume per Employee data are from the month of November 2024. Source: ARK Investment Management LLC, 2025, based on data from Token Terminal, Coin Gecko, and Pitchbook as of November30, 2024. For 
informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results.
0
500
1000
1500
2000
2500
OKX Coinbase Crypto.com Binance Bybit Orca Uniswap Aerodrome Pancakeswap Raydium
$ Millions
Monthly Spot Volume Per Employee*
62
Centralized Exchanges Decentralized Exchanges
ARK’s Big Ideas 2025: Scaling Blockchains
    62/148
    Liquid staking and restaking have become the preferred methods for generating yield on Ether. Thanks to their yield-bearing properties, 
liquidity, and accessibility, they now account for 40% of ETH staking. Demand for rehypothecating staked ETH to increase yield has 
created restaking platforms that now account for ~5.5 million Ether, or 17% of staked ETH. 
Liquid Staking And Restaking Protocols Are The Preferred Solutions For 
Staking And Rehypothecating ETH
0
1
2
3
4
5
6
Jun-23
Jul-23
Aug-23
Sep-23
Oct-23
Nov-23
Dec-23
Jan-24
Feb-24
Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Dec-24
Millions
ETH Staked In Leading Restaking Platforms
Eigenlayer Symbiotic Karak
Note: Staking involves locking up a cryptocurrency in a wallet to support the operations of a blockchain network, typically for Proof-of-Stake (PoS) or similar consensus mechanisms. Liquid staking is an enhancement of traditional 
staking where users can stake their tokens while still having access to liquidity. Restaking refers to using the same staked assets or their derivatives for multiple purposes or layers of staking. Rehypothecation occurs when assets (such as staked tokens or derivatives) are used as collateral multiple times by various entities. Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes 
only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 63
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Jan-21
Mar-21 May-21 Jul-21
Sep-21 Nov-21 Jan-22
Mar-22 May-22 Jul-22
Sep-22 Nov-22 Jan-23Mar-23 May-23 Jul-23
Sep-23 Nov-23 Jan-24
Mar-24 May-24 Jul-24
Sep-24 Nov-24
Share
Ethereum Staking Methods
Liquid Staking CEXs Unidentified
Staking Pools Liquid Restaking Solo Stakers
ARK’s Big Ideas 2025: Scaling Blockchains
    63/148
    Led By Polymarket, Prediction Markets Became The Breakout Consumer 
Application In 2024
During 2024, Polymarket surpassed three million cumulative users and $1.2 billion in monthly volume, with elections and politics driving 
70% of its activity. After a brief post-election pull-back, the number of unique daily users has recovered to 50,000 and volume is trending 
back up with Sports being the most important category.
0
0.5
1
1.5
2
2.5
3
3.5
0
10
20
30
40
50
60
Jan-24
Feb-24Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Dec-24
Millions
Thousands
Number Of Unique Addresses Betting On Polymarket
Unique Addresses (lhs) Cumulative Unique Addresses (rhs)
$0
$200
$400
$600
$800
$1,000
$1,200
20-May-24
3-Jun-24
17-Jun-24
1-Jul-24
15-Jul-24
29-Jul-24
12-Aug-24
26-Aug-24
9-Sep-24
23-Sep-24
7-Oct-24
21-Oct-24
4-Nov-24
18-Nov-24
2-Dec-24
16-Dec-24
$ Millions
Polymarket Volume By Category
Elections Politics Unspecified Sports
Science Pop Culture Crypto Business
Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 64
ARK’s Big Ideas 2025: Scaling Blockchains
    64/148
    0
100
200
300
400
500
600
700
800
900
Oct-23
Nov-23
Dec-23
Jan-24
Feb-24Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Dec-24
Number
Transactions Per Second
Solana Ethereum + L2s
Thanks to Layer 2s and the EIP 
4844 upgrade, Ethereum’s 
throughput has more than 
doubled to ~200 transactions per 
second (TPS). Nonetheless, Solana 
continues to outperform, with an 
average throughput of ~800TPS. 
Solana’s success is based on a set 
of trade-offs, including more 
expensive hardware 
requirements and parallel 
transaction processing.
Solana’s new Firedancer client 
could increase its throughput to 
hundreds of thousands of TPS. 
Solana’s Throughput Is Higher Than Ethereum’s, Thanks To 
High-Performance Design And Strategic Trade-Offs
EIP 4844 Upgrade
Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 65
ARK’s Big Ideas 2025: Scaling Blockchains
    65/148
    After hitting bear market lows at $8 in 2023, Solana has turned around dramatically relative to other Layer 1s. Daily active users, revenue, 
transaction count, and total value locked (TVL) reached all-time highs or grew by an order of magnitude. Solana is the only Layer 1 that 
competes with Ethereum and Bitcoin on metrics like daily active addresses and revenue.
Thanks To Retail Adoption, Solana Has Gained Share 
Based On Several Metrics
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Oct-23
Dec-23
Feb-24
Apr-24
Jun-24
Aug-24
Oct-24
Dec-24
Share
Daily Active Addresses
Solana Bitcoin Ethereum + L2s
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Sep-23
Oct-23
Nov-23
Dec-23
Jan-24
Feb-24Mar-24
Apr-24
May-24
Jun-24
Jul-24
Aug-24
Sep-24
Oct-24
Nov-24
Share
Fees Generated
Solana Bitcoin Ethereum + L2s
Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, 
sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 66
ARK’s Big Ideas 2025: Scaling Blockchains
    66/148
    Of the 39,139 new crypto developers in 2024, Solana led the way with 7,625, surpassing Ethereum Mainnet. With 4,287 developers in total, 
Base was sixth overall, surpassing both Arbitrum and Starknet as the leading Layer 2 solution on Ethereum.
Solana And Base Are Leading The Way In Developer Adoption And Mindshare
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
Solana
Ethereum
Other
Internet Computer
Aptos
Base
Bitcoin
Sui
Near
Polkadot
Polygon
Starknet
Arbitrum
Number Of Net New Crypto Developers In 2024 
0
2000
4000
6000
8000
10000
12000
14000
16000
Ethereum Mainnet
Base
Arbitrum
Starknet
Optimism
Scroll
Zksync
Polygon Zkevm
Number Of Crypto Developers
In The Ethereum Ecosystem In 2024
Source: ARK Investment Management LLC, 2025, based on data from Electric Capital 2024 as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to 
buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 67
ARK’s Big Ideas 2025: Scaling Blockchains
    67/148
    Robotaxis
Transforming Personal 
Mobility While Lowering 
Costs And Enhancing Safety
Tasha Keeney, CFA Daniel Maguire, ACA
DIRECTOR OF INVESTMENT ANALYSIS
& INSTITUTIONAL STRATEGIES
RESEARCH ANALYST, AUTONOMOUS 
TECHNOLOGY & ROBOTICS
68
ARK’s Big Ideas 2025
    68/148
    Electric Vehicles Continue To Take Share From 
Internal Combustion Engine Vehicles
4% 2%
-2% -5%
-16%
-1%
-10%
4%
-1%
57%
67% 70%
18%
33%
113%
59%
30%
12%
-40%
-20%
0%
20%
40%
60%
80%
100%
120%
2016 2017 2018 2019 2020 2021 2022 2023 2024e
Year-Over-Year Percent Change
Global Vehicle Sales Growth
Internal Combustion Battery Electric
10
20
30
40
50
60
70
80
90
100
2016 2017 2018 2019 2020 2021 2022 2023 2024e
Units (Millions)
Global Vehicle Sales
Internal Combustion Battery Electric
Slide by Sam Korus, Director and Akaash TK, Associate, Autonomous Technology & Robotics. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including EVVolumes as of January 3, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 
Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 69
ARK’s Big Ideas 2025: Robotaxis
    69/148
    The Future Is Autonomous Electric Vehicles
Despite the recent slowdown in the growth of electric vehicles, the trend is clear: the internal combustion engine is on its way out.
Slide by Sam Korus, Director and Akaash TK, Associate, Autonomous Technology & Robotics. Note: Percentages may not total 100% due to rounding. Source: ARK Investment Management LLC, 2025, based on 
data from EV-Volumes as of January 3, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance 
is not indicative of future results. 70
92% 87%
81% 77% 72%
4% 5%
6% 7% 8%
1% 2%
4%
5% 7%
3% 6%
10%
12%
13%
50
100
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024e
Unit (Millions)
Global Vehicle Sales And Share By Segment
Internal Combustion Hybrid Plug-in Hybrid Battery
ARK’s Big Ideas 2025: Robotaxis
    70/148
    After More Than A Decade, Most Automakers Have Dropped Out Of The 
Autonomy Race. Tesla And Waymo Remain.
ARK’s 2014 Estimates Of Autonomous Commercial Launch
In 2014, many automakers expected to debut an autonomous vehicle by 2020. Only Waymo delivered, launching its first commercial 
autonomous rides in 2018. Tesla plans to launch in 2025. Waymo’s launch and Tesla’s internal testing of robotaxis suggest that 2025 
could be the standout year in which consumers and businesses agree that the future of transportation is autonomous.
Outcome
Tesla
Volvo
Google
GM
Audi
Mercedes
Nissan
BMW
Daimler
Continental
2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
2025 LAUNCH PLANNED
LAUNCHED IN 2018
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
NO EXPECTATION OF LAUNCH
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 71
ARK’s Big Ideas 2025: Robotaxis
    71/148
    $2.10
$1.10 $1.10 $1.10
$0.25
1871 1934 1950 2016 2021
Cost Per Mile Of A Personally Owned Vehicle
(2024 $)
Autonomous Ride-Hail Should Increase Access To Convenient 
Point-To-Point Transportation
Adjusted for inflation, the cost of owning and operating a personal car has not changed since the Model T rolled off the first assembly 
line more than 100 years ago. ARK estimates that autonomous taxis at scale could cost consumers as little as $0.25 per mile, spurring 
widespread adoption.
Note: Figures are rounded. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Ulvog 2012, Model T Ford Forum 2010, American Automobile Association 
1950, and American Automobile Association 2024, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to 
buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 72
2024 2035
ARK’s Big Ideas 2025: Robotaxis
    72/148
    Autonomous Ride-Hail Will Lower The Cost Of Point-To-Point Transportation 
In early days, robotaxi companies are likely to price rides close to those for human-driven ride-hail. With scale and increased utilization, 
autonomous technology is likely to drive the cost per mile below that of personal vehicle travel. 
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 73
$2.00 
$1.10 
$0.35 $0.25 
Human Driven Ride-Hail
US, 2024
Personal Car 
US, 2024
Human Driven Ride-Hail
China, 2024
Robotaxi 
Global, 2035
Cost Per Mile
(2024 $)
Human-Driven Ride-Hail
US, 2024
ARK’s Big Ideas 2025: Robotaxis
    73/148
    The US And China Are Blazing The Robotaxi Trail, 
Followed By The Middle East And Europe
74
United States
• Tesla
• Waymo
• Zoox
• Wayve
• Pony.ai
• WeRide
China
Saudi Arabia & UAE
United Kingdom
• Wayve
• Baidu Apollo Go
• Tesla
• Pony.ai
• WeRide
• Pony.ai
• Baidu Apollo Go
• WeRide
Price competition in China could continue 
to push local players into more lucrative 
partnerships abroad. 
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: Robotaxis
    74/148
    Robotaxis Already Have Hit 12 Million Rides At An Annual Rate
Waymo and Baidu are responsible for most of the autonomous rides in the US and China, respectively, as shown on the left below. With 
the largest data lake of all players globally, its proprietary data advantage should give Tesla a competitive edge if it launches ride-hail 
operations as planned in 2025.
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
Waymo Baidu Apollo Go Pony.ai Tesla Zoox
Million Rides (Annualized)
Autonomous Commercial Rides
Run Rate
75
~ 37 ~ 16 ~1
~ 3,500 *
~0.1
 -
 500
 1,000
 1,500
 2,000
 2,500
 3,000
 3,500
 4,000
Waymo Baidu Apollo
Go
Pony.ai Tesla Zoox
Million Miles (Annualized)
Autonomous Miles
Run Rate
Baidu Apollo Go
*Tesla Full Self-Driving (FSD) currently is supervised and requires a human behind the wheel. Waymo opened driverless operations to the pubic in October 2020 and now completes ~175,000 weekly rides. Source: ARK Investment 
Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Krafcik 2020, as of January 29, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotaxis
    75/148
    Theoretically, Tesla Alone Could Hit ARK’s Robotaxi Industry Forecast In 2030
By 2030, ARK’s research suggests that the number of robotaxis will have scaled to roughly 50 million globally, with Tesla at ~50% market 
share. Competitors could face barriers to scale, given the intense price competition in China and the lack of alignment between 
technology providers and traditional automakers. Separately, if it were to allocate 100% of its future vehicle production to robotaxi 
service—instead of the ~30-40% assumed—Tesla alone could hit ARK’s robotaxi industry forecast in 2030.
 -
 10
 20
 30
 40
 50
 60
2024 2025 2026 2027 2028 2029 2030
Autonomous Vehicle Fleet (Millions)
Global Autonomous Vehicle Adoption Forecast
Tesla Robotaxis Other (Waymo, Baidu Apollo Go, etc.)
76
Note: Autonomous adoption curve overestimates figures in initial years. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be 
provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future 
results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotaxis
    76/148
    Autonomous Electric Vehicles Could Rapidly 
Shift Vehicle Miles Traveled (VMT) To Electric
Autonomous electric vehicles could exceed 
50% utilization, pushing a disproportionate 
number of miles traveled from gas-powered 
vehicles to electric drivetrains.
0%
5%
10%
15%
20%
25%
30%
35%
40%
2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030
Share
Global Autonomous Electric Miles As A Share Of Total Miles
 Electric Vehicle Miles Traveled Without Autonomous Forecasted Electric Vehicle Miles Traveled Without Autonomous
 Autonomous Electric Vehicle Miles Traveled Forecasted Autonomous Electric Vehicle Miles Traveled
Slide by Sam Korus, Director, Daniel Maguire, Analyst, and Akaash TK, Associate, Autonomous Technology & Robotics. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of 
external data sources as of January 27, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold 
any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 77
ARK’s Big Ideas 2025: Robotaxis
    77/148
    Wright’s Law Suggests That A $15,000 Robotaxi Is Possible
 $1
 $10
 $100
 $1,000
 $10,000
 100,000 1,000,000 10,000,000 100,000,000 1,000,000,000 10,000,000,000
$/kWh
Cumulative kWh
Battery Cost Decline
Nickel Cells Nickel Forecast LFP Cells LFP Forecast
Battery Cost ($/kWh)
$100 $75 $50 
Miles/kWh
4 $6,250 $4,688 $3,125 
4.5 $5,556 $4,167 $2,778 
5 $5,000 $3,750 $2,500 
5.5 $4,545 $3,409 $2,273 
Assumes a 250-mile range vehicle
Tesla has suggested that the Cybercab will have an efficiency of 5.5 miles/kWh. Paired with the battery cost decline anticipated by 
Wright’s Law, that efficiency suggests that the battery could cost as little as $2,300. Typically, the drivetrain of a vehicle is ~20% of the 
total vehicle cost, which suggests a Cybercab could be profitable at a price of $15,000 or below.
Slide by Sam Korus, Director of Research, Autonomous Technology & Robotics Research. Note: Wright’s Law states that for every cumulative doubling of units produced, costs will fall by a constant percentage. 
See Winton 2019. “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of 
external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold 
any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 78
ARK’s Big Ideas 2025: Robotaxis
    78/148
    Tesla’s Cost Per Mile Is ~30-40% Lower Than Waymo’s
 $-
 $0.20
 $0.40
 $0.60
 $0.80
 $1.00
 $1.20
Waymo 5th Gen
(Early Launch)
Model 3
(Early Launch)
Waymo 6th Gen
(At Scale)
Cybercab
(At Scale)
Robotaxi Incremental Cost Per Mile
 Depreciation Operations Support Insurance
 Energy Maintenance & Tires Parking
Tax
Its higher cost structure is likely to make Waymo less competitive or less profitable than a Cybercab robotaxi service. Waymo’s 
dependence on LIDAR and higher cost auto manufacturers are two drags on its competitive positioning. 
 -
 50,000
 100,000
 150,000
 200,000
 250,000
 300,000
 350,000
Waymo Hyundai Ioniqs Tesla Cybercabs
Number Of Vehicles
Vehicles Per $5 Billion In Capex
Note: “Capex”/“Capital Expenditure” is a one-off upfront cost required to build or deploy an asset. “Early launch” assumes ten cars per remote operator with utilization rates roughly equivalent to ride-hail today. “At scale” 
assumes 100 cars per remote operator with an improved autonomous operation utilization rate. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 
2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not 
indicative of future results. Forecasts are inherently limited and cannot be relied upon. 79
ARK’s Big Ideas 2025: Robotaxis
    79/148
     $-
 $0.05
 $0.10
 $0.15
 $0.20
 $0.25
Robotaxi Forecast
Cost Per Mile
Fees
Robotaxi Platform Take
 Depreciation
 Parking
 Energy
 Maintenance & Tires
Deadhead Cost
 Insurance
 $-
 $0.20
 $0.40
 $0.60
 $0.80
 $1.00
 $1.20
 $1.40
 $1.60
 $1.80
 $2.00
Human Ride-Hail
Cost Per Mile
Uber Take
Booking Fee (Ex Insurance)
Driver Take
Deadhead Cost
Maintenance & Tires
Depreciation
Fuel
Insurance
Financing
Parking
Autonomous Ride-Hail Should Reduce Costs, Including Deadhead Miles
Autonomous electric travel should reduce costs, including fuel, insurance, and deadhead miles, thanks to higher utilization, increased 
safety and efficiency, and electric drivetrains. Deadhead miles are the number of miles driven without passengers, which account for 
45% of ride-hail miles today. Autonomous vehicles could reduce deadhead miles by removing the human-in-the-loop and harnessing 
AI tools to predict demand and traffic patterns.
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 80
Fees
Robotaxi Platform Take
ARK’s Big Ideas 2025: Robotaxis
    80/148
    Tesla Expects Its Autonomous Software To Outperform 
Human Drivers On Safety
Soon, Tesla expects more than 10,000 miles between critical interventions, or roughly one year’s worth of human driving before the 
average person would need to touch the wheel in Full Self-Driving (FSD) mode. Later in 2025, Tesla plans to outperform the national 
accident rate—or ~700,000 miles between critical interventions—paving the way for conversations with regulators. In our view, Tesla 
could launch a robotaxi service with remote operators before hitting that milestone.
Waymo’s commercial robotaxi service has surpassed 400,000 miles between police-reported collisions—impressive performance for 
a fleet of its size. 
10
1,000
10,000
National Average Human Accident Rate (700,000) 700,000
One Year Of Personal Driving (10,000)
Waymo Police Reported Collisions (400,000)
1
10
100
1,000
10,000
100,000
1,000,000
16,214 66,736 120,000 190,000
(Log Scale)
Compute In H100 Equivalent GPUs
Tesla Miles Between Critical Intervention
Jan 24
11.4.8.1
Oct 24 
12.5.4.1
100x
Mar 25
13.4.5
10x
Oct 25 
13.?
70X
Date 
FSD Version
Improvement
81
Note: “Miles Between Critical Intervention” measures the distance traveled before a human driver must take control to avoid an accident. Source: ARK Investment Management LLC, 2025. This ARK analysis draws 
on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, 
sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotaxis
    81/148
    Ride-Hail Could Create A ~$10 Trillion Market
$0.00
$0.50
$1.00
$1.50
$2.00
$2.50
$3.00
$3.50
$4.00
$4.50
Price Per Mile
Ride-Hail Addressable Market*
$4B 4.00 3.00 2.00 1.10 0.60 0.50 0.25
$30B
$100B
$1T: Addressable ridership in Western markets at ~$1
$2.4T: Non-commuting miles in higher income 
countries priced at ~$0.60 per mile
$1.92T: Long tail of demand priced at human driven ridehail prices of $0.35 per mile in lower income countries $5T: Low cost, accessible autonomous 
travel at $0.25 per mile 
$134B: Existing addressable market for 
ride-hail companies charging $2-$4 per 
mile in Western Markets
1 5 10 0 30 
Price Points ($):
0.35
At $0.25 cents per mile, autonomous transportation could serve a much larger population than human-driven ride-hail does today. 
Meanwhile, based on how consumers value their time, demand could be significant even at higher price points.
Miles (Trillions)
82
*$10 trillion is the addressable market, not the revenue we expect in 2030, as we do not expect autonomy to penetrate all addressable miles. Source: ARK Investment Management LLC, 2025. This ARK analysis 
draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation 
to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotaxis
    82/148
    500 40 500 1,400 200 1,000
2,600 2,200
33,600
$0
$5,000
$10,000
$15,000
$20,000
$25,000
$30,000
$35,000
$40,000
Revenue (Net) EBIT Enterprise Value
Billions
Forecasted Revenue, Earnings, And Enterprise Value In 2030
Autonomous Electric
Auto Manufacturers
Fleet Owners Autonomous Platform Providers
Robotaxis Could Generate ~$34 Trillion In Enterprise Value By 2030
Multiples: 13X EBIT 5X EBIT 15X EBIT
83
Note: Numbers are rounded. “EBIT”: Earnings Before Interest and Tax. This slide depicts net revenue in 2030, the portion of gross revenue that the platform operator receives, while the previous slide depicts the 
total addressable market. Source: ARK Investment Management LLC, 2025This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For 
informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are 
inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotaxis
    83/148
    Cutting Costs, Revolutionizing 
Supply Chains, And Reshaping 
Consumer Behavior
Autonomous 
Logistics
Tasha Keeney, CFA Daniel Maguire, ACA
DIRECTOR OF INVESTMENT ANALYSIS
& INSTITUTIONAL STRATEGIES
RESEARCH ANALYST, 
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
84
ARK’s Big Ideas 2025
    84/148
    $5.40 
$0.35 
Human-Driven
App Delivery
Forecasted
Drone Delivery
Local Small Item Delivery Cost
(Per Trip)
Autonomous Vehicles That Roll And 
Fly Could Lower Supply Chain Costs 
Dramatically
$2.40 
$0.40 
Human-Driven
Delivery
Rolling Integrated
Traffic Robot Delivery
Local Batch Delivery Cost
(Per Trip)
Autonomous vehicles should operate at higher utilization 
rates than human-in-the-loop systems, creating more costeffective last-mile delivery ecosystems.
-83%
-94%
$0.07 
$0.03 
$0.00
$0.01
$0.02
$0.03
$0.04
$0.05
$0.06
$0.07
$0.08
Human-Driven
Diesel Truck
Forecasted Autonomous
Electric Truck
Truckload Delivery Cost
(Per Ton-Mile)
-57%
Human-Driven 
Vehicle
Forecasted 
Robot Delivery
ARK’s Big Ideas 2025: Autonomous Logistics
$6.00
$5.00
$4.00
$3.00
$2.00
$1.00
$0.00
85
Note: We have updated our drone price per mile with our latest assumptions for replacement costs, launching and charging infrastructure, insurance, and labor costs. Fees for drone and robot delivery are shown net of 
infrastructure costs (outside of charging and launch/land), which we believe either could be borne by the drone or robot delivery operators or shared with logistics or retail partners. Source: ARK Investment Management LLC, 
2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    85/148
    Robotrucks And Drones Face More Hurdles 
To Commercialization Than Robotaxis
Robotaxis Robotrucks Drones
Technical
Hurdles
Regulatory
Hurdles
Customer
Partnerships
Manufacturing
Partnerships
Data Collection/
Testing/Validation
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 86
ARK’s Big Ideas 2025: Autonomous Logistics
    86/148
    While Tesla has generated billions of miles 
of customer data for training, Waymo and 
Baidu have access only to millions of miles. 
In trucking, autonomous players have taken 
a route-by-route approach, building their 
data libraries as they add more routes. For 
drones, while many flights have occurred in 
Africa, companies continue to test with 
partners in the US and Europe, with limited 
services offered, thus far.
In the robotaxi industry, Tesla is vertically 
integrated while companies like Waymo and 
Baidu have partnered with Geely and 
Hyundai/BAIC, respectively. In the trucking 
space, players like Aurora have partnered 
both with OEMs* like Paccar and Volvo and 
with integrators like Continental. Meanwhile, 
many drone providers manufacture in-house 
but have not hit mass manufacturing scale. 
Robotaxis sell rides to end consumers, 
while robotrucking and drone 
companies need to forge partnerships 
with retailers, food delivery providers, 
and other logistics operators. 
Robotaxis were first to commercialize, but 
most robotrucks still require a human 
operator, with a number of companies 
pushing to remove the driver this year. 
Meanwhile, drones have been able to fly 
autonomously for a number of years: the 
difficulty lies more in logistics than on the 
autonomous flight technology.
Waymo and Baidu have paved the 
regulatory path for robotaxis in the US 
and China, respectively. In the US, state 
regulators could give way to federal 
legislation in 2025. Trucking is 
influenced by lobbyists, while drones 
are regulated by the FAA, which 
historically has missed many of its own 
timelines for developing a framework 
for drone delivery.
Manufacturing Partnerships: Data/Testing/Validation: 
Technical: Regulation: Customer Partnerships:
Hurdles To Commercialization
* ”OEM”: Original Equipment Manufacturer. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. 
For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 87
ARK’s Big Ideas 2025: Autonomous Logistics
    87/148
    Company Region Of 
Operations
Customer US Regulatory Approval* Cumulative Number Of Commercial Flights
Zipline US 
Africa
Hospital &
Consumer Part 135 (Paid)
Wing
US 
Australia
EU UK
Hospital &
Consumer Part 135 (Paid)
Meituan China
UAE Consumer N/A
Manna US
EU
Hospital &
Consumer Part 107 (Unpaid)
FlyTrex US Consumer Part 107 (Unpaid)^
Matternet
US
EU
UAE
Hospital &
Consumer
Part 107 (Unpaid)^
Amazon US Consumer Part 135 (Paid) 5,000 e
50,000 
100,000 
280,000 e
400,000 
400,000 
1,320,000 
Logistics Drone Companies 
Are Beginning To Break 
Through Regulatory Barriers
• Thus far, most drone deliveries have taken place in rural areas 
outside the US to serve unmet medical needs.
• Companies with more real-world data and regulatory 
approvals to fly beyond-visual-line-of-sight without observers 
will pull away from the competition. 
88
*In the US, paid drone delivery requires Federal Aviation Administration (FAA) Part 135 certification; Part 107 allows operation but prohibits compensation. That said, the FAA is working on updating drone regulation to 
standardize beyond visual line of sight operations for commercial drones in the US under Part 108. ^FlyTrex and Matternet have Part 135 partners. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a 
range of external data sources as of December 17, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any 
particular security. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: Autonomous Logistics
    88/148
    1,520,000 
2,200,000 
3,100,000 
3,500,000 
125,000,000 
Gatik
Aurora
Pony.ai
Kodiak
Cumulative Number Of Miles Traveled
• In the US, commercial autonomous trucking requires a 
safety driver today.
• China leads in generating real-world driving data, but 
companies in the US should advance considerably once 
regulators approve the removal of safety drivers in 2025.
Autonomous Trucking Companies 
With The Most Real-World Miles 
Should Be Best Equipped For Scale



Inceptio  Technology
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 89

ARK’s Big Ideas 2025: Autonomous Logistics
    89/148
    Company Countries Sidewalk 
Or Road?
College Campus 
Or Beyond?
Number Of Deliveries Completed
Starship 
Technologies
US GB DE 
FI EE Sidewalk Mostly Campus
Neolix CN Road Beyond
Meituan CN Road Beyond
Coco* US FI Sidewalk Beyond
Avride KR US UAE Sidewalk Beyond
Cartken JP US Sidewalk/
Indoor Mostly Campus
Serve Robotics US Sidewalk Beyond
Robot Delivery Companies With 
Broad-Based Experience Should Be 
In A Superior Competitive Position
• In the US, most sidewalk deliveries are on college 
campuses, which do not seem to be productive training 
ground for urban and residential areas.
• In China, road robots are commonplace.
50,000 
200,000 e
200,000 
500,000 
4,000,000 
6,000,000 
7,000,000 
90
*Coco has evolved from human-driven to hybrid robots and is more reliant on humans than are its competitors. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as 
of December 17, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past 
performance is not indicative of future results.
ARK’s Big Ideas 2025: Autonomous Logistics
    90/148
    Autonomous Delivery Revenue 
Could Approach ~$900 Billion 
Globally In 2030
280
160
420
Robots & Drones
Last Mile
Trucks
Middle Mile & Over-The-Road
Forecasted Autonomous Delivery Revenue
($ Billions, 2030)
Parcels Food All Goods
Technology-enabled delivery is likely to reshape consumer 
buying habits, as the cost of transporting goods falls. Our 
research suggests that robot/drone delivery fees and 
autonomous trucking revenues could reach ~$440 billion 
and ~$420 billion, respectively, in 2030.
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 17, 2024, which may be provided upon request. For informational purposes only and should not 
be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 91
ARK’s Big Ideas 2025: Autonomous Logistics
    91/148
    The US Needs Low-Cost, 
AI-Powered Aircraft
• The current state of warfare has proven that high-cost aircraft are 
unsustainable against low-cost internationally made drones. 
• Meanwhile, AI is improving decision-making and efficiency, reducing 
the need for human troops.
“Smaller And Cheaper AI-controlled 
Unmanned Jets Are The Way Ahead.”
US F-16
~$30-60 million
Chinese DJI 
~$2,000-20,000
US SM-2 Missile 
~$2 million
Ukraine
Red Sea:
Iranian Shahed 136
~$20,000
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security.
“Replicator will galvanize progress in the 
too-slow shift of U.S. military innovation to 
leverage platforms that are small, smart, 
cheap, and many.”
“We need them potentially at very large 
scale and very quickly.”
“manned fighter jets are obsolete 
in the age of drones...”
“I want to turn the Taiwan Strait into 
an unmanned hellscape…”
“If we're shooting down a $50,000 one-way 
drone with a $3 million missile, that's not a 
good cost equation." 
Bill LaPlante, 
Under Secretary of Defense for 
Acquisition and Sustainment
Frank Kendall, 
26th Secretary of the Air Force
Doug Bush,
Army Acquisition Chief, on buying 
drones in large quantities
Kathleen Hicks,
Deputy Secretary of Defense
Samuel Paparo,
US Indo-Pacific Command Chief 
Navy Admiral
Elon Musk 
92
ARK’s Big Ideas 2025: Autonomous Logistics
    92/148
    Powering The Artificial 
Intelligence Revolution
Energy
Sam Korus Daniel Maguire, ACA
DIRECTOR OF RESEARCH 
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
RESEARCH ANALYST
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
Akaash TK
RESEARCH ASSOCIATE 
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
93
ARK’s Big Ideas 2025
    93/148
     50
 100
 150
 200
 250
 300
 350
 400
ARK Forecast: Incremental Power 
Necessary For AI Hardware
2030
Power Capacity Added In China
2023
GW
Incremental Power Necessary For AI Hardware In 2030 
VS. Power Capacity That China Added In 2023
Energy Production Is Important To Economic Growth
According to our research, China added more power capacity in 2023 than would be necessary to meet our estimate 
of incremental power demand from AI data centers through 2030. In other words, the buildout rate necessary to 
accommodate AI power demand is achievable.
 5,000
 10,000
 15,000
 20,000
 25,000
 30,000
 35,000
 40,000
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
20052007
2009
2011
2013
2015
2017
2019
2021
2023
2025e
2027e
2029e
Billion kWh
Global Electricity Generation
Global Electricity Generation
Forecast Of Global Electricity Generation Extrapolated From 5 Year Trailing CAGR
Forecast Of Global Electricity Generation Including That Necessary For AI Hardware Demand
Note: “CAGR”: Compound Annual Growth Rate. “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. “GW”: Gigawatt, a unit of power equal to one billion watts. Source: ARK 
Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered 
investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
Power Capacity Added
In China 2023
Incremental Power Necessary
For AI Hardware In 2030e
94
ARK’s Big Ideas 2025: Energy
    94/148
    Nuclear Energy Regulation Ended 
The Declines In Electricity Prices
1893
1944
1974
2021
$0.10
$1.00
$10.00
 1 10 100 1,000 10,000 100,000 1,000,000 10,000,000
$/kWh 2023 Dollars
Cumulative Megawatts
US Electricity Prices
Informed by Wright’s Law,* ARK’s research indicates that, apart from 
WWII, US electricity prices fell consistently from the late 1800s until 
1974. The U.S. Energy Reorganization Act, enacted in 1974, reversed 
the drop in nuclear construction costs.
 $1
 $10
 $100
 $1,000
 $10,000
 $100,000
 1 10 100 1,000 10,000 100,000 1,000,000
Overnight Construction Cost (2023 $/kW)
Cumulative Power Deployed (MW)
United States & Japan Nuclear Construction Cost Trends 
(Log Log)
U.S. Before Significant Regulatory Change (1953 - 1974) U.S. Post Significant Regulatory Change (1975-2023)
Japan Before Significant Regulatory Change (1961-1977) Japan Post Significant Regulatory Change (1978-2009)
Japan Nuclear Reactor 
Regulation Revision (1978)
U.S. Energy 
Reorganization 
Act (1974)
Note: “MW”: Megawatt, a unit of power equal to one million watts. “kW”: Kilowatt, a unit of power equal to one thousand watts. “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 
hour. *Wright’s Law states that for every cumulative doubling of units produced, costs will fall by a constant percentage. See Winton 2019. Source: ARK Investment Management LLC, 2025, based on data from Smil 2000 and Cleveland 
2023 (left chart) and Lovering et al. 2016 (right chart). For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not 
indicative of future results. Forecasts are inherently limited and cannot be relied upon. 95
ARK’s Big Ideas 2025: Energy
    95/148
    20
40
60
80
100
120
140
160
2006200720082009201020112012201320142015201620172018201920202021202220232024
Transcripts Mentioning "Nuclear Energy"
"Nuclear Energy" Mentions
The Nuclear Tide Is Rising
Interest in nuclear energy is surging as companies seek 
a clean, reliable power source to sustain AI initiatives 
and meet ambitious net-zero carbon goals.
Microsoft in deal for Three Mile Island 
nuclear power to meet AI demand
- Financial Times, Sept 20, 2024
Google orders small modular nuclear 
reactors for its data centers
- Financial Times, Oct 14, 2024
Amazon buys stake in nuclear energy 
developer in push to power data centers
- Financial Times, Oct 16, 2024
Fukushima 
Nuclear 
Accident
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Capital IQ Pro, as of December 18, 2024, which may be provided upon request. 
For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 
Forecasts are inherently limited and cannot be relied upon. 96
ARK’s Big Ideas 2025: Energy
    96/148
    Supported by clear regulations, nuclear fission could be a cost-competitive, timely solution to increasing power demands globally.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
 $2,000
 $4,000
 $6,000
 $8,000
 $10,000
 $12,000
 $14,000
Utility Solar Onshore 
Wind
Natural Gas 
Combined 
Cycle
Nuclear
Capacity Factor
$/kW
Overnight Capital Cost
2024 Prices (lhs) 1974 Price (2024 $ lhs) Capacity Factor* (rhs)
 1
 2
 3
 4
 5
Small Hydro Biomass Wind OnshoreWind Offshore Solar PV Nuclear (5MW)
Vallecitos (1957)
Years
Average Commissioning Time 
For Small Scale Renewable Energy Projects (1-10 MW)^
Wind
Onshore
Wind
Offshore
97
Nuclear Fission Projects Could Deliver Economics Superior 
To Other Renewable Energy Projects
Note: In 1957, the Vallecitos Boiling Water Reactor (VBWR) in Sunol, California, was the first nuclear power plant privately owned and operated to supply electricity to a public utility grid. “MW”: Megawatt, a unit of power equal to 
one million watts. “kW”: Kilowatt, a unit of power equal to one thousand watts. ^Average across 48 countries between 2005-2022. *Capacity Factor is defined as the ratio of the electrical energy produced by a generating unit for the 
period of time considered to the electrical energy that could have been produced at continuous full power operation during the same period. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of 
external data sources, including World Nuclear Association 2024, Gumber et al. 2023, and Lovering et al. 2016, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Energy
    97/148
    While Most Small Modular 
Reactor (SMR) Startups Will Fail, 
The Survivors Should Thrive
SMRs are early stage and face significant execution risks. To triple nuclear 
capacity by 2050, the US must build large reactors, SMRs, and microreactors. 
Shorter construction timelines, modular design, and scalability could enable 
smaller reactors to meet near-term demand. Selecting from more than 30 
US designs will be important to reigniting Wright’s Law and lowering 
electricity costs.
0
1
10
100
1,000
10,000
Net Power In Megawatts Electric (MWe)
Size Of US Advanced Reactors
(Log Scale)
NRC Approved Designs Designs Under Development
SMR
Micro
Gen III Gen III+ Gen IV
Large
Advanced
Note: “NRC”: U.S. Nuclear Regulatory Commission. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Small Modular Reactor (SMR) Global Tracker as of 
November 14, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance 
is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
98
Global SMR Projects
Early Stage
Under Construction
Operable
Number
0
10
20
30
40
50
60
70
80
90
100
Global SMR Projects
Global SMR Projects Under Construction
ARK’s Big Ideas 2025: Energy
    98/148
    Nuclear Fusion Faces Significant 
Hurdles To Commercialization
JET 1997
Forecasted SPARC 2027
Forecasted ITER 2039
1E+12
1E+13
1E+14
1E+15
1E+16
1E+17
1E+18
1E+19
1E+20
1E+21
1E+22
1950 1970 1990 2010 2030
Triple Product (Log Scale)
[Density x Time x Temperature]
Fusion Project Progress Toward Scientific Breakeven*
Scientific Breakeven Q=1 (D-T Fuel)
Nuclear Fission Milestones
Demonstration Of SelfSustaining Nuclear Fission 
1942
First Nuclear Power Plant To 
Produce Usable Electricity
1951
15 Years
Time From Key Milestone To Commercialization
First Commercial 
Nuclear Power 
Plant In The US
1946 1957
Creation Of The U.S. Atomic 
Energy Commission
Despite ~75 years of research, nuclear fusion has not generated net 
energy in a meaningful way. While private fusion companies are 
promising breakthroughs as early as 2027, commercialization could 
take another ~15 years, if fission history is any guide. Commercial
Breakeven
The fusion reaction
generates surplus
power to cover the
cost of building and
operating the plant.
Engineering
Breakeven
The fusion reaction
generates the same
power as the reactor
consumes.
Scientific
Breakeven
The fusion reaction
generates the same
power as injected
into the fuel.
99
Note: The ‘Triple Product” displayed on the Y axis above is a single metric that measures nuclear fusion progress by considering density, confinement time, and temperature. It's a very large number, so we use scientific notation, 
e.g., "1E+21," meaning 1 followed by 21 zeros. “SPARC”: Smallest Possible Affordable Robust Compact reactor. SPARC is a Tokamak design. “ITER”: International Thermonuclear Experimental Reactor. *Excluding Laser-Based Inertial 
Confinement. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Energy
    99/148
    Solar, wind, and stationary energy storage are just getting started. One solution will not meet all demands. Interestingly, Tesla accounts 
for ~19% of global energy storage today.
Nuclear And Other Renewables Can Serve As Viable Solutions
20
40
60
80
100
120
140
160
180
2018 2019 2020 2021 2022 2023 2024
Gigawatt-Hours (GWh)
Global Stationary Energy Storage
Tesla Energy Storage Non-Tesla
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results.
4%
8%
12%
16%
1,000
2,000
3,000
4,000
5,000
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024e
Share (%)
Terawatt-Hours (TWh)
Global Renewable Energy Generation
Wind (lhs) Solar (lhs) Percent of Total Generation (rhs)
100
Percent Of Total Generation (rhs)
ARK’s Big Ideas 2025: Energy
    100/148
     $0.05
 $0.10
 $0.15
 $0.20
 $0.25
 $0.30
30% 50% 70% 90%
$/kWh
Uptime Capacity
Lowest Levelized Cost Of Energy 
For Desired Uptime
Current Cost For Solar And Battery
ARK's Estimated Future Cost For Solar And Battery
As Battery Costs Decline, Intermittent Energy Sources Become Less Problematic
Declining battery costs should make intermittent energy systems economically attractive with 100% uptime. ARK’s modeling suggests 
that energy-intensive industries will accelerate the pairing of renewables with energy storage. Bitcoin mining is a prime example, 
showcasing how energy-intensive industries with business models orthogonal to power generation can accelerate the shift to 100% 
renewable grid integration.
Note: “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. “MWh”: Megawatt-hour, a unit of energy, equal to 1,000 kilowatt-hours, representing the use or generation of 1 
megawatt of power for 1 hour. Source: ARK Investment Management LLC, 2025, based on data from National Renewable Energy Laboratory as of December 31, 2024. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 101 Solar Installation Power (Megawatts)
Battery System (MWh)
Solar, Battery, Bitcoin Systems Required To Meet Grid 
Dispatch Targets At Constant Cost Of Electricity
(Size Of Bubble = Bitcoin Mining Power)
ARK’s Big Ideas 2025: Energy
    101/148
    0
20
40
60
80
100
120
140
160
180
Gigawatt-Hours (GWh)
Status Of US Energy Generation Capacity In Interconnection Queue 
(2009 VS. 2023)
Solar Wind Gas
Between 2009 and 2023, US interconnection approval wait times increased ~80%, from 31 months to 56 months. Quicker approvals 
could lower project costs and accelerate the scaling of energy-intensive projects.
Regulatory Reform Should Boost All Renewables
 20
 40
 60
 80
2009
2011
2013
20152017
2019
2021
2023
Median Duration (Months)
US Interconnection Wait Times 
For New Energy Generation
Solar Wind Gas Combined
In Queue/
Active
Complete/ Withdrawn
Operational
In Queue/
Active
Withdrawn Complete/
Operational
2009 2023
102
*An interconnection queue is a system used by grid operators to manage and prioritize requests from energy developers to connect new power generation projects to the electricity grid, involving assessments of each 
project's impact on grid reliability, planning for necessary infrastructure upgrades, and determining cost responsibilities for integrating new plants. Source: ARK Investment Management LLC, 2025, based on data from 
Lawrence Berkeley National Laboratory as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past 
performance is not indicative of future results.
ARK’s Big Ideas 2025: Energy
    102/148
    103
Decoupling Physical 
Labor From Output
Robotics
Sam Korus
DIRECTOR OF RESEARCH
AUTONOMOUS TECHNOLOGY
& ROBOTICS
Daniel Maguire, ACA
RESEARCH ANALYST 
AUTONOMOUS TECHNOLOGY
& ROBOTICS
Akaash TK
RESEARCH ASSOCIATE 
AUTONOMOUS TECHNOLOGY
& ROBOTICS
ARK’s Big Ideas 2025
    103/148
    Robots Should Boost Productivity And Transform Industries
104
2
4
6
8
10
12
14
16
Before 
Washing Machines
After 
Washing Machines
Hours
Time To Do Laundry
0
2
4
6
8
10
12
14
Before
Assembly Line
After
Assembly Line
Hours
Time To Manufacture A Car
10
20
30
40
50
60
70
80
Before
Kiva Robots
After
Kiva Robots
Minutes
Time From Click To Ship 
At An Amazon Warehouse
-87% -88% -78%
Automation’s impact on productivity has transformed industries, with specialized robots as simple as washing machines reducing 
dramatically the time required for tasks.
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: Robotics
    104/148
    Generalizable Robotics Represent A $26+ Trillion Global Revenue Opportunity
105
~2.3 Hours Of Unpaid Work Per Day
~2.8 Billion Working Age Population
~$12 Weighted Average Hourly Wage
~$13+ Trillion Opportunity
½ Value Attributed To Free Time VS. Paid Time
Household Robotics Manufacturing Robotics
~$13+ Trillion Opportunity
(Average Of The Green Cells)
ARK Forecasts Global Manufacturing GDP 
At ~$32 Trillion In 2030
×
=
Revenue Opportunity*
($ Billions)
=
×
×
Productivity Uplift
10% 25% 50% 100% 200% 400%
Take Rate
10% 320 800 1,600 3,200 6,400 12,800 
20% 640 1,600 3,200 6,400 12,800 25,600 
50% 1,600 4,000 8,000 16,000 32,000 64,000 
Note: The $12 hourly wage is an estimate of average global wage, not US wage + benefits. *The cells highlighted in green represent what ARK believes to be a reasonable or likely outcome. ^We define “Take Rate” as the 
percentage of a transactions value that the business keeps. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon 
request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are 
inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotics
    105/148
    Humanoid Robots Are Debuting Around The World
Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 106
Why the human form factor? Key is that a humanoid robot is generalizable. While a wrench can tighten nuts better than a human hand 
can, it is not a generalizable tool. The human hand is generalizable, particularly in an environment built by and designed for humans.
Tesla
Optimus
Sanctuary AI
Phoenix
Figure AI
Figure 02
Unitree
G1
Boston Dynamics
Atlas Agility
Digit
Fourier
GR-1
1x Technologies
NEO
Apptronik
Apollo
ARK’s Big Ideas 2025: Robotics
    106/148
     $-
 $250,000
 $500,000
 $750,000
 $1,000,000
 $1,250,000
0% 50% 100% 150% 200%
Upfront Robot Cost
Productivity Uplift
Net Present Value (NPV) Of A Robot Operating The Same Number 
Of Hours As A Human Worker Over Ten Years
Robot Price For NPV=0 Tesla Optimus Price Expectation At Scale
107
The Adoption Of Humanoid Robots Should Increase As Their Costs Fall 
And Their Productivity Increases 
A 100% productivity uplift is equivalent to replacing a human working one shift over ten years.
Human Labor Costs On 
Average In The US
Hourly
Annually $92,420
$46, including benefits
$924,200
~20,000
10-Year Cumulative
Total Hours 
~$550,000,
not including turnover
Net Present 
Value Of Costs
Not Worth Investing In 
A Humanoid Robot
Worth Investing In 
A Humanoid Robot
Note: Per hour salary based on Bureau of Labor Statistics Employer Costs For Employee Compensation press release on September 10, 2024: Average employer costs for all civilian workers = $46.21 per hour; Wages average = 
$31.80; average benefit costs = $14.41. We assume a 40-hour work week and a 50-week work year. A positive net present value in this calculation suggests that it would be worthwhile to invest in a humanoid robot at that 
upfront cost and productivity uplift. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational 
purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot 
be relied upon.
ARK’s Big Ideas 2025: Robotics
    107/148
    Jobs Are Bundles Of Tasks
108
What does a professional dishwasher do?
$8 $14
$246
$50
$100
$150
$200
$250
$300
Dishwashing Machines Professional
Dishwasher Pay
Value of Unpaid Time
Spent Washing Dishes
At Home
$ Billions
Annual US Dishwashing Costs
In a restaurant with a dishwashing machine, washing dishes is one of thirteen tasks required of a “dishwasher.” The home is not much 
different. The dishwasher is one of the most common “robots” in a house. According to our research, automating activities at home like 
loading and unloading dishes and cleaning countertops represents a $250 billion revenue opportunity in the US alone.
Professional Pay 
For Dishwasher
Value Of Unpaid 
Time Spent Washing 
Dishes At Home
• Wash dishes, glassware, flatware, 
pots, and pans, using dishwashers or 
by hand.
• Place clean dishes, utensils, and
cooking equipment in storage areas.
• Sort and remove trash, placing it in 
designated pickup areas.
• Sweep and scrub floors.
• Maintain kitchen work areas, 
equipment, and utensils in clean and 
orderly condition.
• Clean garbage cans with water or 
steam.
• Receive and store supplies.
• Stock supplies like food and utensils 
in serving stations, cupboards, 
refrigerators, and salad bars.
• Transfer supplies and equipment 
between storage and work areas, by 
hand or using hand trucks.
• Clean and prepare various foods for 
cooking or serving.
• Prepare and package individual 
place settings.
• Load and unload trucks that deliver 
or pick up food and supplies.
• Set up banquet tables.
Tasks:
Note: For the value of unpaid time spent washing dishes, we assume, based on a study, that people value their free time at half their wage. From the Bureau of Labor Statistics, we take a median wage of ~$35. The Bureau of 
Labor Statistics reports that people spend 0.65 hours per day on food prep and cleanup. We assume .22 hours is spent on cleaning. We then multiply by the working age population. Source: ARK Investment Management 
LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotics
    108/148
    Small Businesses Should Benefit Disproportionately From Humanoid Robots
Large manufacturing firms are organized by specialized and automated 
tasks. Specialization and automation give large firms the wherewithal to scale 
significantly and, in turn, lower labor costs as a share of revenue. Consequently, 
and somewhat counter-intuitively, large companies typically pay higher wages 
than small firms because automating specific tasks boosts productivity in large 
firms more than in small firms. 
Because generalized automation solutions—those for multiple tasks—have not 
evolved as quickly as automation solutions for specific tasks, small firms 
typically have a disproportionate number of automatable-but-not-yetautomated tasks that would benefit from generalizable solutions like 
humanoid robots. Moreover, a high percentage of manufacturing employees in 
the US work in small firms.
1
2
3
4
5
6
 <500
 500-749
 750-999
 1,000-1,499
 1,500-1,999
 2,000-2,499
 2,500-4,999
 5,000-9,999
 10,000-19,999
 20,000+
Millions
Firm Size By Number Of Employees
US Manufacturing Employment
Sample Small Firm Sample Large Firm
Revenue $1,000 $5,000
Labor Share Of 
Revenue 40% 20%
Number Of Employees 2 4
Wage Per Employee $200
($1,000 x 40%/2)
$250
($5,000 x 20%/4)
Tasks Per Employee 4 2
Benefit Of Automating 
A Single Task
$50 
($200 x (1/4 tasks))
$125 
($250 x (1/2 tasks))
Valued Automation 
Solution Generalizable Single Task
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including U.S. Census Bureau 2023, as of January 10, 2025, which may be provided upon request. For 
informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are 
inherently limited and cannot be relied upon. 109
ARK’s Big Ideas 2025: Robotics
    109/148
    Automation Turns Non-Market Activity Into Revenue-Generating Activity 
And Gross Domestic Product (GDP)
110
Family 
Farmworkers
Hired 
Farmworkers
1
2
3
4
5
6
7
8
Total Breakdown
Number (Millions)
Farm Worker Jobs “Lost”
1950-2000
Of the 82% who “lost" their jobs between 1950 and 2000, most farm workers were unpaid family members.
2
4
6
8
10
12
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Number (Millions)
Farm Workers
Paid Workers Unpaid Family Workers
Source: ARK Investment Management LLC, 2025, based on data from U.S. Department of Agriculture 2025 as of January 10, 2025. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results.
ARK’s Big Ideas 2025: Robotics
    110/148
    Consumers Continue To “Contribute” Unpaid Labor To The US Economy
111
If US consumers were to value their time at half their hourly wages, food preparation and cleanup would add ~$1 trillion to GDP annually. 
Food Cost
Prep And 
Cleanup 
Cost
 $2
 $4
 $6
 $8
 $10
 $12
 $14
Home Cooked Fast Casual
Cost Of A Meal
Food Away 
From 
Home Food At 
Home
Food Prep 
And 
Cleanup
500
1,000
1,500
2,000
2,500
2023 Addressable 
Market
$ Billions
Total US Food Cost
~$1 Trillion
200
400
600
800
1,000
1,200
1,400
1,600
1997
2000
2003
2006
2009
20122015
2018
2021
$ Billions
US Food Expenditure
Food At Home
Food Away From Home
Food Away From Home Less Delivery
Home-Cooked
Note: The Food Prep and Cleanup Cost is based on the following: The Bureau of Labor Statistics (BLS) reports that time spent on food prep and cleanup is .65 hours per day. Based on a study, we assume people value their free 
time at half their hourly wage; for this, we use BLS data of ~$35. We use the population above 18 years old of ~260 million. Source: ARK Investment Management LLC, 2025, based on data from U.S. Department of Agriculture 2025 
and Franchisee Resource Center 2021 as of January 10, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is 
not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Robotics
    111/148
    3D Printing Revenues Could 
Grow ~40% At An Annual Rate 
To $180 Billion By 2030
20
40
60
80
100
120
140
160
180
200
2024 2025 2026 2027 2028 2029 2030
$ Billions
3D Printing Revenue Forecast
• The 3D printing industry consolidated 
significantly in 2024, highlighted by Nano 
Dimension’s acquisition of Desktop Metal and 
Markforged.
• If manufacturers decide to bring the 
technology in house, as did General Electric, 
consolidation could continue. 
• The Trump administration’s efforts to bolster 
US manufacturing could benefit 3D Printing.
• Industries like drone manufacturing and 
nuclear power, which should scale rapidly, 
could use 3D printing to accelerate time to 
market. 
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
Slide by Tasha Keeney, Director of Investment Analysis & Institutional Strategies
112
ARK’s Big Ideas 2025: Robotics
    112/148
    3D Printing Makes Parts For Nuclear Applications More Efficiently
• Westinghouse uses 3D printing for nuclear debris filters.
• Those components improved on the measure of 
resistance to debris by 30%, thanks to the design 
flexibility offered by 3D printing. 
• The US Navy engaged Australian additive manufacturing 
company AML3D to 3D-print tailpiece components for its 
nuclear submarine program.
• AML3D produced those parts in fewer than five weeks, ~15X 
more quickly than traditional manufacturers.
Source: ARK Investment Management LLC, 2025, based on data from Shaikhnag 2024 and Tyrer-Jones 2025 as of January 10, 2025. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
Slide by Tasha Keeney, Director of Investment Analysis & Institutional Strategies
113
ARK’s Big Ideas 2025: Robotics
    113/148
    Reusable 
Rockets
Sam Korus
DIRECTOR OF RESEARCH 
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
Akaash TK
Daniel Maguire, ACA
RESEARCH ANALYST
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
RESEARCH ASSOCIATE 
AUTONOMOUS TECHNOLOGY 
& ROBOTICS
114
ARK’s Big Ideas 2025
    114/148
    SpaceX Has A Ten-Year Head Start On Reusable Orbital Rockets
“Past experience with launch vehicle reusability 
has been mixed at best in terms of achieving 
sustainable cost reductions. So I am a skeptic 
with regard to many of the claims that have 
been made for cost reductions.”
–David W. Thompson, 
CEO, Orbital ATK, 2016
“We don’t have any plans to look at reusability 
at all right now…The reality is that, for the next 
ten years, other than for Falcon, I don’t think 
any of us will look at this.”
-Phil Slack, 
President of International Launch Services, 2014
“We have not really changed our assessment 
over the last couple of years because we have 
yet to see the other forms of reusability—flyback 
or propulsive return to Earth—demonstrate 
economic sustainability on a recurring basis.” 
–Tory Bruno, 
CEO, United Launch Alliance, 2020
2015
SpaceX Lands 
First Falcon 9
2017
SpaceX Launches
1st Flight Proven 
Booster
2024
No Orbital 
Reusable Rocket
Besides SpaceX
2025+
Blue Origin
Rocket Lab
China
Stoke Space
Firefly
Relativity
20
40
60
80
100
120
140
160
180
200
2010 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025e
Number
Falcon Launch Rate
New Boosters Reused Boosters
Forecast
115
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Klotz 2020, de Selding 2016, and Faust 2014, as of January 10, 2025, which may be provided 
upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future 
results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Reusable Rockets
    115/148
    SpaceX Is Refurbishing Rockets In Record Time
ARK’s Big Ideas 2025: Reusable Rockets
When the Space Shuttle cost ~$1.5 billion per launch, industry experts assumed that an economic reusable rocket would be impossible. 
SpaceX then flipped the script. According to ARK’s research, the first stage of the Falcon 9 cost <$1 million to refurbish. Now, rocket 
turnaround time should be proportional to the cost required to refurbish a rocket booster, the key metric in tracking launch cost declines. 
 50
 100
 150
 200
 250
2016 2018 2020 2022 2024
Days Between Reuse
Falcon 9 Average Reuse Days
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results.
252
54
356
27 21 25 14 50
100
150
200
250
300
350
400
Space Shuttle:
Average 1972-
2011
Space Shuttle:
Fastest 1985
First SpaceX
Turnaround
Time
SpaceX 2021
Fastest
Turnaround
Time
SpaceX 2022
Fastest
Turnaround
Time
SpaceX 2023
Fastest
Turnaround
Time
SpaceX 2024
Fastest
Turnaround
Time
Days Between Reuse
Rocket Turnaround Time
116
    116/148
    Rocket Reuse And Size Are Driving Rocket Launch Costs Down On A $/kg Basis
ARK’s Big Ideas 2025: Reusable Rockets
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
 $1
 $10
 $100
 $1,000
 $10,000
 $100,000
 0 1 10 100 1,000 10,000 100,000 1,000,000
$/kg To Low Earth Orbit (2023 $)
Cumulative Upmass (Metric Tons)
SpaceX Launch Cost Decline
Historical 2023 ARK Calculation Forecast
0.1
117
    117/148
    SpaceX Continues To Drive Satellite Bandwidth 
Costs Down In Line With Wright’s Law
Anik-F2 (2004)
WildBlue-1 (2006)
ViaSat-1 (2011)
ViaSat-2 (2017)
ViaSat-3 (2020)
Starlink (2021)
(estimated) 
Starlink (2024)
(estimated)
Starlink Forecast
Starship 2025/2026
Starlink Forecast
Rapidly Reusable Starship
 1
 10
 100
 1,000
 10,000
 100,000
 1,000,000
 10,000,000
 100,000,000
 1,000,000,000
 1 10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000
$/Gbps In Orbit 
(2024 $)
Cumulative Gbps
Satellite Bandwidth Cost Decline
According to Wright’s Law, satellite bandwidth costs should decline roughly ~45% for every cumulative doubling in gigabits per second 
(Gbps) in orbit. Since 2004, the cost of satellite bandwidth has dropped 15,000-fold from $300,000,000 to ~$20,000/Gbps. According to 
ARK’s research, 1 Gbps can serve 200 customers. At a capital cost of ~$1,000/Gbps, SpaceX could recoup its Starship and satellite 
investment with a one-time charge of $5 per customer.
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 118
ARK’s Big Ideas 2025: Reusable Rockets
    118/148
    Lower Satellite Launch Costs Should 
Enable Continuous Global Coverage 
With Low Latency
ARK’s Big Ideas 2025: Reusable Rockets
While latency precluded geostationary orbit (GEO) satellites 
from offering a compelling broadband internet solution, now 
thousands of low-cost, low earth orbit (LEO) satellites can 
provide service with low latency, continuous global coverage, 
and direct-to-mobile device connectivity.
GEO
~22,000 miles
700 ms latency
LEO
~300 miles
<40 ms 
latency
Debris will fall 
back to earth 
within ~5 years 
Debris will fall back 
to earth within 
1,000+ years 
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 119
    119/148
    Starship Will Help The Starlink Constellation Achieve Its Potential
ARK’s Big Ideas 2025: Reusable Rockets
500
1,000
1,500
2,000
2,500
2023 2024e
Metric Tons
Spacecraft Upmass
SpaceX Other Providers
Starship’s payload capacity to LEO is ~5x that of the Falcon 9. While impressive, given the five-year life of its satellites, Starship still will 
have to fly every 2.3 days to maintain its target constellation of 42,000 Starlink satellites. As of January 2025, SpaceX has a constellation 
of over 7,000 satellites.
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
3.8 
(2023 
Cadence)
2.7 
(2024 
Cadence)
2.1 
(2025 Goal)
1.0 
(Modeled)
0.5 
(Modeled)
Metric Tons
Falcon 9 Annual Upmass 
Capability By Launch Cadence
Upmass necessary to maintain
a 42,000 satellite constellation
2.3 (Modeled)
Starship Annual Upmass 
Capability By Launch Cadence
Days Between Launches
120
    120/148
    Despite A Proliferation In Company 
Creation, Small Launch Providers May 
Not Be The Winners
ARK’s Big Ideas 2025: Reusable Rockets
After capital spending booms, industries tend to consolidate. 
In the space industry, while launch capability is critical, the 
larger opportunity could be in the services enabled by low 
launch costs. Today, only 17 of the 194 small launch providers 
created since 1996 are operational.
Source: ARK Investment Management LLC, 2025, based on data from NewSpace Index as of January 10, 2025. For informational purposes only and should not be considered investment advice or a recommendation 
to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
5
10
15
20
25
30
19961997
1998
1999
2001
2002 2003
2004 20052006 2008 200920102011
20122013
20142015
20162017
2018
2019
20202021
20222023
2024
Number Of New Small Launch Providers Founded
10
27
10
47
56
27
17
50
100
150
200
Stage Of Small Launch Providers Created Since 1996
Operational Development <5 years
Development >5 years Dormant
Concept Cancelled
Retired
121
    121/148
    SpaceX Has Converted The Addressable Market Into Revenue 
With Increased Starlink Capacity
ARK’s Big Ideas 2025: Reusable Rockets
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including SpaceX 2024, as of January 10, 2025, which may be provided upon request. For informational purposes only 
and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
~ 20 24 
96 
1,000 
200
400
600
800
1,000
1,200
V1 V1.5 V2 Mini V3
Gbps
SpaceX Bandwidth Per Satellite
100
200
300
400
1
2
3
4
5
2021 2022 2023 2024 2025
Terabits Per Second (Tbps)
Millions
Starlink Subscribers And Bandwidth Capacity
Starlink Subscribers (lhs) Cumulative Bandwidth
122
    122/148
    Thanks to declining rocket launch costs, telecom operators could provide 100% of cellular subscribers satellite connectivity 
anywhere in the world.
Satellite Connectivity Revenues Could Exceed $130 Billion Per Year
0.001%
0.010%
0.100%
1.000%
10.000%
100.000%
1997 2007 2017 2027
Log Scale
Growth Of Satellite Subscribers 
Relative To Cellular Subscriber
RVs 11 Million $1,620 ~$18 Billion
25 Thousand $225,000 ~$6 Billion Commercial 
Aircraft Fleet
100 Thousand $60,000 ~$6 Billion
Cruise Ships,
Warships,
Commercial Ships
Recreational Boats 8.5 Million $1,620 ~$14 Billion
Global Households
Without Access To 
Broadband
600 Million $60 ~$40 Billion
Addressable Subscribers* Annual Revenue*
Direct-To-Device 8 Billion $6 ~$48 Billion
Annual Addressable Market*
Total: ~$132 Billion
*Forecasts. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
SpaceX Direct To Device 
With T-Mobile, At Scale
123
ARK’s Big Ideas 2025: Reusable Rockets
    123/148
    By 2030, Hypersonic Flight Could Be A ~$35 Billion Market, 
Ready To Scale To ~$350 Billion Longer Term
ARK’s Big Ideas 2025: Reusable Rockets
According to the US Department of Transportation, leisure travelers are willing to 
spend 60%-90% of their estimated hourly household income to save one hour.*
Compared to conventional flights that can take 28 hours roundtrip, ARK estimates 
that hypersonic flights could take just 6 hours, saving each traveler ~22 hours.
Given the typical cost and potential time savings, ARK’s research suggests that a firstclass passenger should be willing to spend $44,000 roundtrip for a hypersonic flight.
If launch costs decline in line with ARK’s expectations, early adopters of hypersonic 
flight could generate $35 billion revenue by 2030.
110kg $200/kg 
to LEO
$44,000
Starship Price
2
Roundtrip
Total Number Of Airline Passengers Worldwide: 6.7 Billion
5% Of Flights Are Long-Haul
5% Of Passengers Are First-Class
Number Of Passengers On Long-Haul Flights: ~335 Million
Number Of Passengers Flying First-Class: ~16 Million
50% Adoption At Maturity
Number Of Passengers Flying Hypersonic: ~8 Million
$44,000 Roundtrip Ticket
Annual Addressable Market: ~$350 Billion
Building Blocks Of Addressable Market Forecast
*Forecasts. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 124
    124/148
    Multiomics
Operationalizing Data With AI 
To Transform Diagnostics, Drug 
Discovery, And Therapies
Nemo Despot Marjanovic, PhD
Rong Guo, PhD
Brett Winton
ANALYST, MULTIOMICS
RESEARCH ASSOCIATE, 
MULTIOMICS
CHIEF FUTURIST
125
ARK’s Big Ideas 2025
    125/148
    Multiomics Performance Should Improve By Orders Of Magnitude By 2030*
10x increase in cancer monitoring market
20x increase in screening productivity
1.6x faster commercialization
4x decrease in cost to develop
5x higher return on R&D
1,000x more performant 
at same cost
20x more valuable than standard of care drugs
2.4x more valuable than best-in-class precision 
drugs
Generate Data
Run
Tests
Infer And Inform 
Decisions
Run Experiments
Infers And Informs
Development
Address 
Small Disease 
Populations 
Economically
Identify Patients And Diseases More Precisely • Intellia Therapeutics
• CRISPR Therapeutics
• Beam Therapeutics
• Prime Medicine
• Natera
• Guardant Health
• Recursion 
Pharmaceuticals
• Absci
• Isomorphic Labs
Biological Insights
Generates 
Molecular Diagnostics
• Adaptive 
Biotechnologies
• Veracyte
AI Developed Drugs
Cures
AI
Multiomic Tools
100x decrease in cost to read DNA et al.
1,000x decrease in cost to write DNA et al.
• Twist Biosciences
• Tempus AI
• 10X Genomics
• Illumina
126
*The multiomics performance statistics provided on this slide represent ARK’s research-based forecasts for 2030, which may not be realized. The companies listed represent companies that are currently working toward achieving 
the forecasted results, but the list does not include all companies that may be pursuing the same goals, and which may do so more successfully. The companies listed may or may not be held in ARK portfolios. The information 
provided should not be used as the basis for any investment decision, and it should not be assumed that an investment in any of the companies listed was or will be profitable. Source: ARK Investment Management LLC, 2025. This 
ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to 
buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    126/148
    Biological Information Is 
Encoded In Three Layers
Sequence Layer
The static genetic blueprint 
within DNA, RNA, and proteins.
STATIC
Illumina
Oxford Nanopore
PacBio
QuantumSi
Bruker
DeepMind
Isomorphic Labs
ThermoFisher
10X Genomics
Akoya
ScaleBio
Vizgen
Biological Information Measuring Tools* 
DYNAMIC
The linear arrangement of nucleotides 
in DNA or amino acids in proteins
Examples:
• DNA sequences store genetic 
information, determining an 
organism's primary traits.
• Protein sequences define structure 
and function in cellular processes.
Transcription
Translation
DNA
mRNA
Protein
3'
Tyr Ser Gly Ser
5'
5' 3'
3'
Codons
5'
Amino acids
G A C T A G T C T G
Sequence
The three-dimensional organization of biomolecules like proteins and nucleic acids
Examples
• Protein structures determine binding 
sites and catalytic activity.
• Nucleic acid structures affect stability, 
replication, and transcription.
Agglomeration of molecules
Examples
• Gene regulation: mRNA and 
protein abundance vary, 
controlling gene expression.
Structure
Systems
t-SNE2
-20
-10
0
10
20
-30
-30 -20 -10 0 10 20 30
t-SNE1
30
Label
Label
Label
Label
Label
Label
Label
Label
Single Cell 
Genomics
AlphaFold
AI
DNA/Protein 
Sequencing
Advances in sequencing, AI, and single-cell technologies enable us to read, 
model, and predict complexity. Biological information is organized by three 
interconnected layers that drive life processes.
Systems Layer
The dynamic abundance of 
biomolecules responding to 
cellular and environmental 
cues. 
Structure Layer
The 3D folding of biomolecules, 
defining their function.
127
*The companies listed are examples of companies that employ the measuring tools described, but the list does not include all companies that may be utilizing those tools, and which may do so more successfully. The 
information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies listed was or will be profitable. Source: ARK Investment Management 
LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a 
recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    127/148
    $1
$1,000
$1,000,000
$1,000 M
$1,000 B
1990 1995 2000 2005 2010 2015 2020 2025 2030
Cost Per 1B Base Pairs (Log Scale)
DNA Sequencing Cost Per Whole Human Genome
Actual Projected Moore's Law Projection
$0
$1
$10
$100
$1,000
$10,000
$100,000
$1 M
$10 M
$100 M
$1 B
$10 B
1990 2000 2010 2020 2030
DNA Synthesis Cost Per 1B Base Pairs
Actual Projected Moore's Law Projection
The Costs Of DNA Sequencing And Synthesis Are Falling Precipitously
Illumina’s introduction 
of high-density flow cell
DNA sequencing and synthesis costs are declining faster than Moore's Law* predicted. The cost of reading biology through multiomics
tools like DNA sequencing has plummeted by 10 billion-fold (10¹⁰) in its 35-year history. DNA synthesis costs have dropped 100,000-fold
(10⁵) and could continue to fall 10 million-fold (10⁷) by 2030. Twist Bioscience’s silicon-based DNA synthesis, combined with
miniaturization and parallelization, has reduced costs by increasing efficiency, reducing reagent usage, and scaling production.
Pressure From 
Competitors
Illumina's 
acquisition 
of Solexa 
Twist chip 
miniaturization 
for DNA storage
$1 M
$1 K
$100 K
$10 K
$1 K
128
*Moore’s Law predicts that the number of transistors on a chip doubles every 18 months to two years. See Winton 2019. Note: The total price of sequencing the first human genome was estimated at ~2.7B, which included all the 
costs of developing and improving sequencing tools, actual sequencing, and all analysis. In the above graph, we are calculating only actual sequencing cost per base pair. Source: ARK Investment Management LLC, 2025. This ARK 
analysis draws on a range of external data sources, including BusinessWire 2024 and Carlson 2022 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered 
investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    128/148
    The Productivity Of Genomics Analysis Has Skyrocketed
Since 2001, the compute time required to analyze a human genome has plummeted from 180 days to 10 minutes. Today we can 
analyze ~1.4 billion genomes for the cost of analyzing a single genome in 2001. Alongside the plummeting costs of sequencing and
synthesis, as noted on the previous slide, genome analysis is commoditizing, driving a revolution in biological research and 
applications, from personalized medicine to global-scale genomics.
~180 Days
~70 Days
~15 Days
~15 Hours
~60 Min
~34 Min
<10 Min
1
10
100
1,000
10,000
100,000
1,000,000
2001 2006 2010 2012 2015 2020 2024
Minutes
Compute Time For Analysis Of One Human Genome
1
~1.4 Billion
1
10
100
1,000
10,000
100,000
1,000,000
10,000,000
100,000,000
1,000,000,000
10,000,000,000
2001 2024
Number
Genomes Analyzed For $450,000 Budget
129
Note: 2001 marks the publication of the first draft of the human genome, covering approximately 90% of the genome. In 2003, a near-complete sequence (99.99%) was published, while the fully complete genome was finalized 
in 2022. Although sequencing the first human genome took about 13 years in total, this slide focuses solely on the time required for the computational analysis, excluding the experimental phase, which was a significantly larger 
bottleneck. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Saunders et al. 2021, Miller et al. 2015, Gorzynski et al. 2022, and Clifford 2024 as of December 31, 
2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not 
indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    129/148
    15
408
2,880
112,000
800,000
1
10
100
1,000
10,000
100,000
1,000,000
Used to Train LLaMa 405b Available in UK biobank
Largest publicly available
whole genome datasource
Generated Annually if we
sequenced every newborn in
the US
Generated annually if we
sequenced every newborn
globally
Devote 1% of global
healthcare budget to
sequence a billion people*
Trillion Tokens (Log)
Biological Data Is Big Data
LLaMa 405b LLaMa 405b UK Biobank US Newborns Sequenced
Per Year
Forecast
Global Newborns Sequenced 
Per Year
Forecast
A Billion People
Sequenced
Forecast
Multiomics tools create volumes of data. The largest publicly-available genomics database, the UK Biobank, includes a half-million 
patients with 27x more data than the 15 trillion tokens of text powering large AI language models. If newborn sequencing were
to become the standard globally, as we expect, data volumes would explode by 1,000x or more.
Data Volumes Should Explode As Measurement Costs Decline 
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources , including UK Biobank as of December 31, 2024, which may be provided upon request. For informational purposes only 
and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 130
ARK’s Big Ideas 2025: Multiomics
    130/148
    0.001
0.01
0.1
1
10
100
1000
$ (Log Scale)
Number Of Cells Profiled (Approximations)
Cost Per Single Cell
1
2
100
1K
3K
23K
123K
33M
277M
1B
2.5B
4.2B
7B
13B
20B
27B
2.6T
9T
11T
10^5
10^6
10^7
10^8
10^9
2024 2028
$ (Log Scale)
Cellular Foundation Model Training
$240 Million
$2 Million
Share of Cost:
Compute
Sequencing
Single Cell
Virtual Cell Foundation Models Are Collapsing The Cost Of Single-Cell Genomics
Single-cell genomics costs are dropping exponentially, in line with Wright’s Law.* According to ARK’s research, trillion-cell experiments 
are feasible now that single-cell sequencing costs have dropped to $0.01 per cell. Virtual cell foundation models are likely to become the 
killer AI application in drug discovery and precision medicine. By 2028, virtual cell costs could drop ~120x. Companies like 10x Genomics 
and Scale Bio are leading the way.
2024 ~$0.01
2024 ~$0.05
2019 ~$0.2
2014 ~$9
2009 ~$400
120x cost 
reduction
42%
42%
16%
58%
20%
22%
131
Note: The companies mentioned are examples of companies developing virtual cell models, but the list does not include all companies that may be developing these models, and which may do so more successfully. The 
information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be profitable. *Wright’s Law states that for every 
cumulative doubling of units produced, costs will fall by a constant percentage. See Winton 2019. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Tang et al. 
2009, Jovic et al. 2022, Ding et al. 2020, and Rood et al. 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation 
to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    131/148
    AlphaFold Is Redefining The Prediction Of Molecular Structures
AlphaFold has transformed the prediction of protein structures, surpassing traditional methods like X-ray crystallography and Cryo-EM 
in both productivity* and scope.** Each version of AlphaFold has delivered exponential gains, not only improving efficiency, but also 
unlocking entirely new structural predictions. AlphaFold1, for example, predicted single-chain proteins (∼60M structures), but 
AlphaFold3 can predict more than 1 billion structures, including multi-chain complexes, protein-ligand, protein-DNA/RNA interactions, 
and post-translational modifications.
1.1
289
508
0
100
200
300
400
500
600
Productivity Gains: Alphafold VS.
Traditional Methods
60
140
1000
0
200
400
600
800
1000
1200
Millions
Structures Predicted
AlphaFold
2018
AlphaFold2
2020
AlphaFold3
2024
AlphaFold
2018
AlphaFold2
2020
AlphaFold3
2024
Unit Of Productivity Gains
132
*Productivity gains are defined as the ratio of AlphaFold's Efficiency Metric—accuracy per cost per speed—to the average efficiency of traditional methods. **Scope is calculated as the Efficiency Metric × Number of Structures 
Predicted. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Jumper et al. 2021, CASP13 2019, and Liu et al. 2024, as of December 31, 2024, which may be provided 
upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are 
inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    132/148
    Minimal residual disease (MRD) testing can detect the recurrence of cancer up to 20 months earlier than traditional imaging. Yet, more 
than eight million suitable cancer patients in the US currently are not reimbursed for MRD testing. Now, at only 10% penetration, MRD 
testing should become the standard of care for every cancer patient, with multiple tests per year and follow-ups for five years. ARK’s 
research suggests that MRD testing could produce 700x more data than the largest genomics project to date, the UK Biobank, enabling 
advanced predictive and prescriptive analytics that could revolutionize how we understand, monitor, and manage cancer.
MRD Testing Is Catalyzing The Generation Of Genomics Data
0.3
0.7
8.1
Number (Millions)
0
1
2
3
4
5
6
7
8
9
10
MRD Patient Population
30
4320
21600
0
5000
10000
15000
20000
25000
PetaBytes
MRD Testing Data
MRD Tested Patients
2024
Current Patient
TAM
Future Patient
TAM
UK Biobank MRD TAM
Annual
MRD TAM Five-Year
Forecast
5-fold
144-fold
133
Note: “TAM”: Total addressable market. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Wan et al. 2017, Luskin et al. 2017, and Borfitz 2024 as of December 
31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not 
indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    133/148
    100
240
$0
$50
$100
$150
$200
$250
$300
Traditional Cancer
Screening
Universal MultiCancer Screening
Market Potential For Screening Methods
($ Billions)
0.82
3.15
0
0.5
1
1.5
2
2.5
3
3.5
Lives Saved
(Millions)
$1,500 
$300 
$0
$200
$400
$600
$800
$1,000
$1,200
$1,400
$1,600
Cost Efficiency Of Screening Methods
($ Billions)
Multi-Cancer Screening Could Save Lives And Cut Costs
Powered by liquid biopsy technology, multi-cancer screening is a revolutionary approach to early cancer detection. Unlike traditional 
methods, a single blood test can screen for multiple cancers simultaneously, streamlining screening and expanding its scope. The FDA’s 
recent approval of Guardant Health’s “Shield” colorectal cancer screening test suggests that the technology is proving its potential. 
Pending resolution of regulatory and reimbursement paths, universal multi-cancer screening could more than double the Total 
Addressable Market (TAM) in the US.
~2x increase
~5x 
Reduction
~4x 
Increase
Traditional Cancer
Screening
Universal MultiCancer Screening
Traditional Cancer
Screening
Universal MultiCancer Screening
134
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Philipson et al. 2023, Goddard et al. 2024, and Mannucci and Goel 2024 as of December 31, 2024, which 
may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future 
results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    134/148
    Target 
Hypothesis TargetTo-Hit
Hit-ToLead
Lead 
Optimization Preclinical Phase I Phase II Phase III Submission 
To Launch
~1 year ~1.5 years ~1.5 years ~1 year ~1.5 years ~2.5 years ~2.5 years
DISCOVERY DEVELOPMENT
~1.5 years
The “virtual cell” is a cutting-edge achievement that integrates 
artificial intelligence and single-cell genomics. By leveraging multimodal, multi-scale data, virtual cells use foundation models to 
simulate cellular functions and predict responses to perturbations 
across various biological states. 
A transformative leap in drug discovery—particularly target 
identification and validation—virtual cells enable in-silico 
experiments to test hypotheses, simulate perturbations, and identify 
viable targets efficiently and cost-effectively. They also predict how 
specific mutations or drugs influence cellular pathways, potentially 
accelerating the pipeline from target discovery to validation and 
reducing reliance on costly and time-consuming wet-lab 
experiments. We expect the same timeline reductions as AlphaFold 
achieved in the prediction of molecular structure. 
Companies like 10X Genomics, Illumina’s FluentBio, ScaleBio, and 
ParseBio are providing technologies that build Virtual Cells.
Single-Cell Genomics And 
AI Are Creating Virtual Cells 
And Transforming Drug 
Discovery
Patient Cell Single Cell Genomics +AI Virtual Cell
Academic Studies
That Can Take
Decades
135
Note: The companies mentioned are examples of companies developing virtual cell models, but those examples do not include all companies that may be developing these models, and which may do so more successfully. The 
information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be profitable. Source: ARK Investment Management 
LLC, 2025. This ARK analysis draws on a range of external data sources, including Bunne et al. 2024 and Rood et al. 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
Lead
Candidate IND
Hit
    135/148
    A self-driving lab (SDL) combines not only multiomics tools 
that read biology at scale, but also automation that enables 
high-throughput experimentation and large language models 
(LLMs) that analyze data and design the next set of 
experiments. The SDL’s integrated approach should enable 
the transition from automated to autonomous drug discovery 
and research, potentially accelerating the cycle of discovery 
and reducing costs/time while increasing efficiency. According 
to our research, SDLs could save up to two years and 
hundreds of millions of dollars in the discovery process. 
Currently achieving ~200x productivity gains, Recursion 
Pharmaceuticals is pioneering this approach. 
Self-Driving Labs Are Using 
Scalable Biology And AI To 
Revolutionize Drug Discovery
Self-Driving Lab
Target 
Hypothesis
Lead 
Optimization Preclinical Phase I Phase II Phase III Submission 
To Launch
~1 year ~1.5 years ~1.5 years ~1 year ~1.5 years ~2.5 years ~2.5 years
DISCOVERY DEVELOPMENT
~1.5 years
Lead
Candidate IND
Hit
136
Academic Studies
That Can Take
Decades
TargetTo-HitHit-ToLead
Note: Recursion Pharmaceuticals is the company currently pioneering the use of SDLs, but there may be other companies that utilize SDLs and which may ultimately do so more successfully. The information provided should not 
be used as the basis for any investment decision, and it should not be assumed that an investment in the company mentioned was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a 
range of external data sources, including Recursion 2024, Abolhasani and Kumacheva 2023, and Martin et al. 2023 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be 
considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    136/148
    Often called the "valley of death" in drug discovery, 
preclinical studies are prone to high failure rates that 
stem from predictive models like 2D cell systems and 
animal testing. Thanks to advances in microfluidics, 3D 
bioprinting, and organoids, next-generation preclinical 
models like organs-on-chips are evolving. The new 
systems are more physiological, scalable, and capable 
of supporting higher-throughput testing. Integrating 
Gen 2 models—grown from patient-derived biopsies—
with AI-powered analyses of vast datasets could lead 
to precise, patient-specific predictions that improve 
drug discovery outcomes dramatically.
Large Language Models are beginning to impact and 
accelerate clinical trials by optimizing regulatory 
filings, patient selection, trial designs, and data 
analysis.
Organs-On-Chips And AI Are 
Rescuing Drug Discovery 
From The Valley Of Death Target 
Hypothesis
Lead 
Optimization Preclinical Phase I Phase II Phase III Submission 
To Launch
~1 year ~1.5 years ~1.5 years ~1 year ~1.5 years ~2.5 years ~2.5 years
DISCOVERY
~1.5 years
DEVELOPMENT
Lead
Candidate IND
Hit
Percentage responding(%) 25
50
75
100
ED5 0 TD5 0
Therapeutic
effect Toxic
effect
Therapeutic
index
Gen 1
2D Human Systems
Animal Models
Organs On Chip
3D Human
Systems
Human Biopsy
Gen 2 Gen 3
3D Printing/
Microfluidics AI
Valley of Drug Discovery Death
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Zhou et al. 2023, Tong et al. 2024, and Thng et al. 2024 as of December 31, 2024, which may be provided 
upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts 
are inherently limited and cannot be relied upon. 137
Academic Studies
That Can Take
Decades
TargetTo-HitHit-ToLead
ARK’s Big Ideas 2025: Multiomics
    137/148
    Biotech valuations include little-to-nothing for the preclinical and Phase 1 stages of development, but increasingly they are likely to 
incorporate the higher probability of longer patent lives. AI-driven drug development could reduce time-to-market nearly 40%, from 
13 years to 8 years, while reducing total drug costs 4-fold, from $2.4 billion to $0.6 billion.
0
100
200
300
400
500
600
13 12 11 10 9 8 7 6 5 4 3 2 1 11 10 9 8 7 6 5 4 3 2 1 8 7 6 5 4 3 2 1
$ Millions Per Phase
Years To Commercialization
Drug Development Average CostHit-to-Lead
Lead 
Optimization
Pre-clinical
Safety
Phase I
Efficacy
Phase II
Phase III
Confirmation
Target-to-Lead
AI Is Transforming The Economics Of Drug Development
Time Cost Of Money
Failed Candidates
Successful Candidates
Licensing
INDUSTRY AVERAGE 2024
• $2.4 billion total
• $1.1 billion pre-clinical
• 7.5 human trial failures
• 13 years time-to-market
INITIAL AI-DRIVEN DRUG EFFORTS
• $1.5 billion total
• $0.6 billion pre-clinical
• 4.3 human trial failures
• 11 years time-to-market
FUTURE AI-DRIVEN DRUG DESIGN
• $0.6 billion total
• $0.13 billion pre-clinical
• 1.4 human trial failures
• 8 years time-to-market
ARK’s Big Ideas 2025: Multiomics
138
Note: 10% discount rate, 2024 dollars. For initial AI-driven drug efforts, assumes 70+% failure reduction in phase I that moderates to 20% in phase II and phase III. For future AI-driven drug design, assumes demonstrated 70+% 
reduction in phase I failure rates that moderates to a 50% failure reduction in phase II and 25% failure reduction in phase III. Assumes that pre-clinical efficiency is similar to what Absci has indicated is achievable. Also assumes 
that licensing timeframe reduces to 12 months from 18 months for future drugs. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may 
be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 
Forecasts are inherently limited and cannot be relied upon.
    138/148
    Reducing Development 
Timelines Should Increase 
The Value Of Patents And 
Commercialization
Industry Average
Current AI methods 
accelerate time-to-market by 
2 to 3 years, increasing the 
value of intellectual property 
by 30% to 50% 
Emerging AI methods could 
accelerate time-to-market by 
4 to 5 years, increasing the 
value of intellectual property 
by 70% to 80%
0.5x
1.0x
1.5x
2.0x
3 4 5 6 7 8 9 10
Patent Life Years Remaining At Commercialization
Value Of A Drug At Commercialization By Patent Life Remaining
Relative To Industry Average
Faster commercialization extends patent-protected revenue, allowing companies 
to generate higher returns for longer. Current AI efforts shorten development 
timelines by 2-3 years, increasing a drug's lifetime value by 30-50% compared to 
the industry average of five years. Emerging AI designs could accelerate time-tomarket by 4-5 years and boost the value by 70-80%.
Note: 10% discount rate. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only 
and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 139
ARK’s Big Ideas 2025: Multiomics
    139/148
    Industry Average Current AI methods 
accelerate time to 
market by 2 to 3 years, 
increasing IP value by 
30% to 50% 
Emerging AI 
methods could 
accelerate time to 
market by 4 to 5 
years, increasing IP 
value by 70% to 
80%
0.5x
1.0x
1.5x
2.0x
3 4 5 6 7 8 9 10
Patent Life Years Remaining At Commercialization
Value Of A Drug At Commercialization by Patent Life Remaining
Relative To Industry Average
-$1,500
-$500
$500
$1,500
$2,500
$3,500
$4,500
Years From Start Of Development
Drug Development Cumulative Cashflow
($ Millions)
Drug Development Cost Efficiencies Could Boost Cash Flow Significantly
AI Designed Drugs
Forecast
The value of AI-driven drug development should compound thanks to lower costs, an acceleration in time-to-market, and a longer 
timeline of patent-protected revenue. Over a 30-year period, the cumulative cashflow for the average AI-designed drug could reach $4 
billion, compared to less than $1 billion for traditional drug models. Even before a traditional drug breaks even, an AI-developed drug
could generate $2.5 billion in cashflow.
Traditional Drug
5 10 15 20 25 30
140
Note: 10% discount rate, 2024 dollars. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. 
For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts 
are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    140/148
    Cures for diseases can command much higher prices than traditional treatments. The average price of a cure today exceeds $1 million, 
nearly 15x the the lifetime prescription cost that would be necessary to manage the disease. Capturing revenue upfront and addressing 
much of the patient population well before patent expiration suggests that cures could become more valuable—potentially 20x more
valuable—than a typical drug and 2.4x more valuable than a chronic prescription that abates the disease.
Biological Cures Could Become Quite Valuable
$8,000 $75,000 
$390,000 
$620,000 $1,100,000 
$0
$200,000
$400,000
$600,000
$800,000
$1,000,000
$1,200,000
Price Of A Cure For Rare Disease
(Per Patient, On Average)
Typical 
Prescription 
Cost 
Per Year
Lifetime 
Prescription 
Cost
Lifetime 
Non-Drug 
Health 
Service Costs
Saved
Life Years 
Gained
Value Of 
Cure
$0
$500
$1,000
$1,500
$2,000
$2,500
$3,000
$3,500
Value Of A Cure For Rare Disease
($ Millions, On Average)
Cures avoid 
competition, 
with lifetime 
value 
captured on 
the first dose
Cures 
frontload cash 
in a drug’s 
commercial 
lifecycle
Typical 
Prescription 
Medicine
Drug That
Displaces 
Health Service 
Costs
Chronic 
Prescription 
That 
Abates 
The Disease
One Time 
Cure
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not 
be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 141
ARK’s Big Ideas 2025: Multiomics
    141/148
    0% 15% 30% 45%
Drug Development 
Return On Invested Capital
Cures And Accelerated 
Development Should 
Transform The Economics 
Of Drug Development
At inception Preclinical Phase I Phase II Phase III Commercial
(300) (200) 0 200 500 1,000 
200 600 1,000 1,400 2,000 3,000 
300 500 600 800 1,200 1,800 
600 700 900 1,000 1,300 1,800 
1,500 1,800 2,300 2,500 3,400 4,400 
1,800 2,000 2,500 2,700 3,500 4,400 
Value Of Assets at Different Clinical Phases
($ Millions)
The combination of AI-accelerated drug development and the value of cures for 
disease could boost returns on research and development (R&D) and improve the value 
of pre-clinical pipeline assets materially. According to our research, an AI-accelerated 
cure in the first phase of human testing could be worth more than $2 billion. 
Traditional assets typically pay back only the cost of capital for Phase 1 pipelines.
Representative 
Companies*
Industry Average
Beam, Crispr Tx, Editas, 
Intellia, Prime
Abcellera , Absci, 
Generate Bio, Isomorphic 
Lab, Lantern Pharma, 
Recursion, Relation, 
Roivant Sciences, Xtalpi
In Development
Traditional Drug, 
Traditional Development
Cure,
Traditional Development
Traditional Drug, 
AI Speedup, Traditional Fail Rate
Traditional Drug, 
AI Speedup And AI Success Rate
Cure,
AI Speedup, Traditional Fail Rate
Cure,
AI Speedup And AI Success Rate
142
Note: 10% discount rate. *Companies pursuing this strategy for drug development, though this does not imply that every asset the company has at that particular stage is worth that amount or that company return on R&D 
dollars will meet the modeled value. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be 
profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and 
should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
ARK’s Big Ideas 2025: Multiomics
    142/148
    AI-Driven Drug Development And Cures For Disease Could 
Reverse A Multi-Decade Decay In Pharma/Biotech Returns
-10%
0%
10%
20%
30%
40%
50%
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
2010
2012
2014
2016
2018
2020
2022
As AI and other technologies Return On Drug Development R&D
impact the drug development 
pipeline, industry-wide returns 
should turn around. 
Meanwhile, many development 
projects in place today could 
become sunk costs as more 
nimble AI drug development 
platforms beat them to market 
with cures that eliminate 
disease.
A bifurcation could grow 
between the returns generated 
by AI driven, cure-seeking firms 
and those traditional 
pharma/biotech companies 
that do not adapt.
14%
22%
30%
47%
Cure,
Traditional 
Development
Traditional Drug, 
AI Speedup, 
Traditional Fail Rate
Traditional Drug, 
AI Speedup and AI 
Success Rate
Cure,
AI Speedup, 
Traditional Fail Rate
Cure
AI Speedup And 
AI Success Rate
Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should 
not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 143
ARK’s Big Ideas 2025: Multiomics
    143/148
    Works Cited
144
ARK’s Big Ideas 2025
    144/148
    Abolhasani, M. and E. Kumacheva. 2023. “The Rise of Self-driving Labs in Chemical and Materials Sciences.” Nature 
Synthesis.
American Automobile Association. 2050. “Your Driving Costs.”
American Automobile Association. 2024. “How Much Does it Really Cost to Own a New Car?”
Artemis Terminal. 2024. 
Bitcoin Treasuries. 2024. 
Borfitz, D. 2024. “UK Biobank Pivoting to Platform-Only Model For Big Data Sharing.” Bio IT World.
Bunne, C. et al. 2024. “How to Build the Virtual Cell with Artificial Intelligence: Priorities and opportunities.” Cell.
BusinessWire. 2024. “Ultima Announces UG 100™ and Reveals Disruptive Cost and Accuracy Profile to Enable the 
Era of the $100 Genome and Beyond.”
Calma, J. 2024. “Google Orders Small Nuclear Reactors for its Data Centers.” Financial Times.
Carlson, R. 2022. “DNA Cost and Productivity Data, aka ‘Carlson Curves.’” Synthesis.
Capital IQ Pro. 2024.
CASP13. 2019. “13th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure 
Prediction.”
Circle. 2024. 
Cleveland, C. 2023. “United States Electricity History in Four Charts.” Boston University Institute for Global 
sustainability.
Clifford, J. 2024. “Boost Multi-Omics Analysis with GPU-Acceleration and Generative AI.” Nvidia Developer.
CoinGecko. 2024. 
Dediu, H. 2017. “Share of United States Households Using Specific Technologies.” Our World in Data.
DeLong, B. 1998. "Estimating World GDP, One Million B.C. — Present." Brad De Long’s homepage. Department of 
Economics, University of California, Berkeley.
de Selding, P.B. “Orbital ATK Believes in Satellite Servicing, but Not Rocket Reusability.” Space News.
Ding, J. et al. 2020. “Systematic Comparison of Single-cell and Single-nucleus RNA-sequencing methods.” Nature 
Biotechnology.
Dune Analytics. 2024.
Electric Capital. 2024. “Developer Report 2023.”
EV Volumes. 2025. 
Foust, J. 2014. “Reusability and Other Issues Facing the Launch Industry.” The Space Review.
Franchisee Resource Center. 2021. “Fast Casual Industry Analysis 2020 - Cost & Trends.” Franchise Help.
Glassnode. 2024.
Goddard, K. et al. 2024. “Estimation of Cancer Deaths Averted From Prevention, Screening, and Treatment Efforts, 
1975-2020.” JAMA Oncology.
Goddard, T. 2024. “AlphaFold Performance: Molecule Size, Speed, Memory, and GPU.” UCSF Resource for 
Biocomputing, Visualization, and Informatics.
Gorzynski, J. et al. 2022. “Ultrarapid Nanopore Genome Sequencing in a Critical Care Setting.” The New England 
Journal of Medicine.
Gumber, A. et al. 2024. “A Global Analysis of Renewable Energy Project Commissioning Timelines.” Applied 
Energy.
Hartig, K. 1998. “TUNING IN: Communications Technologies Historically Have Had Broad Appeal for Consumers,” 
The Wall Street Journal. 
Jiminez, C. et al. 2024. “SWE-bench: Can Language Models Resolve Real-World GitHub Issues?” arXiv.
Jovic, D. et al. 2022. “Single-cell RNA Sequencing Technologies and Applications: A Brief Overview.” Clinical and 
Translational Medicine.
Jumper, J. et al. 2021. “Highly Accurate Protein Structure Prediction with AlphaFold.” Nature.
Jurvetson, Steve 2024. “The Moore’s Law Update.” X.
Klotz, I. 2020. “Economics Of Rocket Reuse Still Up In The Air.” Aviation Week.
Kurzweil, R. 2006. “The Singularity Is Near: When Humans Transcend Biology.” Duckworth.
Lawrence Berkeley National Laboratory. 2024. “Energy Markets & Policy.” 
Liu, L. et al. 2024. “Technical Report of HelixFold3 for Biomolecular Structure Prediction.” arXiv.
Lovering, J. et al. 2016. “Historical Construction Costs of Global Nuclear Power Reactors.” Energy Policy.
Luskin, M. et al. 2018. “Targeting Minimal Residual Disease: A Path to Cure?” Nature Reviews Cancer.
Mannucci, A. and A. Goel. 2024. “Stool and Blood Biomarkers for Colorectal Cancer Management: An Update on 
Screening and Disease monitoring.” Molecular Cancer.
ARK’s Big Ideas 2025: Works Cited
    145/148
    Martin, H.G. et al. 2023. “Perspectives for Self-driving Labs in Synthetic Biology.” Current Opinion in Biotechnology.
McCormick, M. and J. Smith. 2025. “Microsoft in Deal for Three Mile Island Nuclear Power to Meet AI Demand.” 
Financial Times.
Miller, N.A. et al. 2015. “A 26-hour System of Highly Sensitive Whole Genome Sequencing for Emergency 
Management of Genetic Diseases.” Genome Medicine.
Model T Ford Forum. 2010. “How Many Miles Can a T Drive Down, Before it Needs a Rebuild?”
Morningstar Direct. 2024.
National Renewable Energy Laboratory (NREL). 2024. “Solar Resource Data.” 
NewSpace Index. 2025. “Small Satellite Launchers.”
OpenAI. 2024.
OpenPhilanthropy. 2025.
Philipson, T. et al. 2023. “The Aggregate Value of Cancer Screenings in the United States: Full Potential Value and 
Value Considering Adherence.” BMC Health Services Research.
Pitchbook. 2024.
Recursion. 2024. “Recursion Unveils LOWE Drug Discovery Software at the J.P. Morgan Healthcare Conference.”
Rood, J. et al. 2024. “Toward a Foundation Model of Causal Cell and Tissue Biology with a Perturbation Cell and 
Tissue Atlas.” Cell.
Rwa.xyz. 2024
Saunders, C.J. et al. 2021. “Rapid Whole-genome Sequencing for Genetic Disease Diagnosis in Neonatal Intensive 
Care Units.” Science and Translational Medicine.
Shaikhnag, A. 2024. “Westinghouse Facilitates Improved Nuclear Power with Its 3D Printed Nozzles.” 3D printing 
Industry.
Sidoti, O. et al. 2024. “Mobile Fact Sheet.” Pew Research Center.
Small Modular Reactor (SMR) Global Tracker. 2024.
Smil, V. 2000. “Energy in the Twentieth Century: Resources, Conversions, Costs, Uses, and Consequences.” Annual 
Review of Energy Environment.
Smyth, J. 2024 “Amazon Buys Stake in Nuclear Energy Developer in Push to Power Data Centres.” Financial Times.
SpaceX. 2024. “STARSHIPS ARE MEANT TO FLY—2024 PROGRESS REPORT.” 
Tang, F. et al. 2009. “mRNA-Seq Whole-transcriptome Analysis of a Single Cell.” Nature Methods.
Tether. 2024.
The Block. 2024.
Thng, D. et al. 2024. “A Functional Personalised Oncology Approach Against Metastatic Colorectal Cancer in 
Matched Patient Derived Organoids.” NPJ Precision Oncology.
TicData Treasury. 2024. “Table 5: Major Foreign Holders of Treasury Securities.”
Token Terminal. 2024.
Tong, L. et al. 2024. “Patient-derived Organoids in Precision Cancer Medicine.” MED.
TradingView. 2024. 
Tyrer-Jones, A. 2025. “3D Printer Nuclear Submarine Components from AML3D Delivered to US Navy.” 3D PRINTING 
INDUSTRY.
Ulvog, J. 2012. “Travel Cost by Stagecoach in 1870s – Part One.” Outrun Change.
UK Biobank. 2024.
U.S. Department of Agriculture. 2025. “Food Expenditure Series.”
U.S. Department of Agriculture. 2025. “Farm Labor.”
U.S. CENSUS BUREAU. 2023. “2021 SUSB Annual Data Tables by Establishment Industry.”
Visa OnChain Analytics. 2024.
Wallstreet Zen. 2024.
Wan, J. et al. 2017. “Liquid Biopsies Come of Age: Towards Implementation of Circulating Tumour DNA.” Nature 
Reviews Cancer.
Winton, B. 2019. “Moore’s Law Isn’t Dead: It’s Wrong—Long Live Wright’s Law.” ARK Investment Management LLC.
World Nuclear Association. 2024. “Small Modular Reactor Global Tracker.” 
Worldpay Inc. 2018. “2018 Global Payments Report.” 
Worldpay Inc. 2019. “2019 Global Payments Report.” 
Worldpay Inc. 2020. “2020 Global Payments Report.” 
Worldpay Inc. 2021. “2021 Global Payments Report.” 
Worldpay Inc. 2022. “2022 Global Payments Report.” 
ARK’s Big Ideas 2025: Works Cited
    146/148
    Worldpay Inc. 2023. “2023 Global Payments Report.” 
World Gold Council. 2024. “Gold ETFs, holdings and flows.”
World Nuclear Association. 2024. “United State of America Vallecitos Permanent Shutdown.”
Zhou, Z. et al. 2023. “Harnessing 3D In Vitro Systems to Model Immune Responses to Solid Tumours: A Step 
Towards Improving and Creating Personalized Immunotherapies.” Nature Reviews Immunology.
ZipRecruiter. 2024.
ARK’s Big Ideas 2025: Works Cited
    147/148
    ARK’s Big Ideas 2025: Disclosure
For more research on disruptive innovation visit www.ark-invest.com
©2021-2026, ARK Investment Management LLC. No part of this material may be reproduced in any form, or referred to in any other publication, without the express written
permission of ARK Investment Management LLC (“ARK”).
Please note, companies that ARK believes are capitalizing on disruptive innovation and developing technologies to displace older technologies or create new markets may not
in fact do so and/or may face political or legal attacks from competitors, industry groups, or local and national governments.
ARK aims to educate investors and to size the potential opportunity of Disruptive Innovation, noting that risks and uncertainties may impact our projections and research
models. Investors should use the content presented for informational purposes only, and be aware of market risk, disruptive innovation risk, regulatory risk, and risks related
to Deep Learning, Digital Wallets, Battery Technology, Autonomous Technologies, Drones, DNA Sequencing, CRISPR, Robotics, 3D Printing, Bitcoin, Blockchain Technology, etc.
Cryptocurrency Risk. Cryptocurrencies (also referred to as “virtual currencies” and “digital currencies”) are digital assets designed to act as a medium of exchange.
Cryptocurrency is an emerging asset class. There are thousands of cryptocurrencies, the most well-known of which is bitcoin. Cryptocurrency generally operates without
central authority (such as a bank) and is not backed by any government. Cryptocurrency is not legal tender. Federal, state and/or foreign governments may restrict the use and
exchange of cryptocurrency, and regulation in the U.S. is still developing. The market price of bitcoin and other cryptocurrencies have been subject to extreme fluctuations.
Similar to fiat currencies (i.e., a currency that is backed by a central bank or a national, supra-national or quasi-national organization), cryptocurrencies are susceptible to theft,
loss and destruction. Cryptocurrency exchanges and other trading venues on which cryptocurrencies trade are relatively new and, in most cases, largely unregulated and may
therefore be more exposed to fraud and failure than established, regulated exchanges for securities, derivatives and other currencies. Cryptocurrency exchanges may stop
operating or permanently shut down due to fraud, technical glitches, hackers or malware, which may also affect the price of cryptocurrencies. Cryptocurrency Tax Risk. Many
significant aspects of the U.S. federal income tax treatment of investments in bitcoin and other cryptocurrencies are uncertain and still evolving.
The content of this presentation is for informational purposes only and is subject to change without notice. This presentation does not constitute, either explicitly or
implicitly, any provision of services or products by ARK and investors are encouraged to consult counsel and/or other investment professionals as to whether a particular
investment management service is suitable for their investment needs. All statements made regarding companies or securities are strictly beliefs and points of view held by
ARK and are not endorsements by ARK of any company or security or recommendations by ARK to buy, sell or hold any security. Historical results are not indications of future
results. Certain of the statements contained in this presentation may be statements of future expectations and other forward-looking statements that are based on ARK's
current views and assumptions and involve known and unknown risks and uncertainties that could cause actual results, performance or events to differ materially from those
expressed or implied in such statements. The matters discussed in this presentation may also involve risks and uncertainties described from time to time in ARK's filings with
the U.S. Securities and Exchange Commission. ARK assumes no obligation to update any forward-looking information contained in this presentation. Certain information was
obtained from sources that ARK believes to be reliable; however, ARK does not guarantee the accuracy or completeness of any information obtained from any third party. ARK
and its clients as well as its related persons may (but do not necessarily) have financial interests in securities or issuers that are discussed.
ARK Investment Management LLC
    148/148

    ARK's Big Ideas 2025: Unlocking Exponential Growth

    • 1. FEBRUARY 04, 2025 RESEARCH REPORT FOR INFORMATIONAL PURPOSES ONLY WWW.ARK-INVEST.COM ARK Investment Management LLC. This is not a recommendation in relation to any named particular securities/cryptocurrencies and no warranty or guarantee is provided. Any references to particular securities/cryptocurrencies are for illustrative purposes only. There is no assurance that the Adviser will make any investments with the same or similar characteristics as any investment presented. The reader should not assume that an investment identified was or will be profitable. PAST PERFORMANCE IS NOT INDICATIVE OF FUTURE PERFORMANCE, FUTURE RETURNS ARE NOT GUARANTEED. BIG IDEAS 2025
    • 2. Risks Of Investing In Innovation Please note: Companies that ARK believes are capitalizing on disruptive innovation and developing technologies to displace older technologies or create new markets may not in fact do so. ARK aims to educate investors and seeks to size the potential investment opportunity, noting that risks and uncertainties may impact our projections and research models. Investors should use the content presented for informational purposes only, and be aware of market risk, disruptive innovation risk, regulatory risk, and risks related to certain innovation areas. Please read risk disclosure carefully. DISRUPTIVE INNOVATION RAPID PACE OF CHANGE UNCERTAINTY AND UNKNOWNS EXPOSURE ACROSS SECTORS AND MARKET CAP RISK OF INVESTING IN INNOVATION REGULATORY HURDLES COMPETITIVE LANDSCAPE POLITICAL OR LEGAL PRESSURE Sources: ARK Investment Management LLC, 2023. à Aim to understand the regulatory, market, sector, and company risks. (See Disclosure Page) à Aim for a cross-sector understanding of technology and combine top-down and bottom-up research. ARK’s Big Ideas 2025
    • 3. ARK’s Big Ideas 2025 Table Of Contents Introduction Convergence AI Agents Bitcoin Stablecoins Scaling Blockchains Robotaxis Autonomous Logistics Energy Robotics Reusable Rockets Multiomics 04 05 15 29 41 54 68 84 93 103 114 125
    • 4. ARK’s Big Ideas 2025 BIG IDEAS 2025 4 As we stand on the cusp of a new era of unprecedented growth, ARK Invest’s “Big Ideas 2025” illuminates the complex convergences between and among the five technologically enabled innovation platforms evolving today: Artificial Intelligence, Robotics, Energy Storage, Public Blockchains, and Multiomic Sequencing. These platforms should drive exponential advances across industries and catalyze a step change in global economic growth. Presented in this year’s report are 11 Big Ideas illustrating the massive transformations occurring today. Our research suggests that these Big Ideas are poised to boost productivity dramatically, to revolutionize industries, and to create long-term investment opportunities. Join our journey into the future of innovation and its profound implications for investors, businesses, and society at large. Welcome to ARK’s Big Ideas 2025. Unlocking Exponential Growth Through Disruptive Innovation
    • 5. The Second Half Of The Chessboard Convergence Among Technology Platforms Leading To Significant Acceleration In Macroeconomic Growth Brett Winton CHIEF FUTURIST ARK’s Big Ideas 2025 5
    • 6. Five Innovation Platforms Are Catalyzing Accelerated Growth Public Blockchains Upon large-scale adoption, all money and contracts will likely migrate onto Public Blockchains that enable and verify digital scarcity and proof of ownership. The financial ecosystem is likely to reconfigure to accommodate the rise of Cryptocurrencies and Smart Contracts. These technologies increase transparency, reduce the influence of capital and regulatory controls, and collapse the costs of contract execution. In such a world, Digital Wallets will become increasingly necessary as more assets become money-like and corporations and consumers adapt to the new financial infrastructure. Corporate structures may be called into question. Multiomics The cost to gather, sequence, and understand digital biological data is falling precipitously. Multiomic Technologies provide research scientists, therapeutic organizations, and health platforms with unprecedented access to DNA, RNA, protein, and digital health data. Cancer care should transform with pan-cancer blood tests. Fed by rich multiomic data and powered by Programmable Biology, AI systems running autonomous labs could collapse the cost of drug discovery, development, and trial, transforming returns in a sector that has stagnated. Biological discoveries should power novel Precision Therapies that target and cure rare diseases and chronic conditions, unlocking profound economics. Over time, the design and synthesis of novel biological constructs will yield advances in agriculture, material science, and even computation. AI Computational systems and software that evolve with data can solve intractable problems, automate knowledge work, and accelerate technology’s integration into every economic sector. The adoption of Neural Networks should prove more momentous than electrification and potentially create tens of trillion dollars of value. At scale, these systems will require unprecedented computational resources, and AI-specific compute hardware should dominate the Next Gen Cloud datacenters that train and operate AI models. The potential for end-users is clear: a constellation of AI-driven Intelligent Devices that permeate people's lives, changing the way that they spend, work, and play. The adoption of artificial intelligence should transform every sector, impact every business, and catalyze every innovation platform. Robotics Catalyzed by artificial intelligence, Humanoid Robots should operate alongside humans and navigate legacy infrastructure, changing the way products are made and sold, and eventually the way we live our lives at home. 3D Printing should contribute to the digitization of manufacturing, increasing not only the performance and precision of end-use parts, but also the resilience of supply chains. Meanwhile, the world’s fastest robots, Reusable Rockets, should continue to reduce the cost of launching satellite constellations and enable uninterruptible connectivity and earth observation. A nascent innovation platform, robotics could collapse the cost of transporting across distance, with hypersonic travel, the cost of manufacturing complexity with 3D printers, and the cost of physical work with AI-guided robots. Energy Storage The declining costs of Advanced Battery Technology should cause an explosion in form factors, enabling Autonomous Mobility systems that collapse the cost of transportation. Electric drivetrain cost declines should unlock micro-mobility and aerial systems, including flying taxis, enabling business models that transform cities. Autonomy should reduce the cost of taxi, delivery, and surveillance by an order of magnitude, enabling frictionless transport that will increase the velocity of ecommerce and make individual car ownership the exception rather than the rule. These innovations combined with large-scale stationary batteries and Distributed Energy Generation, notably solar and small-scale fission, should cause a transformation in energy, substituting electricity for liquid fuel and increasing system-wide resilience, reliability, and production. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Convergence 6
    • 7. Row 2X Wheat Grains Value Year Computers Crossed The Same Compounding Threshold 8 128 2 teaspoons of wheat 1975 16 65 thousand 9 loaves of bread 1983 24 17 million 4 year’s sustenance 1998 32 4 billion 7 pounds of gold 2008 40 1 trillion 1 ton of gold 2018 48 300 trillion 20% of India GDP in 600 AD 2023 56 70 quadrillion 50x Global GDP in 600 AD 2027e 64 18 quintillion 9x Global GDP in 2024 2030e By lore, when he agreed to the reward, the Indian ruler thought it quite reasonable. By the 6th row of the chessboard, however, he had exhausted India’s treasury, having paid just ~.001% of the bill. Today, computational advance has completed the equivalent of the 6th row on the chessboard. Entering the AI cycle in 2018, computers had crossed 40 performance doublings and in 2023 surpassed 48. By the end of this decade, thanks to the acceleration in AI, computation could reach the end of the chessboard. In Compounding, The Real Action Takes Place On The Second Half Of The Chessboard I don’t require much, your highness, just a grain of wheat on the first square of the board, doubling consecutively on each square thereafter. Legendarily requesting his reward for inventing chess, the 6th century advisor to the Indian ruler: Note: “GDP”: Gross Domestic Product, aggregate global production for the global datapoints. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, Kurzweil 2006 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Convergence 7
    • 8. 1975 8 performance doublings 1983 16 performance doublings 1998 24 performance doublings 2008 Halfway through the chessboard Google announces the TPU 2023 48 performance doublings 2027 56 performance doublings 2030 End of the chessboard 2018 40 performance doublings 2028 48 performance doublings 1E-02 1E+00 1E+02 1E+04 1E+06 1E+08 1E+10 1E+12 1E+14 1E+16 1E+18 Compute Capability (Per 2024 $) Cumulative Compute Purchased (Unit Of xAI Colossus-Sized Datacenters) Integrated Circuit Performance Improvement Curve 100 Billionth 1 Billionth 100 Millionth 1 Millionth 10 Thousandth 1 Hundredth 1 100 10,000 1,000,000 AI Is Accelerating The Performance Doubling Rate, Closing In On The End Of The Chessboard Enabled primarily by architectural improvements in AI systems, performance per dollar of AI compute is expected to improve >1000x by 2030. At that time, we expect that compute performance will have doubled 64 times since the advent of the integrated circuit. Improvements in AI Architecture Driving Compute Performance Compute Hardware Driving Performance Improvements Each Circle Represents 1 Performance Doubling Hollow Circles Are Forecasts ARK’s Big Ideas 2025: Convergence 8 Note: The “Compute Capability per $2024 Dollar” displayed on the Y axis above is a single metric that measures 1000s of computations per second purchasable for $1. For AI compute performance, computational impact is adjusted upwards by the rate of AI architectural performance per dollar improvements according to ARK’s research. It's a very large number, so we use scientific notation, e.g., "1E+21," meaning 1 followed by 21 zeros. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Kurzweil 2005 and Jurvetson 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    • 9. 2023 2024 0 50 100 150 200 250 0 500 1000 1500 # Of Hypotheses Tested Per FTE Researcher H100 Equivalents Compute In Recursion’s Datacenter Recursion Drug Development Efficiency 10/1/23 4/18/24 11/4/24 5/23/25 Log Error Rate Scale SWE-Bench Verified Best In Class AI Software Engineering AI Advances Should Unlock Massive Market Opportunities 72% 54% As AI continues to accelerate, robotaxis should proliferate, drug development timelines and costs should collapse, and AI agents should solve software engineering challenges autonomously, monitoring and modifying systems around the clock. 4% 22% 38% 45% FSD 11.5 12/31/23 FSD 12.5 10/24/24 FSD 13.4 1Q25* FSD 13.6 2H25* 1 10 100 1000 10000 100000 1000000 - 100,000 200,000 Miles Per Intervention H100 Equivalent Compute In Tesla’s Datacenter Tesla FSD Performance Waymo intervention rate post commercialization Human driver accident rate ARK’s Big Ideas 2025: Convergence 9 *Estimates. Note: “FSD”: Full Self-Driving. SWE-bench is a dataset that tests systems' ability to solve GitHub issues automatically. The dataset collects 2,294 Issue-Pull Request pairs from 12 popular Python repositories. Evaluation is performed by unit test verification using post-PR behavior as the reference solution. See Jiminez et al. 2024. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Tesla, Recursion, and SWEbench.com, and OpenAI as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    • 10. Convergence Is Accelerating The Technology Revolution ARK measures the degree to which technologies serve as catalysts between and among innovation platforms. The convergences among them are increasing, with network density up 30% in the past year. AI is proving more critical to unlocking the value of Precision Therapies and Multiomic Technologies. Smart Contract ecosystems are serving as test beds in which autonomous AI agents can be renumerated for sharpening their capabilities. Next Gen Cloud energy demand is pulling forward timelines for Distributed Energy Generation. Catalyzed Increase In Market Opportunity Order Of Magnitude+ Multiples Very Meaningful Meaningful Potential ARK’s Big Ideas 2025: Convergence 10 Note: “Network density” measures the degree of interconnectedness between nodes relative to the maximum potential interconnectedness. In our research, if every technology were expected to catalyze another technology to increase in value by an order of magnitude or more, that would equate to a fully interconnected network. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    • 11. The Acceleration In Neural Networks Is Accelerating Every Other Disruptive Technology Cryptocurrencies Smart Contracts Digital Wallets Precision Therapies Multiomic Technology Programmable Biology Neural Networks Next Gen Cloud Intelligent Devices Autonomous Mobility Advanced Battery Technology Reusable Rockets Humanoid Robotics 3D Printing Distributed Energy Generation CryptocurrenciesSmart Contracts Digital Wallets Precision Therapies Multiomic Technology Neural Networks Next Gen Cloud Intelligent Devices Autonomous Mobility Advanced Battery Technology Reusable Rockets Humanoid Robotics 3D Printing Programmable Biology Distributed Energy Generation Technology Catalyzing Technology Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Convergence 11
    • 12. Advances In AI, Energy Storage, And Public Blockchains Are Critical To The Pace Of Technological Advance Technology Importance As A Catalyst Advances another technology by 1 order of magnitude Neural Networks Intelligent Devices Next Gen Cloud Advanced Battery Technology Digital Wallets Reusable Rockets 3D Printing Cryptocurrencies Smart Contracts Autonomous Mobility Distributed Energy Generation Multiomic Technologies Precision Therapies Programmable Biology Humanoid Robots Among disruptive technologies, Neural Networks are the most important catalyst. According to our research, advances in neural networks will increase the value of six of the other 14 technologies by at least an order of magnitude, creating massive market expansions for Next Gen Cloud, Intelligent Devices, Autonomous Mobility, Humanoid Robots, Precision Medicine, and Multiomic Technology. The AI-led revolution in technology is likely to lead to dramatic productivity gains and a stepfunction increase in economic growth. Technology Platform AI Energy Storage Multiomics Robotics Public Blockchains SCALE Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Convergence 12
    • 13. Technology Inflection Should Lead To GDP Inflection Structural changes in the underlying rate of macroeconomic growth are the historical rule, not the exception. After 100,000 years of economic stagnation leading to year 1, innovations, notably writing, allowed empires to stitch together continents, which quadrupled real growth through 1000 AD. Agricultural innovations enabled denser populations and labor specialization, leading to a doubling of growth through 1500 to 0.3% per annum. During the 400 years through 1900, annual GDP growth doubled again to 0.6% as the Enlightenment and Industrial Revolution swept the globe. The Second Industrial Revolution—marked by electrification, the automobile, and telephony—ushered in the modern age and quintupled growth to 3% on average over the last 125 years. Technological breakthroughs in AI, Autonomous Mobility, and Humanoid Robots are likely to turbocharge productivity again, causing another step function in growth over the next 5 to 10 years. 0.0% 0.1% 1.0% 10.0% 100.0% Compounded Annual Growth (Log Scale) Global Real GDP Growth Log Years Until 2050 100,000 BC 1 1000 1500 1900 2024 .037% 0.14% 0.3% 0.6% 3% IMF Forecast 3.1% ARK Forecast 7.3% 2030 ARK’s Big Ideas 2025: Convergence 13 Note: “Real GDP”: Aggregate Global Production adjusted for changes in price levels. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources including DeLong 1998 and Open Philanthropy as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    • 14. - 50 100 150 200 250 300 $ Trillions (Nominal) Innovation Should Drive A Momentous Business Cycle Disruptive Innovation Could Grow To Dominate Markets 2014 2019 2024 2030e Disruptive innovation could command more than two-thirds of the global equity market, compounding at a 38% rate through 2030. Within innovation, market value could broaden out beyond the so-called Mag 6. Non-innovation businesses could begin to shrink as technological deflation associated with innovation threatens margins and competitiveness, even amid rapid macroeconomic growth. GDP Non-disruptive Mag 6** Other Disruptive Innovation* 45 1.7 4.7 73 92 127 220 38% CAGR ARK’s Big Ideas 2025: Convergence 65 5 9 103 16 12 89 -2% CAGR 60 24% CAGR 140 50% CAGR 14 Note: “CAGR”: Compounded Annual Growth Rate. *We include cryptocurrencies and crypto assets in disruptive innovation capitalization. **Microsoft, Nvidia, Meta, Apple, Amazon, and Alphabet. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. Global Market Capitalization
    • 15. Redefining Consumer Interactions And Business Workflows ARK’s Big Ideas 2025 AI Agents Nicholas Grous Frank Downing ASSOCIATE PORTFOLIO MANAGER DIRECTOR OF RESEARCH, NEXT GENERATION INTERNET Varshika Prasanna Jozef Soja RESEARCH ASSOCIATE, NEXT GENERATION INTERNET RESEARCH ANALYST, NEXT GENERATION INTERNET 15
    • 16. AI Agents: Understand intent through natural language Plan using reasoning and appropriate context Take action using tools to accomplish the intent Improve through iteration and continuous learning What Are AI Agents? Platform-Level Agents Siloed, Single-Purpose Agents Usefulness Level Of Intelligence Universal Agents AI agents are poised to accelerate the adoption of digital applications and create an epochal shift in human-computer interaction. ARK’s Big Ideas 2025: AI Agents Smarter models using more tools to accomplish highervalue tasks Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 16
    • 17. 0 1 2 3 4 5 6 7 8 9 10 - 2 4 6 8 10 12 14 Revenue ($B) Years Since Product Launch AI Monetization is Outpacing Social Media OpenAI Instagram Facebook TikTok YouTube WeChat AI Is Accelerating The Adoption of Hardware And Software OpenAI could surpass $10 billion in revenue in 2025, monetizing at a faster rate than social media companies over the past decade. If the adoption of ChatGPT is an indicator, AI should drive rapid demand for a range of new technologies. ARK’s Big Ideas 2025: AI Agents $0 $1 $2 $3 $4 $5 $6 $7 $8 $9 $10 - 5 10 15 Revenue ($ Billions) Years Since Product Launch AI Monetization Is Outpacing Social Media OpenAI Instagram Facebook TikTok YouTube WeChat 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1974 1983 1992 2001 2010 2019 2028E 2037E Percent Penetration US Consumer Hardware Adoption PC Cellphone Smartphone Smart Home Smartwatch VR Headset AI Hardware Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Hartig 1998, Dediu 2017, and Sidoti et al. 2024, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 17 Starting in 2025, we believe most hardware sales will be AI-enabled, leading to fast mass adoption.
    • 18. Embedded in the operating systems of consumer hardware, AI agents empower consumers to delegate all discovery and research to AI—a massive time-saver. Curated AI results will contextualize digital ad impressions. AI Agents Will Transform Consumer Search And Discovery TEXT / VOICE COMMAN D CHECKOU T AWARENESS AUTONOMOUS DISCOVERY + RESEARCH CONVERSION QUERY PERSONAL IZED AI AGENT CURATED RESULTS SEARCH ENGINE 1 SEARCH ENGINE 2 SEARCH ENGINE 3 MARKETPLACE 1 MARKETPLACE 2 MARKETPLACE 3 = Ad Opportunities Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 18 ARK’s Big Ideas 2025: AI Agents
    • 19. AI Agent Revenue Opportunities AI-Mediated Ads Should Take The Lion’s Share Of Digital Ad Revenue by 2030 If search shifts to personal AI agents, AI-mediated ad revenue could surge. By 2030, we believe AI ad revenue could account for more than 54% of the $1.1 trillion digital ad market. ARK’s Big Ideas 2025: AI Agents 0% 1% 2% 3% 4% 5% 6% 7% 8% August-21 August-22 August-23 August-24 Share Of Page Views ChatGPT, Claude, Perplexity, And Bing Share Of Search VS. Google (Based On 7-Day Rolling Average Of Page Views) Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 19 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% $- $200 $400 $600 $800 $1,000 $1,200 2024 2025E 2026E 2027E 2028E 2029E 2030E Gross Spend ($ Billions) Global Advertising Forecast AI Ad Revenue Digital Ad Spend AI Ad Revenue % of Total Digital Spend
    • 20. AI-Mediated Shopping Is Likely To Approach 25% Of Addressable Online Sales Globally By 2030 The growing use of AI agents in consumer shopping should streamline product discovery, personalization, and purchasing. ARK research suggests that AI agents could facilitate nearly $9 trillion in global gross online consumption by 2030. Note: ARK measures “total addressable online sales” using internal and external estimates for global e-commerce and addressable services spending out to 2030. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 20 0% 5% 10% 15% 20% 25% 30% $- $1 $2 $3 $4 $5 $6 $7 $8 $9 $10 2023 2024 2025E 2026E 2027E 2028E 2029E 2030E Percent $ Trillions AI-Facilitated Online Spend AI-Facilitated Spend (LHS) % of Total Addressable Online Sales ARK’s Big Ideas 2025: AI Agents
    • 21. Digital Wallets Are Positioned For Continued Share Gains In E-Commerce ARK’s research suggests that digital wallets empowered by AI purchasing agents—taking share from payment methods like credit and debit cards—could account for 72% of all ecommerce transactions by 2030. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2018 2019 2020 2021 2022 2023 2024 2025E 2026E 2027E 2028E 2029E 2030E Percent Share Of E-Commerce Purchase Volume Digital Wallet Other Payment Methods 21 Note: Other Payment Methods for E-commerce purchase volume include Credit Card, Debit Card, Bank Transfer, Cash on Delivery, and Buy Now Pay Later (BNPL). “Credit Card” and “Debit Card” refer to stored card credentials, while “Digital Wallet” includes credit/debit cards stored within the digital wallet. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including WorldPay as of December 31, 2024, which may be provided upon request. 2018-2023 figures are documented in the Worldpay Global Payments Reports. 2024-2030 figures are ARK's adoption estimates for e-commerce payment methods. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: AI Agents
    • 22. Digital Wallets Are Consolidating Financial Services And E-Commerce Based on their consumer-facing operations, the market is valuing leading digital wallet platforms like Block, Robinhood, and SoFi at $1,800 per user today. $1,800 $250 $1,000 $300 $150 $100 $- $200 $400 $600 $800 $1,000 $1,200 $1,400 $1,600 $1,800 $2,000 Checking & Savings Lending Insurance Investing E-Commerce Total $ Digital Wallet Enterprise Value Per User (US) 22 Note: “Checking & Savings” data include Checking & Savings, Debit Card, and Peer to Peer Payments. “Lending” data include Loans & Mortgages, Credit Card, and Buy Now Pay Later (BNPL). Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources representing the revenues, margins, and multiples of leading digital wallets and established incumbents. as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: AI Agents
    • 23. Agentic lead generation should move digital wallets upstream to capture market share in e-commerce and digital spending globally. “One-click checkout” should give way to “one-query purchasing.” Digital Wallet Purchasing Agents Could Become Central To The Shopping Journey Interaction With Digital Wallet Existing System System With Purchasing Agent Consumer Search Engine / 3rd-Party Aggregator Untailored Results Shopping Cart / Bag Checkout Pay With Digital Wallet SEARCH LAYER AGGREGATION LAYER CHECKOUT LAYER SEARCH LAYER AGGREGATION LAYER CHECKOUT LAYER Consumer Marketplaces Personalized Results Automatic Checkout Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 23 ARK’s Big Ideas 2025: AI Agents
    • 24. In 2030, AI-powered purchasing agents could add between $50 and $200 per user to the enterprise value (EV) of digital wallets in the US. Purchasing Agents Should Increase The Enterprise Value Of Digital Wallets, Notably In E-Commerce 2030 Forecast Global Gross Spend Powered By AI Agents ~ $9 Trillion Percent Of AI Agent Spend In Digital Wallets 5% 10% 20% Lead Generation Take-Rate For Digital Wallets 2.5% $10B $20B $40B 5% $20B $40B $90B 10% $40B $90B $200B Based on lead-generation take rates, AI purchasing agents could generate global revenue between $40 billion and $200 billion—ARK’s base and bull cases, respectively—for digital wallet platforms in 2030. $100 $50 $200 $- $50 $100 $150 $200 $250 $300 $350 2023 2030E - Base Case 2030E - Bull Case $ AI-Driven Digital Wallet EV Per User In E-Commerce (US) E-Commerce Forecast - Additional Purchasing Agent Opportunity Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 24 ARK’s Big Ideas 2025: AI Agents
    • 25. In The Enterprise, Agents Will Increase Productivity Via Software INBOUND REQUESTS AI Agent Platform SOFTWARE INVESTMENT LABOR SAVINGS Traditional Customer Support AI-Enabled Customer Support Companies that deploy agents should be able to increase unit volume with the same workforce and/or optimize their workforce toward higher-value activities. As AI evolves, agents are likely to handle a higher percentage of workloads and to complete highervalue tasks independently. HUMAN AGENTS INBOUND REQUESTS CUSTOMERS CUSTOMERS HUMAN AGENTS Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 25 ARK’s Big Ideas 2025: AI Agents
    • 26. AI Cost Declines Should Impact Agent Economics Significantly New products from OpenAI and Salesforce are supplementing human customer service representatives cost-effectively. Even at a fixed cost of $1 per conversation, AI agents could save enterprises significant sums once they can handle 35% of customer service inquiries. AI agents also should lower onboarding and hiring costs, as well as seat-based software costs, while scaling more easily than human labor. Potential Enterprise Customer Service Savings** ($ Billions) Dollar Cost Per Conversation 8x $290 $430 $500 $540 4x $140 $220 $250 $270 2x $70 $110 $130 $130 1x $40 $50 $60 $70 $1.00 $0.50 $0.25 $0.125 $0 $20 $40 $60 $80 $100 $120 $140 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% $ Billions Percent Of Customer Interactions Handled Successfully By AI AI Improvements In Customer Service Should Shift Labor Costs To Software Costs* Human Labor Costs AI Software Costs Savings Current U.S. Human Labor Spend Demand For Customer Service Relative To Today *Assumes fixed cost of $1 per customer interaction. **Assuming 70% of interactions can be handled by AI in all scenarios, corresponding to the 70% bar in the left-hand chart. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including ZipRecruiter and OpenAI, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 26 ARK’s Big Ideas 2025: AI Agents
    • 27. 0 -1 7 11 -7 -10 -10 -5 0 5 10 15 Share Shift (Percentage Points) PP 19% 31% 23% 39% 15% 27% 0% 20% 40% 60% Spend CAGRs Estimated Enterprise Software Stack Evolution Infrastructure Layer Platform Layer Application Layer o3 (OpenAI) Q Developer Agent (Amazon) Claude 3.5 Sonnet + Tools (Anthropic) Genie (Cosine) SWE-agent + GPT 4 (Princeton, OpenAI) Claude 2 + RAG (Anthropic) 0% 10% 20% 30% 40% 50% 60% 70% 80% Sep-23 Dec-23 Mar-24 Jun-24 Sep-24 Dec-24 SWE-Bench Verified Score* Percentage Of Real World Coding Tasks Solved By Coding Agents The coding skills of AI agents are improving rapidly, accelerating the software development lifecycle. As the cost to create software declines, software production should accelerate and sway enterprise “build vs. buy” decisions, displacing traditional software incumbents that are slow to adapt. As customer software proliferates, growth in all layers of the software stack should accelerate, even as share shifts toward the platform layer. AI Is Reshaping The Software Value Chain 2019 To 2024 (Actual) 2024 To 2030 (Forecast) Note: “CAGR”: Compound Annual Growth Rate. *SWE-Bench is a benchmark that measures AI agents’ ability to write code autonomously. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 27 ARK’s Big Ideas 2025: AI Agents
    • 28. AI is driving a Cambrian explosion in software. Through 2030, we expect the amount of software deployed per knowledge worker to grow considerably as businesses invest in productivity solutions. Depending on adoption rates, global spend on software could accelerate from an annual rate of 14% over the last ten years to annual rates of 18% to 48%. AI Will Supercharge Knowledge Work Modest Investment Accelerated Investment Rapid Mass Adoption Annual knowledge worker employment growth 6.3% 3.2% 1.3% Percentage of current working time automated by 2030 31% 61% 81% Reduction in productive working hours 0% 8% 20% Productivity surplus created* $22 trillion $57 trillion $117 trillion Value capture of productivity solutions** 10% 10% 10% New software revenue $2.2 trillion $5.7 trillion $11.7 trillion 2030 Software Market Estimate (current size + AI revenue) $3.5 trillion 18% CAGR $7 trillion 33% CAGR $13 trillion 48% CAGR Adoption Scenarios $0 $10 $20 $30 $40 $50 $60 Current Modest Investment Scenario Accelerated Investment Scenario Rapid Mass Adoption Scenario 2030 Enterprise Spending Forecast Knowledge Work Labor Software 28 Note: “CAGR”: Compound Annual Growth Rate. *Traditional production statistics are unlikely to adequately capture the surplus created by AI software. **Value capture rates are likely to vary across scenarios based on competition in the market and other factors. The rate is held constant in these scenarios for simplicity. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: AI Agents
    • 29. A Maturing Global Monetary System With Sound Network Fundamentals And Growing Institutional Adoption ARK’s Big Ideas 2025 Bitcoin Yassine Elmandjra David Puell FORMER DIRECTOR OF DIGITAL ASSETS RESEARCH TRADING ANALYST & ASSOCIATE PORTFOLIO MANAGER, DIGITAL ASSETS Lorenzo Valente DIRECTOR OF DIGITAL ASSETS 29
    • 30. $30,000 $40,000 $50,000 $60,000 $70,000 $80,000 $90,000 $100,000 $110,000 Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 ($) Bitcoin Price Bitcoin Hit An All-Time High In 2024 Bitcoin Undergoes Its Fourth Halving US Launches Spot Bitcoin ETFs Ethereum’s Dencun Upgrade Activates On Mainnet Sam Bankman Fried Sentenced to 25 Years In Prison US Launches Spot Ether ETFs German Government Sells Out Of 50,000 Seized Bitcoin Bitcoin Hash Rate Hits An All-Time High Genesis Settles SEC Lawsuit Over Gemini Earn Hong Kong Debuts Asia’s First Spot Bitcoin And Ether ETFs SEC Issues Robinhood Crypto With Wells Notice Bitcoin Market Cap Dominance Surges Past 55% For The First Time in Three Years Bitcoin Hits A Technical All-Time High ARK’s Big Ideas 2025: Bitcoin As Donald Trump Wins US Presidential Election, Bitcoin Breaks All-Time High SEC Grants Trading And Listing Bitcoin ETF Options Pennsylvania Is First US State To Propose Bitcoin Strategic Reserve David Sacks Appointed US AI And Crypto Czar Source: ARK Investment Management LLC, 2025, based on data from Glassnode and TradingView as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Mt. Gox Begins $9 Billion Payout To Creditors MicroStrategy Announces $42-Billion Raise Over The Next 3 Years 30
    • 31. The Spot Bitcoin ETF Complex Was The Most Successful ETF Launch In History On their first day of trading, the spot bitcoin ETFs attracted over $4 billion of inflows, a record high for ETF launches, surpassing the $1.2 billion that flowed into the gold ETF in its first month in November 2004. Inflows into the spot bitcoin ETFs have dwarfed the firstmonth inflows into each of the ~6,000 ETFs launched over the past 30 years. $0 $20 $40 $60 $80 $100 $120 Jan-24 Feb-24Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 $ Billions Spot Bitcoin ETFs1 AUM2 IBIT FBTC ARKB BITB BTC HODL BRRR BTCO EZBC BTCW 0 10 20 30 40 50 60 70 80 90 100 1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 $ Billions Number Of Months From Launch Spot Bitcoin ETFs And Gold ETF AUM In Year One and Beyond Bitcoin ETF Cumulative Flows Gold ETF Cumulative Flows [1] GBTC and DEFI are not included. [2] “AUM”: assets under management. Source: ARK Investment Management LLC, 2025, based on data from Glassnode and World Gold Council as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 31 ARK’s Big Ideas 2025: Bitcoin
    • 32. Upon Its Fourth Halving, Bitcoin’s Inflation Rate Dropped Below Gold’s Long-Term Supply Growth Growth in the supply of bitcoin “halved” for the fourth time in history, dropping from ~1.8% at an annual rate to ~0.9%. Critical to its design, the halving highlights bitcoin’s predictable monetary policy and its role as a scarce asset. Seven Lines Of Code Created Bitcoin’s Mathematically-Metered Monetary Policy1 0% 20% 40% 60% 80% 100% 0 5 10 15 20 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Percent BTC, Millions Bitcoin Supply Mined Supply (lhs) Percent to Be Mined (rhs) 0% 1% 10% 100% 1000% 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Annualized Percent Inflation Rate: Bitcoin VS. Gold BTC Supply Growth Gold Supply Growth [1] Screenshot as of December 31, 2024, sourced from (github.com/bitcoin/bitcoin/blob/master/src/validation.cpp) lines 1939-1949. Source: ARK Investment Management LLC, 2025, based on data from Glassnode and World Gold Council as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 32 ARK’s Big Ideas 2025: Bitcoin
    • 33. Bitcoin’s Annual Volatility Fell To An All-Time Low, While Its Risk-Adjusted Returns Remained Superior To Most Major Asset Classes Asset Class1 2024 Return Sharpe Ratio2 Sortino Ratio3 5-Year CAGR4 Bitcoin 122.2% 1.4 4.1 67.2% Gold 26.6% 1.7 4.8 10.4% Equities 19.2% 1.3 2.3 11.7% Emerging Markets 8% 0.2 0.4 2.1% Real Estate 3.9% -0.01 -0.01 1.5% 0% Bonds -3.1% -1.1 -1.2 -2.7% 50% 100% 150% 200% 250% 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Annualized Percent, 1-Year, Rolling Bitcoin Volatility []1 Data used for calculations are as follows: S&P Bitcoin PR USD for bitcoin; S&P GSCI Gold TR for gold; MSCI World GR USD for equities; MSCI EM GR USD for emerging markets; S&P Global REIT TR USD for real estate; and Bloomberg Gbl Agg Govt TR USD for bonds. The returns shown are of unmanaged indices and therefore do not reflect the deduction of any fees or expenses that would be applicable to investors in these asset classes, which would lower the returns shown. [2] Sharpe ratio is calculated using average returns over 2024 adjusted for the risk-free rate, divided by the standard deviation. [3] Sortino ratio is calculated using average returns over 2024 adjusted for the risk-free rate, divided by the downside deviation. [4] “CAGR”: Compound Annual Growth Rate. Source: ARK Investment Management LLC, 2025, based on data from Glassnode and Morningstar Direct as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 33 ARK’s Big Ideas 2025: Bitcoin
    • 34. 0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000 $0 $10 $20 $30 $40 $50 $60 $70 $80 Jan-22 Jul-22 Jan-23 Jul-23 Jan-24 Jul-24 BTC $ Millions Miner Revenue Miner Revenue ($) Miner Revenue (BTC) Despite A Steep Decline In Miner Revenue After The Halving, Bitcoin’s Hash Rate Hit An All-Time High Even though the halving cut bitcoin miner revenue by 50%, its hash rate—a proxy for network security—hit a record high. In other words, miners’ long-term conviction in bitcoin remains strong. Miner revenue cut in half 0 100,000,000 200,000,000 300,000,000 400,000,000 500,000,000 600,000,000 700,000,000 800,000,000 900,000,000 2017 2018 2019 2020 2021 2022 2023 2024 TH/s Hash Rate1 [1] In terahashes per second, a unit of measurement used to determine the estimated rate of processing power securing the Bitcoin network. It equals 1 trillion hashes per second. Source: ARK Investment Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 34 ARK’s Big Ideas 2025: Bitcoin
    • 35. Transaction Count Soared, Thanks To Runes Bitcoin’s daily transaction count hit a record high, thanks to the launch of the Runes protocol. Runes facilitates the creation of fungible tokens directly on the Bitcoin blockchain. 0 100,000 200,000 300,000 400,000 500,000 600,000 700,000 800,000 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Change-Adjusted, 14-Day Average Transaction Count 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% May-23 Aug-23 Nov-23 Feb-24 May-24 Aug-24 Nov-24 Distribution (%) Transaction Count Share1 Runes BRC20 Inscriptions Ex BRC20 Regular Transactions [1] “BRC20”: Token standard that enables the minting and transaction of fungible tokens via the Ordinals protocol on the Bitcoin network. Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics and Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 35 ARK’s Big Ideas 2025: Bitcoin
    • 36. Bitcoin Absorbed Major Selling Pressure In 2024 In January 2024, the German government seized 50,000 bitcoin linked to an online piracy group. Six months later, it sold the bitcoin. The market absorbed the supply successfully, and then the price of bitcoin rallied from $53,000 to $68,000. Also, around mid-year, the long-awaited Mt. Gox1 creditor repayment process distributed more than 109,000 bitcoin, removing its largest overhang. $0 $20,000 $40,000 $60,000 $80,000 $100,000 $120,000 0 50,000 100,000 150,000 200,000 250,000 Jan-24 Apr-24 Jul-24 Oct-24 Bitcoin Balance (BTC) Bitcoin Price ($) Major Bitcoin Distributions VS. Bitcoin Price US Government Balance (BTC) Mt. Gox Balance (BTC) German Government Balance (BTC) Bitcoin Price ($) [1] Mt. Gox was a bitcoin exchange that once handled up to 70% of bitcoin’s total trading volume. It declared bankruptcy in 2014 after it was found to have run a fraudulent fractional reserve, incurring the loss of 850k bitcoin from users’ funds at the time. During 2024, a large portion of reclaimed bitcoin was repaid to creditors, from an estimated 141,686 bitcoin at the start of the year to 27,877 by November 13. Data estimates sourced from Glassnode. Source: ARK Investment Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 36 ARK’s Big Ideas 2025: Bitcoin
    • 37. An Increasing Number Of Public Companies Now Hold Bitcoin Public Company (Top 10) Bitcoin Holdings USD Value (Millions) BTC Holdings As Percentage Of Market Cap Microstrategy (MSTR) 446,400 $41,780 58.7% Marathon Digital (MARA) 44,893 $4,201 73.8% Riot Platforms (RIOT) 17,429 $1,631 46.4% Hut 8 Mining (HUT) 10,096 $944 49.4% Tesla (TSLA) 9,720 $909 0.07% Coinbase (COIN) 9,480 $887 1.4% CleanSpark (CLSK) 9,297 $870 32.3% Block (SQ) 8,363 $782 1.4% Bitcoin Group SE (ADE) 3,589 $335 133.5% Boyaa Interactive (BOYAA/0434.HK) 3,183 $297 10.4% $0 $10 $20 $30 $40 $50 $60 0 100 200 300 400 500 600 700 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec $ Billions BTC, Thousands Bitcoin On Public Company Balance Sheets Bitcoin Balance USD Value Source: ARK Investment Management LLC, 2025, based on data from Bitcoin Treasuries and TradingView as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Seventy-four public companies now hold bitcoin on their balance sheets. The value on corporate balance sheets quintupled during the past year, from $11 billion in 2023 to $55 billion. 37 ARK’s Big Ideas 2025: Bitcoin
    • 38. Bitcoin’s Aggregate Cost Basis Hit A New High In 2024 In 2024, Bitcoin’s realized capitalization—or cost basis—grew by 86%, as its average acquisition cost hit a record high at $40,980 per bitcoin, or $811.7 billion in total. $0 $500 $1,000 $1,500 $2,000 $2,500 2019 2020 2021 2022 2023 2024 $ Billions Bitcoin Cost Basis Market Cap ($) Realized Cap ($) Source: ARK Investment Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 38 ARK’s Big Ideas 2025: Bitcoin
    • 39. Bitcoin’s Transaction Velocity1 And Holding Behavior Underscore Its Role As A Store Of Value Bitcoin’s velocity dropped to a 14-year low in 2024, as supply held for three years or more reached an all-time high. 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Percent Bitcoin Supply (%) Supply Last Active 3+ Years Ago 1 10 100 1,000 2011 2012 2013 2014 2015 2016 2017 20182019 2020 2021 2022 2023 2024 Change-Adjusted, Annualized, 30-Day Average Bitcoin's Velocity [1] “Velocity”: a measure of how quickly bitcoin units circulate in the network, calculated by dividing the on-chain bitcoin transaction volume by the outstanding supply (annualized). Source: ARK Investment Management LLC, 2025, based on data from Glassnode as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 39 ARK’s Big Ideas 2025: Bitcoin
    • 40. $0 $200,000 $400,000 $600,000 $800,000 $1,000,000 $1,200,000 $1,400,000 $1,600,000 2030 Price Target Bear Case 2030 Price Target Base Case 2030 Price Target Bull Case ($) 2030 Bitcoin Price Targets (12/31/24 to 12/31/30 CAGR) Institutional Investment Digital Gold Emerging Market Safe Haven Nation-State Treasury Corporate Treasury Bitcoin On-Chain Financial Services Bitcoin Is On Pace To Meet Our 2030 Price Targets Bear Case Base Case Bull Case Institutional Investment TAM: Global Market Portfolio Ex Gold 1% 2.5% 6.5% Digital Gold TAM: Gold Market Cap 20% 40% 60% Emerging Market Safe Haven TAM: EM M2 Monetary Base 0.5% 2.5% 6% Nation-State Treasury TAM: Global Treasury Reserves Ex Gold 0.5% 2.5% 7% Corporate Treasury TAM: Global Cash & Cash Equivalents 1% 2.5% 10% Bitcoin On-Chain Financial Services TAM: L2s, LN, Sidechains, Restaking & WBTC 20% CAGR 40% CAGR 60% CAGR $300,000 CAGR ~21% $710,000 CAGR ~40% $1.5 million CAGR ~58% Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 40 Price Target Assumptions ARK’s Big Ideas 2025: Bitcoin
    • 41. Reshaping The Digital Asset Space Stablecoins Yassine Elmandjra Lorenzo Valente FORMER DIRECTOR OF DIGITAL ASSETS David Puell RESEARCH TRADING ANALYST & ASSOCIATE PORTFOLIO MANAGER, DIGITAL ASSETS DIRECTOR OF DIGITAL ASSETS Frank Downing DIRECTOR OF RESEARCH, NEXT GENERATION INTERNET 41 ARK’s Big Ideas 2025
    • 42. One Of The Fastest-Growing Segments In Digital Assets, Stablecoins Overtook Mastercard And Visa In Transaction Value In 2024 Despite a two-year bear market and a 70%+ decline in market capitalization, stablecoin growth has remained uninterrupted. In 2024, the annualized transaction value of stablecoins hit $15.6 trillion—roughly 119% and 200% that of Visa and Mastercard, respectively. The number of transactions hit 110 million monthly, roughly 0.41% and 0.72% of those processed by Visa and Mastercard, respectively. In other words, the stablecoin value per transaction is much higher than that for Visa and Mastercard. -0.50 0.50 1.50 2.50 3.50 4.50 0 20 40 60 80 100 120 140 Jan-20 May-20 Sep-20 Jan-21 May-21 Sep-21 Jan-22 May-22 Sep-22 Jan-23 May-23 Sep-23 Jan-24 May-24 Sep-24 $ Trillions Millions Stablecoin Transaction Count VS. Total Digital Asset Market Capitalization Stablecoin Transaction Count (lhs) Digital Asset Market Capitalization (rhs) $0 $2 $4 $6 $8 $10 $12 $14 $16 $18 2019 2020 2021 2022 2023 2024 $ Trillions Transaction Value: Stablecoins VS. Traditional Payment Processors Visa Mastercard American Express Stablecoins 42 Note: In the lefthand chart, stablecoins data are for calendar year, while credit card data are for fiscal year. Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external data sources, including Wall Street Zen, CoinGecko, and Visa OnChain Analytics, as of December 31, 2024, which may be provided upon request, For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Stablecoins
    • 43. 0 0.5 1 1.5 2 2.5 3 Dec-23 Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 $ Trillions Stablecoin Volume By Blockchain Solana Tron Ethereum Base Arbitrum BNB Chain Avalanche Optimism Polygon POS TON Celo Stablecoin Volumes Hit Record Highs In December 2024 Solana, Tron, Ethereum, and Base are the leading blockchains that drove stablecoin volume in 2024. December 2024 set new records with a daily volume of $270 billion and a monthly volume of $2.7 trillion, highlighting the sector's rapid growth. 3.8 x Source: ARK Investment Management LLC, 2025., based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Stablecoins 43
    • 44. Ethena Labs Tokenized A Popular Basis Trade And Amassed $6 Billion In 12 Months Innovation in the stablecoin space is thriving, with Ethena Labs as one of the fastest-growing projects. Despite criticism of its novel design, the project has captured significant share of the non-fiatbacked stablecoin market and has become a key competitor in this $200 billion space. Ethena Labs reached $6 billion in total value locked during its first twelve months. By tokenizing a delta-neutral position, Ethena Labs can offer yields as high as 20%-30%, depending on market conditions. The yield could turn negative in a bear market. Having captured 10% of total Ether open interest, Ethena Labs is expanding into Solana and Bitcoin delta-neutral trades. User provides $100 USD worth of ETH Ethena swaps the collateral for $100 worth of yield bearing ETH such as stETH which is deposited at a custodian An Exchange opens a short position worth $100 with the deposit at the custodian as collateral Delta neutral position in which the short position offsets the spot collateral $100 USDE are minted to the user User Note: A delta neutral position in finance is a strategy designed to hedge against price movements in the underlying asset. Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 1 2 3 4 5 44 ARK’s Big Ideas 2025: Stablecoins
    • 45. UDST (Tether) continues to dominate the stablecoin landscape, followed by USDC (Circle). Together, they account for 90% of the total supply. Stablecoins are “multichain” and have penetrated almost every major Layer 1 blockchain. At $203 billion, the stablecoin supply represents ~0.97% of M2* money supply in the US. During December, active stablecoin addresses hit 23 million, an all-time high. As measured by monthly active addresses, Tron is the leading network, favored in emerging markets for its low transaction fees. After A Drawdown In 2023, The Supply Of Stablecoins And The Number Of Active Stablecoin Addresses Hit All-Time Highs In 2024 ARK’s Big Ideas 2025: Stablecoins *M2 is a measure of the US money stock that includes M1 (currency and coins held by the non-bank public, checkable deposits, and travelers’ checks) plus savings deposits (including money market deposit accounts), small time deposits under $100,000, and shares in retail money market mutual funds. Source: ARK Investment Management LLC, 2025, based on data from rwa.xyz (“Stablecoins”) as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 0 5 10 15 20 25 30 Jan-18 Apr-18 Jul-18 Oct-18 Jan-19Apr-19 Jul-19Oct-19 Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21Oct-21Jan-22 Apr-22 Jul-22 Oct-22 Jan-23 Apr-23 Jul-23 Oct-23 Jan-24 Apr-24 Jul-24Oct-24 $ Millions Monthly Active Stablecoin Addresses By Network TRON Binance Smart Chain Solana Ethereum Arbitrum Celo Polygon Optimism Aptos Avalanche NEAR Mantle XDC Base $0 $50 $100 $150 $200 $250 Jan-20 Jan-21 Jan-22 Jan-23 Jan-24 $ Billions Stablecoin Supply USDT USDC USDS USDe FDUSD 45
    • 46. Digital Assets are undergoing “stablecoinization” and “dollarization,” while China and Japan have sold record volumes of US Treasuries, Saudi Arabia has ended its 45-year petrodollar agreement, and BRICS* nations have reduced their reliance on US dollar payments by circumventing the SWIFT network. Historically, bitcoin and ether were the bridges into the digital asset ecosystem. During the past two years, however, stablecoins have dethroned them and now account for 35% to 50% of on-chain transaction volumes. Dollar-pegged stablecoins dominate, accounting for 98%+ of the supply, followed by Gold-backed and Euro-backed stablecoins, with ~1% and ~0.5%, respectively. ARK’s research suggests that the market will expand and include Asian currency-backed stablecoins. As Some Countries Shift Away From The Dollar, Digital Assets Are Moving Toward It 0% 20% 40% 60% 80% 100% Sep-19 Dec-19 Mar-20 Jun-20 Sep-20 Dec-20 Mar-21 Jun-21 Sep-21 Dec-21 Mar-22 Jun-22 Sep-22 Dec-22 Mar-23 Jun-23 Sep-23 Dec-23 Mar-24 Jun-24 Sep-24 Dec-24 Share Settlement Volume Across Different Blockchains Stablecoins L1 L2 Bitcoin Ethereum *BRICS refers to five major emerging economies—Brazil, Russia, India, China, and South Africa. Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 46 ARK’s Big Ideas 2025: Stablecoins
    • 47. Stablecoins Are Attracting Retail Interest, Thanks To The Lower Cost And Higher Efficiency Of Layer 2s Retail investors are flocking to Layer 2s for less expensive and more convenient stablecoin transactions, increasing the market share of Blockchains like Arbitrum, Base, and Optimism. Meanwhile, whales* and institutions continue to operate on Ethereum's base layer. Transactions below $100 dominate activity on Base and Optimism, while those above $100 dominate Ethereum’s base layer. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan-22 Mar-22 May-22 Jul-22 Sep-22 Nov-22 Jan-23 Mar-23 May-23 Jul-23 Sep-23 Nov-23 Jan-24 Mar-24 May-24 Jul-24 Sep-24 Nov-24 Stablecoin Transaction Count Share By Blockchain Ethereum Arbitrum Base Optimism Celo 0% 20% 40% 60% 80% 100% Celo Optimism Base Arbitrum Ethereum Transaction Size Share By Blockchain In 2024 Less than $100 Between $100 and $1K Between $1K and $10K Between $10k and $100K Between $100k and $1M Between $1M and $10M Over $10M *“Whale”: an individual or entity that holds a large amount of a specific cryptocurrency, usually north of 5-10 million dollars depending on the asset. Source: ARK Investment Management LLC, 2025, based on data from Visa OnChain Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 47 ARK’s Big Ideas 2025: Stablecoins
    • 48. Externally Owned Accounts (EOAs)— standard Ethereum addresses used for peer-to-peer (P2P) transactions and storage—account for 60% of USDC’s usage, while centralized exchanges account for 11%, bridges across Layer 2 solutions for 7%, and decentralized exchanges (DEXs) and money markets each for 1.7%. As DeFi usage surges in the next few years, DEXs, Bridges, and Money Markets are likely to retake market share from P2P. While usage in lending/borrowing markets, DEXs, and Bridges fluctuates with market cycles, P2P transactions and storage are more resilient because of product-market fit beyond trading. Peer-to-Peer Transactions And Personal Wallet Storage Dominate Stablecoin Use Cases Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 48 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan-21 May-21 Sep-21 Jan-22 May-22 Sep-22 Jan-23 May-23 Sep-23 Jan-24 May-24 Sep-24 Share Top Addresses Storing USDC EOA CeFi Other Bridge Treasury DEX Lending Yield Farm ARK’s Big Ideas 2025: Stablecoins
    • 49. Four Stablecoin Issuers Dominate The Revenue Generated In Digital Assets With fewer than 200 employees, Tether reported $5.2 billion in profits during the first half of 2024, including USDT, the rest of its products and services, and unrealized gains on its Digital Assets—clearly one of the most capitalefficient businesses in history. Tether (USDT) and Circle (USDC) account for 60% of the revenue generated by the top five networks and applications. Collectively, stablecoins USDT, USDC, DAI/USDS, and USDE generated revenues of $3.35 billion, or $6.7 billion at an annual rate, during the second half of 2024. 0 500 1000 1500 2000 2500 3000 Tether (usdt) Tron Circle (usdc) Ethereum Solana Aerodrome Sky (usds/dai) Ethena labs (usde) Lido Finance Aave Uniswap Pancakeswap $ Millions Top 12 Blockchain Networks And Applications By Revenue* Stablecoin Issuers Others *Second Half of 2024. Source: ARK Investment Management LLC, 2025, based on data from Token Terminal of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. Tether (USDT) Circle (USDC) Sky (USDS/DAI) Ethena Labs (USDE) 49 ARK’s Big Ideas 2025: Stablecoins
    • 50. Tether’s Financial Performance Is Stunning, Both Absolutely And Relatively Tether employs fewer than 200 people compared to 300,000+ each at JP Morgan and Berkshire Hathaway. The only institutions in the S&P Financial Select Sector Index generating more net income than Tether in the first half of 2024 were Berkshire Hathaway, JP Morgan, Bank of America, and Wells Fargo. 1 10 100 1000 10000 100000 1000000 10000000 100000000 0 5 10 15 20 25 30 35 Tether Visa Mastercard Berkshire Hathaway Goldman Sachs JP Morgan Amercan Express Bank of America Progressive S&P Global Wells Fargo $, Log Scale $ Billions Tether VS. Leading Financial Firms* Net Income (lhs) Net Income Per Employee (rhs) *First Half of 2024. Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external data, sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 50 ARK’s Big Ideas 2025: Stablecoins
    • 51. In a world trending toward deglobalization and de-dollarization, stablecoins could drive steady demand for US Treasury securities. Combined, Tether and Circle already ranked as the 20th largest holder of US Treasuries as of December 2024. In populous emerging markets like Brazil, Nigeria, Turkey, Indonesia, and India, individuals and companies are adopting stablecoins as a store of value, means of payment, and crossborder currency. Stablecoins could become one of the most effective ways to export US dollars. Counterbalancing ”De-Dollarization,” Stablecoins Are Increasing The Demand For US Government Debt As Collateral 0 200 400 600 800 1000 1200 Japan China United Kingdom Cayman Islands Luxembourg Canada Belgium France Ireland Switzerland Taiwan Singapore India Brazil Hong Kong Norway Saudi Arabia Korea, South Germany Tether & Circle Mexico $ Billions Top Foreign Holders Of US Treasury Securities As Of September 2024 Source: ARK Investment Management LLC, 2025. This ARK analysis is based on a range of external data sources, including Tether, Circle, and TicData Treasury, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 51 ARK’s Big Ideas 2025: Stablecoins
    • 52. Circle and Tether have been generating billions in revenue from Treasury bills and other securities serving as collateral for their stablecoins. In 2024, however, in response to competition and demand, stablecoins operating outside of the US began to pass a significant portion of their interest income to users. Circle and Tether are unlikely to follow the trend until absolutely necessary. Though still small, yield-bearing stablecoins are the fastest-growing category in the stablecoin market. Stablecoin Issuers Are Transferring Risk-Free Rates To Users Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 52 0 1 2 3 4 5 6 7 8 Mar-23 May-23 Jul-23 Sep-23 Nov-23 Jan-24 Mar-24 May-24 Jul-24 Sep-24 Nov-24 $ Billions Size Of The Yield Bearing Stablecoin Market sUSDe USD0 sDAI USDY deUSD USDB sFRAX USDM USDL rUSD 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan-23 Apr-23 Jul-23 Oct-23 Jan-24 Apr-24 Jul-24 Oct-24 Share Of Interest-Bearing VS. Non Interest Bearing Stablecoins Yield Bearing Non Yield-Bearing ARK’s Big Ideas 2025: Stablecoins
    • 53. 203 1409 0.00% 0.10% 0.20% 0.30% 0.40% 0.50% 0.60% 0.70% 0.80% 0.90% 1.00% $0 $200 $400 $600 $800 $1,000 $1,200 $1,400 $1,600 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 $ Billions Stablecoin Supply VS. Fiat Global M2 2019 To 2030 (2025-2030 Forecast) Stablecoin Supply (lhs) Stablecoins as Percentage of Global M2 Supply (rhs) Stablecoins Could Increase From 0.17% to 0.9% Of Fiat* Global M2 Supply By 2030 Currently $203 billion, or 0.17% of global M2** supply, stablecoins could grow to $1.4 trillion and 0.9%, respectively, by 2030. If so, stablecoins would be the 13th largest currency in circulation, behind Spain and ahead of Netherlands. 38% CAGR 53 Note: “CAGR”: Compound Annual Growth Rate. *Fiat money is a type of currency that is issued by a government and has value because the government declares it to be legal tender. **M2 is a measure of the US money stock that includes M1 (currency and coins held by the non-bank public, checkable deposits, and travelers’ checks) plus savings deposits (including money market deposit accounts), small time deposits under $100,000, and shares in retail money market mutual funds. Source: ARK Investment Management LLC, 2025, based on data from rwa.xyz (“Stablecoins”) as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Stablecoins
    • 54. Enabling Steep Cost Declines And New Use-Cases At The Application Layer Scaling Blockchains Yassine Elmandjra Lorenzo Valente FORMER DIRECTOR OF DIGITAL ASSETS David Puell RESEARCH TRADING ANALYST & ASSOCIATE PORTFOLIO MANAGER, DIGITAL ASSETS DIRECTOR OF DIGITAL ASSETS 54 Frank Downing DIRECTOR OF RESEARCH, NEXT GENERATION INTERNET ARK’s Big Ideas 2025
    • 55. Navigating The Smart Contract Landscape As the digital asset space becomes increasingly complex, smart contracts are driving innovation across a growing number of sectors. The ecosystem is evolving rapidly to meet diverse and dynamic needs—from user-focused applications like gaming and SocialFi, to advanced financial tools like derivatives and structured products, to decentralized infrastructure networks powering wireless connectivity and energy storage. Consumer Applications Gaming Ticketing Smart Devices Trading Apps Decision Markets SocialFi Music Messaging Prediction Markets Metaverse Community And Culture Memecoins NFTs Media And Entertainment Community Accelerators Conference and Events DAOs Generative Art Clubs And Collectibles DeFi RWA Derivatives And Perpetuals NFT Marketplaces and Tooling Trading Bots Launchpads DEX/CEX/HEX Restaking Liquid Staking Asset Creators/ Issuers NFT Lending And Fractionalization Stablecoins Money Markets OTC/ RFQ Derivatives And Spot Aggregators YieldStructured Products Indexes Insurance Risk Management DePIN Mapping Wireless Storage Environment Compute Data Management Energy Services Crypto x AI Compute Marketplaces AI Agents Data Availability Gaming Privacy–ZK-FHE Coordination Networks CoProcessors Model TrainingInference Model Creation Infrastructure And Developer Tooling Wallets DAO Tooling Oracles Developer Tooling And Standards Clients And Nodes Data And Analytics Bridges And CrossChain Messaging Layer 1s RaaS-Shared Sequencers Web 3 Identity Payment And Blink Tools MEV Clients And Tooling Privacy Testing and Frameworks Custodial Services Layer 2s, Layer 3s, and RaaS Onramp Tools Auditing and Bug Bounties Roll-Up Frameworks PoW And PoS Services Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. ARK’s Big Ideas 2025: Scaling Blockchains 55
    • 56. Applications Seeking Lower Fees And Greater Efficiency Deploy Either On Solana’s High Throughput—Layer 1—Or On Ethereum’s Layer 2s Layer 1 Blockchain Layer 2 Blockchain Application Layer User Aggregation Layer Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 56 Ledger Metamask Coinbase Wallet Arbitrum Base Polygon Solana Ethereum Pendle Interest rate Marketplace GMX Perpetual Exchange Polymarket Prediction Market Aave Money Market Morpho Borrow/Lend Infrastructure Jupiter Derivatives Exchange Jito Payment For Order Flow Orca Spot Exchange ARK’s Big Ideas 2025: Scaling Blockchains
    • 57. 0 0.5 1 1.5 2 2.5 Dec-23 Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 $ Ethereum Layer 2s Median Fee Arbitrum Optimism Base zkSync Scroll Blast Linea One of Ethereum's most significant technical upgrades to date, EIP 4844, impacted Layer 2 networks by enabling faster, less-expensive transactions. Ethereum’s EIP 4844 Technical Upgrade Cut Transaction Costs By 10x, Spurring Use Prior to the EIP 4844 upgrade, the average transaction fee on Ethereum Layer 2s was ~$0.50. Now, users pay ~$0.05. EIP 4844 was the first milestone in a roadmap that should enable a 250- to 1250- fold increase in transactions per second (TPS) from ~400 today to 100,000-500,000. More than 200 Layer 2 projects have launched, with others on the way. EIP 4844 Upgrade Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 57 ARK’s Big Ideas 2025: Scaling Blockchains
    • 58. Plummeting Transaction Costs Have Led To A Boom In Layer 2 Activity, Pulling Users Away From Ethereum’s Base Layer Of the Daily Active Addresses transacting on the Ethereum ecosystem, Layer 2s have captured 85% share. The activity on Layer 2s enabled Ethereum to scale daily transactions by 400% from 3 million to 15 million in 2024. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Dec-21 Mar-22 Jun-22 Sep-22 Dec-22 Mar-23 Jun-23 Sep-23 Dec-23 Mar-24 Jun-24 Sep-24 Dec-24 Share Daily Active Addresses: Ethereum VS. Layer 2s Ethereum Top 10 Ethereum L2s 0 5 10 15 20 25 30 Dec-21 Feb-22 Apr-22 Jun-22 Aug-22 Oct-22 Dec-22 Feb-23 Apr-23 Jun-23 Aug-23 Oct-23 Dec-23 Feb-24 Apr-24 Jun-24 Aug-24 Oct-24 Number (Millions) Daily Transactions: Ethereum VS. Layer 2s Ethereum Top 10 Ethereum L2s Combined Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 58 ARK’s Big Ideas 2025: Scaling Blockchains
    • 59. Within one year of launch, Base outpaced all other Ethereum Layer 2 solutions in growth and market share. In 2024, Base captured 46% of active users and generated 63% of fees among Ethereum Layer 2s. With $15 billion in total value locked and more than 300 applications deployed, Base has been contributing significantly to Coinbase’s cash-flows. Base Is The Fastest Growing Ethereum Layer 2 Blockchain Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 59 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Dec-23 Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Share Layer 2 Daily Active Addresses Base Celo Arbitrum Mantle OP Mainnet Linea zkSync Era Blast Scroll StarkNet 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Dec-23 Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Share Layer 2 Fees Base Arbitrum OP Mainnet Linea Mantle StarkNet zkSync Era Blast Scroll Celo ARK’s Big Ideas 2025: Scaling Blockchains
    • 60. Institutions, high-value users, and whales are settling their transactions primarily on the Ethereum base layer. The unit economics of Ethereum’s base layer, as measured by Total Value Locked (TVL) and Decentralized Exchange (DEX) volume per user, are unparalleled. Despite Migration To Layer 2s, Ethereum's Base Layer Still Dominates High-Value Storage And Settlement Note: “Whale”: an individual or entity that holds a large amount of a specific cryptocurrency, usually north of 10 million dollars depending on the asset. Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 60 $0 $20 $40 $60 $80 $100 $120 $140 $160 $180 $200 Dec-23 Jan-24 Feb-24Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 $ Thousands Total Value Locked Per Daily Active Address Ethereum StarkNet Blast Scroll Celo Linea Mantle OP Mainnet zkSync Era ARK’s Big Ideas 2025: Scaling Blockchains
    • 61. When Binance settled with the U.S. Securities and Exchange Commission (SEC) for more than $4 billion and its CEO stepped down in 2024, centralized exchanges (CEXs) lost some share to decentralized exchanges (DEXs). From January to their peaks, spot and derivatives DeFi volume nearly doubled their market share from 8% and 3%, respectively, to 14% and 8%, reaching all-time highs. During the same period, Binance’s market share among centralized exchanges dropped from 62% to 35%. As Decentralized Finance (DeFi) Surged To All-Time Highs, Decentralized Exchanges Challenged Centralized Venues In Both Spot And Derivatives Trading 0% 2% 4% 6% 8% 10% 12% 14% 16% Jan-20 Apr-20 Jul-20 Oct-20 Jan-21 Apr-21 Jul-21 Oct-21Jan-22 Apr-22 Jul-22 Oct-22 Jan-23 Apr-23 Jul-23 Oct-23 Jan-24 Apr-24 Jul-24 Oct-24 DEX Market Share By Trading Volume Spot Trading Volume Derivatives Trading Volume 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan-23 Mar-23 May-23 Jul-23 Sep-23 Nov-23 Jan-24 Mar-24 May-24 Jul-24 Sep-24 Nov-24 Share Centralized Exchange Market Share By Volume Binance Others Source: ARK Investment Management LLC, 2025, based on data from The Block as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 61 ARK’s Big Ideas 2025: Scaling Blockchains
    • 62. Decentralized exchanges (DEXs) like Uniswap, Aerodrome, and Raydium leverage the efficiency of small, agile teams to develop and maintain core protocol infrastructure. With massive efficiency advantages over centralized venues, DEXs have one-tenth the headcount. With ~9,000 employees, Binance leads the centralized exchange space headcount. Smart Contract-Powered Exchanges Are 5-10 Times More Efficient Than Their Centralized Counterparts *The Spot Volume per Employee data are from the month of November 2024. Source: ARK Investment Management LLC, 2025, based on data from Token Terminal, Coin Gecko, and Pitchbook as of November30, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 0 500 1000 1500 2000 2500 OKX Coinbase Crypto.com Binance Bybit Orca Uniswap Aerodrome Pancakeswap Raydium $ Millions Monthly Spot Volume Per Employee* 62 Centralized Exchanges Decentralized Exchanges ARK’s Big Ideas 2025: Scaling Blockchains
    • 63. Liquid staking and restaking have become the preferred methods for generating yield on Ether. Thanks to their yield-bearing properties, liquidity, and accessibility, they now account for 40% of ETH staking. Demand for rehypothecating staked ETH to increase yield has created restaking platforms that now account for ~5.5 million Ether, or 17% of staked ETH. Liquid Staking And Restaking Protocols Are The Preferred Solutions For Staking And Rehypothecating ETH 0 1 2 3 4 5 6 Jun-23 Jul-23 Aug-23 Sep-23 Oct-23 Nov-23 Dec-23 Jan-24 Feb-24 Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 Millions ETH Staked In Leading Restaking Platforms Eigenlayer Symbiotic Karak Note: Staking involves locking up a cryptocurrency in a wallet to support the operations of a blockchain network, typically for Proof-of-Stake (PoS) or similar consensus mechanisms. Liquid staking is an enhancement of traditional staking where users can stake their tokens while still having access to liquidity. Restaking refers to using the same staked assets or their derivatives for multiple purposes or layers of staking. Rehypothecation occurs when assets (such as staked tokens or derivatives) are used as collateral multiple times by various entities. Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 63 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Jan-21 Mar-21 May-21 Jul-21 Sep-21 Nov-21 Jan-22 Mar-22 May-22 Jul-22 Sep-22 Nov-22 Jan-23Mar-23 May-23 Jul-23 Sep-23 Nov-23 Jan-24 Mar-24 May-24 Jul-24 Sep-24 Nov-24 Share Ethereum Staking Methods Liquid Staking CEXs Unidentified Staking Pools Liquid Restaking Solo Stakers ARK’s Big Ideas 2025: Scaling Blockchains
    • 64. Led By Polymarket, Prediction Markets Became The Breakout Consumer Application In 2024 During 2024, Polymarket surpassed three million cumulative users and $1.2 billion in monthly volume, with elections and politics driving 70% of its activity. After a brief post-election pull-back, the number of unique daily users has recovered to 50,000 and volume is trending back up with Sports being the most important category. 0 0.5 1 1.5 2 2.5 3 3.5 0 10 20 30 40 50 60 Jan-24 Feb-24Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 Millions Thousands Number Of Unique Addresses Betting On Polymarket Unique Addresses (lhs) Cumulative Unique Addresses (rhs) $0 $200 $400 $600 $800 $1,000 $1,200 20-May-24 3-Jun-24 17-Jun-24 1-Jul-24 15-Jul-24 29-Jul-24 12-Aug-24 26-Aug-24 9-Sep-24 23-Sep-24 7-Oct-24 21-Oct-24 4-Nov-24 18-Nov-24 2-Dec-24 16-Dec-24 $ Millions Polymarket Volume By Category Elections Politics Unspecified Sports Science Pop Culture Crypto Business Source: ARK Investment Management LLC, 2025, based on data from Dune Analytics as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 64 ARK’s Big Ideas 2025: Scaling Blockchains
    • 65. 0 100 200 300 400 500 600 700 800 900 Oct-23 Nov-23 Dec-23 Jan-24 Feb-24Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Dec-24 Number Transactions Per Second Solana Ethereum + L2s Thanks to Layer 2s and the EIP 4844 upgrade, Ethereum’s throughput has more than doubled to ~200 transactions per second (TPS). Nonetheless, Solana continues to outperform, with an average throughput of ~800TPS. Solana’s success is based on a set of trade-offs, including more expensive hardware requirements and parallel transaction processing. Solana’s new Firedancer client could increase its throughput to hundreds of thousands of TPS. Solana’s Throughput Is Higher Than Ethereum’s, Thanks To High-Performance Design And Strategic Trade-Offs EIP 4844 Upgrade Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 65 ARK’s Big Ideas 2025: Scaling Blockchains
    • 66. After hitting bear market lows at $8 in 2023, Solana has turned around dramatically relative to other Layer 1s. Daily active users, revenue, transaction count, and total value locked (TVL) reached all-time highs or grew by an order of magnitude. Solana is the only Layer 1 that competes with Ethereum and Bitcoin on metrics like daily active addresses and revenue. Thanks To Retail Adoption, Solana Has Gained Share Based On Several Metrics 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Oct-23 Dec-23 Feb-24 Apr-24 Jun-24 Aug-24 Oct-24 Dec-24 Share Daily Active Addresses Solana Bitcoin Ethereum + L2s 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Sep-23 Oct-23 Nov-23 Dec-23 Jan-24 Feb-24Mar-24 Apr-24 May-24 Jun-24 Jul-24 Aug-24 Sep-24 Oct-24 Nov-24 Share Fees Generated Solana Bitcoin Ethereum + L2s Source: ARK Investment Management LLC, 2025, based on data from Artemis Terminal as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 66 ARK’s Big Ideas 2025: Scaling Blockchains
    • 67. Of the 39,139 new crypto developers in 2024, Solana led the way with 7,625, surpassing Ethereum Mainnet. With 4,287 developers in total, Base was sixth overall, surpassing both Arbitrum and Starknet as the leading Layer 2 solution on Ethereum. Solana And Base Are Leading The Way In Developer Adoption And Mindshare 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 Solana Ethereum Other Internet Computer Aptos Base Bitcoin Sui Near Polkadot Polygon Starknet Arbitrum Number Of Net New Crypto Developers In 2024 0 2000 4000 6000 8000 10000 12000 14000 16000 Ethereum Mainnet Base Arbitrum Starknet Optimism Scroll Zksync Polygon Zkevm Number Of Crypto Developers In The Ethereum Ecosystem In 2024 Source: ARK Investment Management LLC, 2025, based on data from Electric Capital 2024 as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security or cryptocurrency. Past performance is not indicative of future results. 67 ARK’s Big Ideas 2025: Scaling Blockchains
    • 68. Robotaxis Transforming Personal Mobility While Lowering Costs And Enhancing Safety Tasha Keeney, CFA Daniel Maguire, ACA DIRECTOR OF INVESTMENT ANALYSIS & INSTITUTIONAL STRATEGIES RESEARCH ANALYST, AUTONOMOUS TECHNOLOGY & ROBOTICS 68 ARK’s Big Ideas 2025
    • 69. Electric Vehicles Continue To Take Share From Internal Combustion Engine Vehicles 4% 2% -2% -5% -16% -1% -10% 4% -1% 57% 67% 70% 18% 33% 113% 59% 30% 12% -40% -20% 0% 20% 40% 60% 80% 100% 120% 2016 2017 2018 2019 2020 2021 2022 2023 2024e Year-Over-Year Percent Change Global Vehicle Sales Growth Internal Combustion Battery Electric 10 20 30 40 50 60 70 80 90 100 2016 2017 2018 2019 2020 2021 2022 2023 2024e Units (Millions) Global Vehicle Sales Internal Combustion Battery Electric Slide by Sam Korus, Director and Akaash TK, Associate, Autonomous Technology & Robotics. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including EVVolumes as of January 3, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 69 ARK’s Big Ideas 2025: Robotaxis
    • 70. The Future Is Autonomous Electric Vehicles Despite the recent slowdown in the growth of electric vehicles, the trend is clear: the internal combustion engine is on its way out. Slide by Sam Korus, Director and Akaash TK, Associate, Autonomous Technology & Robotics. Note: Percentages may not total 100% due to rounding. Source: ARK Investment Management LLC, 2025, based on data from EV-Volumes as of January 3, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 70 92% 87% 81% 77% 72% 4% 5% 6% 7% 8% 1% 2% 4% 5% 7% 3% 6% 10% 12% 13% 50 100 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024e Unit (Millions) Global Vehicle Sales And Share By Segment Internal Combustion Hybrid Plug-in Hybrid Battery ARK’s Big Ideas 2025: Robotaxis
    • 71. After More Than A Decade, Most Automakers Have Dropped Out Of The Autonomy Race. Tesla And Waymo Remain. ARK’s 2014 Estimates Of Autonomous Commercial Launch In 2014, many automakers expected to debut an autonomous vehicle by 2020. Only Waymo delivered, launching its first commercial autonomous rides in 2018. Tesla plans to launch in 2025. Waymo’s launch and Tesla’s internal testing of robotaxis suggest that 2025 could be the standout year in which consumers and businesses agree that the future of transportation is autonomous. Outcome Tesla Volvo Google GM Audi Mercedes Nissan BMW Daimler Continental 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2025 LAUNCH PLANNED LAUNCHED IN 2018 NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH NO EXPECTATION OF LAUNCH Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 71 ARK’s Big Ideas 2025: Robotaxis
    • 72. $2.10 $1.10 $1.10 $1.10 $0.25 1871 1934 1950 2016 2021 Cost Per Mile Of A Personally Owned Vehicle (2024 $) Autonomous Ride-Hail Should Increase Access To Convenient Point-To-Point Transportation Adjusted for inflation, the cost of owning and operating a personal car has not changed since the Model T rolled off the first assembly line more than 100 years ago. ARK estimates that autonomous taxis at scale could cost consumers as little as $0.25 per mile, spurring widespread adoption. Note: Figures are rounded. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Ulvog 2012, Model T Ford Forum 2010, American Automobile Association 1950, and American Automobile Association 2024, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 72 2024 2035 ARK’s Big Ideas 2025: Robotaxis
    • 73. Autonomous Ride-Hail Will Lower The Cost Of Point-To-Point Transportation In early days, robotaxi companies are likely to price rides close to those for human-driven ride-hail. With scale and increased utilization, autonomous technology is likely to drive the cost per mile below that of personal vehicle travel. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 73 $2.00 $1.10 $0.35 $0.25 Human Driven Ride-Hail US, 2024 Personal Car US, 2024 Human Driven Ride-Hail China, 2024 Robotaxi Global, 2035 Cost Per Mile (2024 $) Human-Driven Ride-Hail US, 2024 ARK’s Big Ideas 2025: Robotaxis
    • 74. The US And China Are Blazing The Robotaxi Trail, Followed By The Middle East And Europe 74 United States • Tesla • Waymo • Zoox • Wayve • Pony.ai • WeRide China Saudi Arabia & UAE United Kingdom • Wayve • Baidu Apollo Go • Tesla • Pony.ai • WeRide • Pony.ai • Baidu Apollo Go • WeRide Price competition in China could continue to push local players into more lucrative partnerships abroad. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Robotaxis
    • 75. Robotaxis Already Have Hit 12 Million Rides At An Annual Rate Waymo and Baidu are responsible for most of the autonomous rides in the US and China, respectively, as shown on the left below. With the largest data lake of all players globally, its proprietary data advantage should give Tesla a competitive edge if it launches ride-hail operations as planned in 2025. 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Waymo Baidu Apollo Go Pony.ai Tesla Zoox Million Rides (Annualized) Autonomous Commercial Rides Run Rate 75 ~ 37 ~ 16 ~1 ~ 3,500 * ~0.1 - 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 Waymo Baidu Apollo Go Pony.ai Tesla Zoox Million Miles (Annualized) Autonomous Miles Run Rate Baidu Apollo Go *Tesla Full Self-Driving (FSD) currently is supervised and requires a human behind the wheel. Waymo opened driverless operations to the pubic in October 2020 and now completes ~175,000 weekly rides. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Krafcik 2020, as of January 29, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotaxis
    • 76. Theoretically, Tesla Alone Could Hit ARK’s Robotaxi Industry Forecast In 2030 By 2030, ARK’s research suggests that the number of robotaxis will have scaled to roughly 50 million globally, with Tesla at ~50% market share. Competitors could face barriers to scale, given the intense price competition in China and the lack of alignment between technology providers and traditional automakers. Separately, if it were to allocate 100% of its future vehicle production to robotaxi service—instead of the ~30-40% assumed—Tesla alone could hit ARK’s robotaxi industry forecast in 2030. - 10 20 30 40 50 60 2024 2025 2026 2027 2028 2029 2030 Autonomous Vehicle Fleet (Millions) Global Autonomous Vehicle Adoption Forecast Tesla Robotaxis Other (Waymo, Baidu Apollo Go, etc.) 76 Note: Autonomous adoption curve overestimates figures in initial years. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotaxis
    • 77. Autonomous Electric Vehicles Could Rapidly Shift Vehicle Miles Traveled (VMT) To Electric Autonomous electric vehicles could exceed 50% utilization, pushing a disproportionate number of miles traveled from gas-powered vehicles to electric drivetrains. 0% 5% 10% 15% 20% 25% 30% 35% 40% 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 Share Global Autonomous Electric Miles As A Share Of Total Miles Electric Vehicle Miles Traveled Without Autonomous Forecasted Electric Vehicle Miles Traveled Without Autonomous Autonomous Electric Vehicle Miles Traveled Forecasted Autonomous Electric Vehicle Miles Traveled Slide by Sam Korus, Director, Daniel Maguire, Analyst, and Akaash TK, Associate, Autonomous Technology & Robotics. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 27, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 77 ARK’s Big Ideas 2025: Robotaxis
    • 78. Wright’s Law Suggests That A $15,000 Robotaxi Is Possible $1 $10 $100 $1,000 $10,000 100,000 1,000,000 10,000,000 100,000,000 1,000,000,000 10,000,000,000 $/kWh Cumulative kWh Battery Cost Decline Nickel Cells Nickel Forecast LFP Cells LFP Forecast Battery Cost ($/kWh) $100 $75 $50 Miles/kWh 4 $6,250 $4,688 $3,125 4.5 $5,556 $4,167 $2,778 5 $5,000 $3,750 $2,500 5.5 $4,545 $3,409 $2,273 Assumes a 250-mile range vehicle Tesla has suggested that the Cybercab will have an efficiency of 5.5 miles/kWh. Paired with the battery cost decline anticipated by Wright’s Law, that efficiency suggests that the battery could cost as little as $2,300. Typically, the drivetrain of a vehicle is ~20% of the total vehicle cost, which suggests a Cybercab could be profitable at a price of $15,000 or below. Slide by Sam Korus, Director of Research, Autonomous Technology & Robotics Research. Note: Wright’s Law states that for every cumulative doubling of units produced, costs will fall by a constant percentage. See Winton 2019. “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 78 ARK’s Big Ideas 2025: Robotaxis
    • 79. Tesla’s Cost Per Mile Is ~30-40% Lower Than Waymo’s $- $0.20 $0.40 $0.60 $0.80 $1.00 $1.20 Waymo 5th Gen (Early Launch) Model 3 (Early Launch) Waymo 6th Gen (At Scale) Cybercab (At Scale) Robotaxi Incremental Cost Per Mile Depreciation Operations Support Insurance Energy Maintenance & Tires Parking Tax Its higher cost structure is likely to make Waymo less competitive or less profitable than a Cybercab robotaxi service. Waymo’s dependence on LIDAR and higher cost auto manufacturers are two drags on its competitive positioning. - 50,000 100,000 150,000 200,000 250,000 300,000 350,000 Waymo Hyundai Ioniqs Tesla Cybercabs Number Of Vehicles Vehicles Per $5 Billion In Capex Note: “Capex”/“Capital Expenditure” is a one-off upfront cost required to build or deploy an asset. “Early launch” assumes ten cars per remote operator with utilization rates roughly equivalent to ride-hail today. “At scale” assumes 100 cars per remote operator with an improved autonomous operation utilization rate. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 79 ARK’s Big Ideas 2025: Robotaxis
    • 80. $- $0.05 $0.10 $0.15 $0.20 $0.25 Robotaxi Forecast Cost Per Mile Fees Robotaxi Platform Take Depreciation Parking Energy Maintenance & Tires Deadhead Cost Insurance $- $0.20 $0.40 $0.60 $0.80 $1.00 $1.20 $1.40 $1.60 $1.80 $2.00 Human Ride-Hail Cost Per Mile Uber Take Booking Fee (Ex Insurance) Driver Take Deadhead Cost Maintenance & Tires Depreciation Fuel Insurance Financing Parking Autonomous Ride-Hail Should Reduce Costs, Including Deadhead Miles Autonomous electric travel should reduce costs, including fuel, insurance, and deadhead miles, thanks to higher utilization, increased safety and efficiency, and electric drivetrains. Deadhead miles are the number of miles driven without passengers, which account for 45% of ride-hail miles today. Autonomous vehicles could reduce deadhead miles by removing the human-in-the-loop and harnessing AI tools to predict demand and traffic patterns. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 80 Fees Robotaxi Platform Take ARK’s Big Ideas 2025: Robotaxis
    • 81. Tesla Expects Its Autonomous Software To Outperform Human Drivers On Safety Soon, Tesla expects more than 10,000 miles between critical interventions, or roughly one year’s worth of human driving before the average person would need to touch the wheel in Full Self-Driving (FSD) mode. Later in 2025, Tesla plans to outperform the national accident rate—or ~700,000 miles between critical interventions—paving the way for conversations with regulators. In our view, Tesla could launch a robotaxi service with remote operators before hitting that milestone. Waymo’s commercial robotaxi service has surpassed 400,000 miles between police-reported collisions—impressive performance for a fleet of its size. 10 1,000 10,000 National Average Human Accident Rate (700,000) 700,000 One Year Of Personal Driving (10,000) Waymo Police Reported Collisions (400,000) 1 10 100 1,000 10,000 100,000 1,000,000 16,214 66,736 120,000 190,000 (Log Scale) Compute In H100 Equivalent GPUs Tesla Miles Between Critical Intervention Jan 24 11.4.8.1 Oct 24 12.5.4.1 100x Mar 25 13.4.5 10x Oct 25 13.? 70X Date FSD Version Improvement 81 Note: “Miles Between Critical Intervention” measures the distance traveled before a human driver must take control to avoid an accident. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotaxis
    • 82. Ride-Hail Could Create A ~$10 Trillion Market $0.00 $0.50 $1.00 $1.50 $2.00 $2.50 $3.00 $3.50 $4.00 $4.50 Price Per Mile Ride-Hail Addressable Market* $4B 4.00 3.00 2.00 1.10 0.60 0.50 0.25 $30B $100B $1T: Addressable ridership in Western markets at ~$1 $2.4T: Non-commuting miles in higher income countries priced at ~$0.60 per mile $1.92T: Long tail of demand priced at human driven ridehail prices of $0.35 per mile in lower income countries $5T: Low cost, accessible autonomous travel at $0.25 per mile $134B: Existing addressable market for ride-hail companies charging $2-$4 per mile in Western Markets 1 5 10 0 30 Price Points ($): 0.35 At $0.25 cents per mile, autonomous transportation could serve a much larger population than human-driven ride-hail does today. Meanwhile, based on how consumers value their time, demand could be significant even at higher price points. Miles (Trillions) 82 *$10 trillion is the addressable market, not the revenue we expect in 2030, as we do not expect autonomy to penetrate all addressable miles. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotaxis
    • 83. 500 40 500 1,400 200 1,000 2,600 2,200 33,600 $0 $5,000 $10,000 $15,000 $20,000 $25,000 $30,000 $35,000 $40,000 Revenue (Net) EBIT Enterprise Value Billions Forecasted Revenue, Earnings, And Enterprise Value In 2030 Autonomous Electric Auto Manufacturers Fleet Owners Autonomous Platform Providers Robotaxis Could Generate ~$34 Trillion In Enterprise Value By 2030 Multiples: 13X EBIT 5X EBIT 15X EBIT 83 Note: Numbers are rounded. “EBIT”: Earnings Before Interest and Tax. This slide depicts net revenue in 2030, the portion of gross revenue that the platform operator receives, while the previous slide depicts the total addressable market. Source: ARK Investment Management LLC, 2025This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotaxis
    • 84. Cutting Costs, Revolutionizing Supply Chains, And Reshaping Consumer Behavior Autonomous Logistics Tasha Keeney, CFA Daniel Maguire, ACA DIRECTOR OF INVESTMENT ANALYSIS & INSTITUTIONAL STRATEGIES RESEARCH ANALYST, AUTONOMOUS TECHNOLOGY & ROBOTICS 84 ARK’s Big Ideas 2025
    • 85. $5.40 $0.35 Human-Driven App Delivery Forecasted Drone Delivery Local Small Item Delivery Cost (Per Trip) Autonomous Vehicles That Roll And Fly Could Lower Supply Chain Costs Dramatically $2.40 $0.40 Human-Driven Delivery Rolling Integrated Traffic Robot Delivery Local Batch Delivery Cost (Per Trip) Autonomous vehicles should operate at higher utilization rates than human-in-the-loop systems, creating more costeffective last-mile delivery ecosystems. -83% -94% $0.07 $0.03 $0.00 $0.01 $0.02 $0.03 $0.04 $0.05 $0.06 $0.07 $0.08 Human-Driven Diesel Truck Forecasted Autonomous Electric Truck Truckload Delivery Cost (Per Ton-Mile) -57% Human-Driven Vehicle Forecasted Robot Delivery ARK’s Big Ideas 2025: Autonomous Logistics $6.00 $5.00 $4.00 $3.00 $2.00 $1.00 $0.00 85 Note: We have updated our drone price per mile with our latest assumptions for replacement costs, launching and charging infrastructure, insurance, and labor costs. Fees for drone and robot delivery are shown net of infrastructure costs (outside of charging and launch/land), which we believe either could be borne by the drone or robot delivery operators or shared with logistics or retail partners. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    • 86. Robotrucks And Drones Face More Hurdles To Commercialization Than Robotaxis Robotaxis Robotrucks Drones Technical Hurdles Regulatory Hurdles Customer Partnerships Manufacturing Partnerships Data Collection/ Testing/Validation Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 86 ARK’s Big Ideas 2025: Autonomous Logistics
    • 87. While Tesla has generated billions of miles of customer data for training, Waymo and Baidu have access only to millions of miles. In trucking, autonomous players have taken a route-by-route approach, building their data libraries as they add more routes. For drones, while many flights have occurred in Africa, companies continue to test with partners in the US and Europe, with limited services offered, thus far. In the robotaxi industry, Tesla is vertically integrated while companies like Waymo and Baidu have partnered with Geely and Hyundai/BAIC, respectively. In the trucking space, players like Aurora have partnered both with OEMs* like Paccar and Volvo and with integrators like Continental. Meanwhile, many drone providers manufacture in-house but have not hit mass manufacturing scale. Robotaxis sell rides to end consumers, while robotrucking and drone companies need to forge partnerships with retailers, food delivery providers, and other logistics operators. Robotaxis were first to commercialize, but most robotrucks still require a human operator, with a number of companies pushing to remove the driver this year. Meanwhile, drones have been able to fly autonomously for a number of years: the difficulty lies more in logistics than on the autonomous flight technology. Waymo and Baidu have paved the regulatory path for robotaxis in the US and China, respectively. In the US, state regulators could give way to federal legislation in 2025. Trucking is influenced by lobbyists, while drones are regulated by the FAA, which historically has missed many of its own timelines for developing a framework for drone delivery. Manufacturing Partnerships: Data/Testing/Validation: Technical: Regulation: Customer Partnerships: Hurdles To Commercialization * ”OEM”: Original Equipment Manufacturer. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 87 ARK’s Big Ideas 2025: Autonomous Logistics
    • 88. Company Region Of Operations Customer US Regulatory Approval* Cumulative Number Of Commercial Flights Zipline US Africa Hospital & Consumer Part 135 (Paid) Wing US Australia EU UK Hospital & Consumer Part 135 (Paid) Meituan China UAE Consumer N/A Manna US EU Hospital & Consumer Part 107 (Unpaid) FlyTrex US Consumer Part 107 (Unpaid)^ Matternet US EU UAE Hospital & Consumer Part 107 (Unpaid)^ Amazon US Consumer Part 135 (Paid) 5,000 e 50,000 100,000 280,000 e 400,000 400,000 1,320,000 Logistics Drone Companies Are Beginning To Break Through Regulatory Barriers • Thus far, most drone deliveries have taken place in rural areas outside the US to serve unmet medical needs. • Companies with more real-world data and regulatory approvals to fly beyond-visual-line-of-sight without observers will pull away from the competition. 88 *In the US, paid drone delivery requires Federal Aviation Administration (FAA) Part 135 certification; Part 107 allows operation but prohibits compensation. That said, the FAA is working on updating drone regulation to standardize beyond visual line of sight operations for commercial drones in the US under Part 108. ^FlyTrex and Matternet have Part 135 partners. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 17, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Autonomous Logistics
    • 89. 1,520,000 2,200,000 3,100,000 3,500,000 125,000,000 Gatik Aurora Pony.ai Kodiak Cumulative Number Of Miles Traveled • In the US, commercial autonomous trucking requires a safety driver today. • China leads in generating real-world driving data, but companies in the US should advance considerably once regulators approve the removal of safety drivers in 2025. Autonomous Trucking Companies With The Most Real-World Miles Should Be Best Equipped For Scale Inceptio Technology Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 89 ARK’s Big Ideas 2025: Autonomous Logistics
    • 90. Company Countries Sidewalk Or Road? College Campus Or Beyond? Number Of Deliveries Completed Starship Technologies US GB DE FI EE Sidewalk Mostly Campus Neolix CN Road Beyond Meituan CN Road Beyond Coco* US FI Sidewalk Beyond Avride KR US UAE Sidewalk Beyond Cartken JP US Sidewalk/ Indoor Mostly Campus Serve Robotics US Sidewalk Beyond Robot Delivery Companies With Broad-Based Experience Should Be In A Superior Competitive Position • In the US, most sidewalk deliveries are on college campuses, which do not seem to be productive training ground for urban and residential areas. • In China, road robots are commonplace. 50,000 200,000 e 200,000 500,000 4,000,000 6,000,000 7,000,000 90 *Coco has evolved from human-driven to hybrid robots and is more reliant on humans than are its competitors. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 17, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Autonomous Logistics
    • 91. Autonomous Delivery Revenue Could Approach ~$900 Billion Globally In 2030 280 160 420 Robots & Drones Last Mile Trucks Middle Mile & Over-The-Road Forecasted Autonomous Delivery Revenue ($ Billions, 2030) Parcels Food All Goods Technology-enabled delivery is likely to reshape consumer buying habits, as the cost of transporting goods falls. Our research suggests that robot/drone delivery fees and autonomous trucking revenues could reach ~$440 billion and ~$420 billion, respectively, in 2030. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 17, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 91 ARK’s Big Ideas 2025: Autonomous Logistics
    • 92. The US Needs Low-Cost, AI-Powered Aircraft • The current state of warfare has proven that high-cost aircraft are unsustainable against low-cost internationally made drones. • Meanwhile, AI is improving decision-making and efficiency, reducing the need for human troops. “Smaller And Cheaper AI-controlled Unmanned Jets Are The Way Ahead.” US F-16 ~$30-60 million Chinese DJI ~$2,000-20,000 US SM-2 Missile ~$2 million Ukraine Red Sea: Iranian Shahed 136 ~$20,000 Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. “Replicator will galvanize progress in the too-slow shift of U.S. military innovation to leverage platforms that are small, smart, cheap, and many.” “We need them potentially at very large scale and very quickly.” “manned fighter jets are obsolete in the age of drones...” “I want to turn the Taiwan Strait into an unmanned hellscape…” “If we're shooting down a $50,000 one-way drone with a $3 million missile, that's not a good cost equation." Bill LaPlante, Under Secretary of Defense for Acquisition and Sustainment Frank Kendall, 26th Secretary of the Air Force Doug Bush, Army Acquisition Chief, on buying drones in large quantities Kathleen Hicks, Deputy Secretary of Defense Samuel Paparo, US Indo-Pacific Command Chief Navy Admiral Elon Musk 92 ARK’s Big Ideas 2025: Autonomous Logistics
    • 93. Powering The Artificial Intelligence Revolution Energy Sam Korus Daniel Maguire, ACA DIRECTOR OF RESEARCH AUTONOMOUS TECHNOLOGY & ROBOTICS RESEARCH ANALYST AUTONOMOUS TECHNOLOGY & ROBOTICS Akaash TK RESEARCH ASSOCIATE AUTONOMOUS TECHNOLOGY & ROBOTICS 93 ARK’s Big Ideas 2025
    • 94. 50 100 150 200 250 300 350 400 ARK Forecast: Incremental Power Necessary For AI Hardware 2030 Power Capacity Added In China 2023 GW Incremental Power Necessary For AI Hardware In 2030 VS. Power Capacity That China Added In 2023 Energy Production Is Important To Economic Growth According to our research, China added more power capacity in 2023 than would be necessary to meet our estimate of incremental power demand from AI data centers through 2030. In other words, the buildout rate necessary to accommodate AI power demand is achievable. 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 20052007 2009 2011 2013 2015 2017 2019 2021 2023 2025e 2027e 2029e Billion kWh Global Electricity Generation Global Electricity Generation Forecast Of Global Electricity Generation Extrapolated From 5 Year Trailing CAGR Forecast Of Global Electricity Generation Including That Necessary For AI Hardware Demand Note: “CAGR”: Compound Annual Growth Rate. “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. “GW”: Gigawatt, a unit of power equal to one billion watts. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. Power Capacity Added In China 2023 Incremental Power Necessary For AI Hardware In 2030e 94 ARK’s Big Ideas 2025: Energy
    • 95. Nuclear Energy Regulation Ended The Declines In Electricity Prices 1893 1944 1974 2021 $0.10 $1.00 $10.00 1 10 100 1,000 10,000 100,000 1,000,000 10,000,000 $/kWh 2023 Dollars Cumulative Megawatts US Electricity Prices Informed by Wright’s Law,* ARK’s research indicates that, apart from WWII, US electricity prices fell consistently from the late 1800s until 1974. The U.S. Energy Reorganization Act, enacted in 1974, reversed the drop in nuclear construction costs. $1 $10 $100 $1,000 $10,000 $100,000 1 10 100 1,000 10,000 100,000 1,000,000 Overnight Construction Cost (2023 $/kW) Cumulative Power Deployed (MW) United States & Japan Nuclear Construction Cost Trends (Log Log) U.S. Before Significant Regulatory Change (1953 - 1974) U.S. Post Significant Regulatory Change (1975-2023) Japan Before Significant Regulatory Change (1961-1977) Japan Post Significant Regulatory Change (1978-2009) Japan Nuclear Reactor Regulation Revision (1978) U.S. Energy Reorganization Act (1974) Note: “MW”: Megawatt, a unit of power equal to one million watts. “kW”: Kilowatt, a unit of power equal to one thousand watts. “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. *Wright’s Law states that for every cumulative doubling of units produced, costs will fall by a constant percentage. See Winton 2019. Source: ARK Investment Management LLC, 2025, based on data from Smil 2000 and Cleveland 2023 (left chart) and Lovering et al. 2016 (right chart). For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 95 ARK’s Big Ideas 2025: Energy
    • 96. 20 40 60 80 100 120 140 160 2006200720082009201020112012201320142015201620172018201920202021202220232024 Transcripts Mentioning "Nuclear Energy" "Nuclear Energy" Mentions The Nuclear Tide Is Rising Interest in nuclear energy is surging as companies seek a clean, reliable power source to sustain AI initiatives and meet ambitious net-zero carbon goals. Microsoft in deal for Three Mile Island nuclear power to meet AI demand - Financial Times, Sept 20, 2024 Google orders small modular nuclear reactors for its data centers - Financial Times, Oct 14, 2024 Amazon buys stake in nuclear energy developer in push to power data centers - Financial Times, Oct 16, 2024 Fukushima Nuclear Accident Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Capital IQ Pro, as of December 18, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 96 ARK’s Big Ideas 2025: Energy
    • 97. Supported by clear regulations, nuclear fission could be a cost-competitive, timely solution to increasing power demands globally. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% $2,000 $4,000 $6,000 $8,000 $10,000 $12,000 $14,000 Utility Solar Onshore Wind Natural Gas Combined Cycle Nuclear Capacity Factor $/kW Overnight Capital Cost 2024 Prices (lhs) 1974 Price (2024 $ lhs) Capacity Factor* (rhs) 1 2 3 4 5 Small Hydro Biomass Wind OnshoreWind Offshore Solar PV Nuclear (5MW) Vallecitos (1957) Years Average Commissioning Time For Small Scale Renewable Energy Projects (1-10 MW)^ Wind Onshore Wind Offshore 97 Nuclear Fission Projects Could Deliver Economics Superior To Other Renewable Energy Projects Note: In 1957, the Vallecitos Boiling Water Reactor (VBWR) in Sunol, California, was the first nuclear power plant privately owned and operated to supply electricity to a public utility grid. “MW”: Megawatt, a unit of power equal to one million watts. “kW”: Kilowatt, a unit of power equal to one thousand watts. ^Average across 48 countries between 2005-2022. *Capacity Factor is defined as the ratio of the electrical energy produced by a generating unit for the period of time considered to the electrical energy that could have been produced at continuous full power operation during the same period. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including World Nuclear Association 2024, Gumber et al. 2023, and Lovering et al. 2016, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Energy
    • 98. While Most Small Modular Reactor (SMR) Startups Will Fail, The Survivors Should Thrive SMRs are early stage and face significant execution risks. To triple nuclear capacity by 2050, the US must build large reactors, SMRs, and microreactors. Shorter construction timelines, modular design, and scalability could enable smaller reactors to meet near-term demand. Selecting from more than 30 US designs will be important to reigniting Wright’s Law and lowering electricity costs. 0 1 10 100 1,000 10,000 Net Power In Megawatts Electric (MWe) Size Of US Advanced Reactors (Log Scale) NRC Approved Designs Designs Under Development SMR Micro Gen III Gen III+ Gen IV Large Advanced Note: “NRC”: U.S. Nuclear Regulatory Commission. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Small Modular Reactor (SMR) Global Tracker as of November 14, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 98 Global SMR Projects Early Stage Under Construction Operable Number 0 10 20 30 40 50 60 70 80 90 100 Global SMR Projects Global SMR Projects Under Construction ARK’s Big Ideas 2025: Energy
    • 99. Nuclear Fusion Faces Significant Hurdles To Commercialization JET 1997 Forecasted SPARC 2027 Forecasted ITER 2039 1E+12 1E+13 1E+14 1E+15 1E+16 1E+17 1E+18 1E+19 1E+20 1E+21 1E+22 1950 1970 1990 2010 2030 Triple Product (Log Scale) [Density x Time x Temperature] Fusion Project Progress Toward Scientific Breakeven* Scientific Breakeven Q=1 (D-T Fuel) Nuclear Fission Milestones Demonstration Of SelfSustaining Nuclear Fission 1942 First Nuclear Power Plant To Produce Usable Electricity 1951 15 Years Time From Key Milestone To Commercialization First Commercial Nuclear Power Plant In The US 1946 1957 Creation Of The U.S. Atomic Energy Commission Despite ~75 years of research, nuclear fusion has not generated net energy in a meaningful way. While private fusion companies are promising breakthroughs as early as 2027, commercialization could take another ~15 years, if fission history is any guide. Commercial Breakeven The fusion reaction generates surplus power to cover the cost of building and operating the plant. Engineering Breakeven The fusion reaction generates the same power as the reactor consumes. Scientific Breakeven The fusion reaction generates the same power as injected into the fuel. 99 Note: The ‘Triple Product” displayed on the Y axis above is a single metric that measures nuclear fusion progress by considering density, confinement time, and temperature. It's a very large number, so we use scientific notation, e.g., "1E+21," meaning 1 followed by 21 zeros. “SPARC”: Smallest Possible Affordable Robust Compact reactor. SPARC is a Tokamak design. “ITER”: International Thermonuclear Experimental Reactor. *Excluding Laser-Based Inertial Confinement. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Energy
    • 100. Solar, wind, and stationary energy storage are just getting started. One solution will not meet all demands. Interestingly, Tesla accounts for ~19% of global energy storage today. Nuclear And Other Renewables Can Serve As Viable Solutions 20 40 60 80 100 120 140 160 180 2018 2019 2020 2021 2022 2023 2024 Gigawatt-Hours (GWh) Global Stationary Energy Storage Tesla Energy Storage Non-Tesla Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 4% 8% 12% 16% 1,000 2,000 3,000 4,000 5,000 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024e Share (%) Terawatt-Hours (TWh) Global Renewable Energy Generation Wind (lhs) Solar (lhs) Percent of Total Generation (rhs) 100 Percent Of Total Generation (rhs) ARK’s Big Ideas 2025: Energy
    • 101. $0.05 $0.10 $0.15 $0.20 $0.25 $0.30 30% 50% 70% 90% $/kWh Uptime Capacity Lowest Levelized Cost Of Energy For Desired Uptime Current Cost For Solar And Battery ARK's Estimated Future Cost For Solar And Battery As Battery Costs Decline, Intermittent Energy Sources Become Less Problematic Declining battery costs should make intermittent energy systems economically attractive with 100% uptime. ARK’s modeling suggests that energy-intensive industries will accelerate the pairing of renewables with energy storage. Bitcoin mining is a prime example, showcasing how energy-intensive industries with business models orthogonal to power generation can accelerate the shift to 100% renewable grid integration. Note: “kWh”: Kilowatt-hour, a unit of energy, representing the use or generation of 1 kilowatt of power for 1 hour. “MWh”: Megawatt-hour, a unit of energy, equal to 1,000 kilowatt-hours, representing the use or generation of 1 megawatt of power for 1 hour. Source: ARK Investment Management LLC, 2025, based on data from National Renewable Energy Laboratory as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 101 Solar Installation Power (Megawatts) Battery System (MWh) Solar, Battery, Bitcoin Systems Required To Meet Grid Dispatch Targets At Constant Cost Of Electricity (Size Of Bubble = Bitcoin Mining Power) ARK’s Big Ideas 2025: Energy
    • 102. 0 20 40 60 80 100 120 140 160 180 Gigawatt-Hours (GWh) Status Of US Energy Generation Capacity In Interconnection Queue (2009 VS. 2023) Solar Wind Gas Between 2009 and 2023, US interconnection approval wait times increased ~80%, from 31 months to 56 months. Quicker approvals could lower project costs and accelerate the scaling of energy-intensive projects. Regulatory Reform Should Boost All Renewables 20 40 60 80 2009 2011 2013 20152017 2019 2021 2023 Median Duration (Months) US Interconnection Wait Times For New Energy Generation Solar Wind Gas Combined In Queue/ Active Complete/ Withdrawn Operational In Queue/ Active Withdrawn Complete/ Operational 2009 2023 102 *An interconnection queue is a system used by grid operators to manage and prioritize requests from energy developers to connect new power generation projects to the electricity grid, involving assessments of each project's impact on grid reliability, planning for necessary infrastructure upgrades, and determining cost responsibilities for integrating new plants. Source: ARK Investment Management LLC, 2025, based on data from Lawrence Berkeley National Laboratory as of December 31, 2024. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Energy
    • 103. 103 Decoupling Physical Labor From Output Robotics Sam Korus DIRECTOR OF RESEARCH AUTONOMOUS TECHNOLOGY & ROBOTICS Daniel Maguire, ACA RESEARCH ANALYST AUTONOMOUS TECHNOLOGY & ROBOTICS Akaash TK RESEARCH ASSOCIATE AUTONOMOUS TECHNOLOGY & ROBOTICS ARK’s Big Ideas 2025
    • 104. Robots Should Boost Productivity And Transform Industries 104 2 4 6 8 10 12 14 16 Before Washing Machines After Washing Machines Hours Time To Do Laundry 0 2 4 6 8 10 12 14 Before Assembly Line After Assembly Line Hours Time To Manufacture A Car 10 20 30 40 50 60 70 80 Before Kiva Robots After Kiva Robots Minutes Time From Click To Ship At An Amazon Warehouse -87% -88% -78% Automation’s impact on productivity has transformed industries, with specialized robots as simple as washing machines reducing dramatically the time required for tasks. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Robotics
    • 105. Generalizable Robotics Represent A $26+ Trillion Global Revenue Opportunity 105 ~2.3 Hours Of Unpaid Work Per Day ~2.8 Billion Working Age Population ~$12 Weighted Average Hourly Wage ~$13+ Trillion Opportunity ½ Value Attributed To Free Time VS. Paid Time Household Robotics Manufacturing Robotics ~$13+ Trillion Opportunity (Average Of The Green Cells) ARK Forecasts Global Manufacturing GDP At ~$32 Trillion In 2030 × = Revenue Opportunity* ($ Billions) = × × Productivity Uplift 10% 25% 50% 100% 200% 400% Take Rate 10% 320 800 1,600 3,200 6,400 12,800 20% 640 1,600 3,200 6,400 12,800 25,600 50% 1,600 4,000 8,000 16,000 32,000 64,000 Note: The $12 hourly wage is an estimate of average global wage, not US wage + benefits. *The cells highlighted in green represent what ARK believes to be a reasonable or likely outcome. ^We define “Take Rate” as the percentage of a transactions value that the business keeps. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotics
    • 106. Humanoid Robots Are Debuting Around The World Source: ARK Investment Management LLC, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. 106 Why the human form factor? Key is that a humanoid robot is generalizable. While a wrench can tighten nuts better than a human hand can, it is not a generalizable tool. The human hand is generalizable, particularly in an environment built by and designed for humans. Tesla Optimus Sanctuary AI Phoenix Figure AI Figure 02 Unitree G1 Boston Dynamics Atlas Agility Digit Fourier GR-1 1x Technologies NEO Apptronik Apollo ARK’s Big Ideas 2025: Robotics
    • 107. $- $250,000 $500,000 $750,000 $1,000,000 $1,250,000 0% 50% 100% 150% 200% Upfront Robot Cost Productivity Uplift Net Present Value (NPV) Of A Robot Operating The Same Number Of Hours As A Human Worker Over Ten Years Robot Price For NPV=0 Tesla Optimus Price Expectation At Scale 107 The Adoption Of Humanoid Robots Should Increase As Their Costs Fall And Their Productivity Increases A 100% productivity uplift is equivalent to replacing a human working one shift over ten years. Human Labor Costs On Average In The US Hourly Annually $92,420 $46, including benefits $924,200 ~20,000 10-Year Cumulative Total Hours ~$550,000, not including turnover Net Present Value Of Costs Not Worth Investing In A Humanoid Robot Worth Investing In A Humanoid Robot Note: Per hour salary based on Bureau of Labor Statistics Employer Costs For Employee Compensation press release on September 10, 2024: Average employer costs for all civilian workers = $46.21 per hour; Wages average = $31.80; average benefit costs = $14.41. We assume a 40-hour work week and a 50-week work year. A positive net present value in this calculation suggests that it would be worthwhile to invest in a humanoid robot at that upfront cost and productivity uplift. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotics
    • 108. Jobs Are Bundles Of Tasks 108 What does a professional dishwasher do? $8 $14 $246 $50 $100 $150 $200 $250 $300 Dishwashing Machines Professional Dishwasher Pay Value of Unpaid Time Spent Washing Dishes At Home $ Billions Annual US Dishwashing Costs In a restaurant with a dishwashing machine, washing dishes is one of thirteen tasks required of a “dishwasher.” The home is not much different. The dishwasher is one of the most common “robots” in a house. According to our research, automating activities at home like loading and unloading dishes and cleaning countertops represents a $250 billion revenue opportunity in the US alone. Professional Pay For Dishwasher Value Of Unpaid Time Spent Washing Dishes At Home • Wash dishes, glassware, flatware, pots, and pans, using dishwashers or by hand. • Place clean dishes, utensils, and cooking equipment in storage areas. • Sort and remove trash, placing it in designated pickup areas. • Sweep and scrub floors. • Maintain kitchen work areas, equipment, and utensils in clean and orderly condition. • Clean garbage cans with water or steam. • Receive and store supplies. • Stock supplies like food and utensils in serving stations, cupboards, refrigerators, and salad bars. • Transfer supplies and equipment between storage and work areas, by hand or using hand trucks. • Clean and prepare various foods for cooking or serving. • Prepare and package individual place settings. • Load and unload trucks that deliver or pick up food and supplies. • Set up banquet tables. Tasks: Note: For the value of unpaid time spent washing dishes, we assume, based on a study, that people value their free time at half their wage. From the Bureau of Labor Statistics, we take a median wage of ~$35. The Bureau of Labor Statistics reports that people spend 0.65 hours per day on food prep and cleanup. We assume .22 hours is spent on cleaning. We then multiply by the working age population. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotics
    • 109. Small Businesses Should Benefit Disproportionately From Humanoid Robots Large manufacturing firms are organized by specialized and automated tasks. Specialization and automation give large firms the wherewithal to scale significantly and, in turn, lower labor costs as a share of revenue. Consequently, and somewhat counter-intuitively, large companies typically pay higher wages than small firms because automating specific tasks boosts productivity in large firms more than in small firms. Because generalized automation solutions—those for multiple tasks—have not evolved as quickly as automation solutions for specific tasks, small firms typically have a disproportionate number of automatable-but-not-yetautomated tasks that would benefit from generalizable solutions like humanoid robots. Moreover, a high percentage of manufacturing employees in the US work in small firms. 1 2 3 4 5 6 <500 500-749 750-999 1,000-1,499 1,500-1,999 2,000-2,499 2,500-4,999 5,000-9,999 10,000-19,999 20,000+ Millions Firm Size By Number Of Employees US Manufacturing Employment Sample Small Firm Sample Large Firm Revenue $1,000 $5,000 Labor Share Of Revenue 40% 20% Number Of Employees 2 4 Wage Per Employee $200 ($1,000 x 40%/2) $250 ($5,000 x 20%/4) Tasks Per Employee 4 2 Benefit Of Automating A Single Task $50 ($200 x (1/4 tasks)) $125 ($250 x (1/2 tasks)) Valued Automation Solution Generalizable Single Task Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including U.S. Census Bureau 2023, as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 109 ARK’s Big Ideas 2025: Robotics
    • 110. Automation Turns Non-Market Activity Into Revenue-Generating Activity And Gross Domestic Product (GDP) 110 Family Farmworkers Hired Farmworkers 1 2 3 4 5 6 7 8 Total Breakdown Number (Millions) Farm Worker Jobs “Lost” 1950-2000 Of the 82% who “lost" their jobs between 1950 and 2000, most farm workers were unpaid family members. 2 4 6 8 10 12 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 Number (Millions) Farm Workers Paid Workers Unpaid Family Workers Source: ARK Investment Management LLC, 2025, based on data from U.S. Department of Agriculture 2025 as of January 10, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. ARK’s Big Ideas 2025: Robotics
    • 111. Consumers Continue To “Contribute” Unpaid Labor To The US Economy 111 If US consumers were to value their time at half their hourly wages, food preparation and cleanup would add ~$1 trillion to GDP annually. Food Cost Prep And Cleanup Cost $2 $4 $6 $8 $10 $12 $14 Home Cooked Fast Casual Cost Of A Meal Food Away From Home Food At Home Food Prep And Cleanup 500 1,000 1,500 2,000 2,500 2023 Addressable Market $ Billions Total US Food Cost ~$1 Trillion 200 400 600 800 1,000 1,200 1,400 1,600 1997 2000 2003 2006 2009 20122015 2018 2021 $ Billions US Food Expenditure Food At Home Food Away From Home Food Away From Home Less Delivery Home-Cooked Note: The Food Prep and Cleanup Cost is based on the following: The Bureau of Labor Statistics (BLS) reports that time spent on food prep and cleanup is .65 hours per day. Based on a study, we assume people value their free time at half their hourly wage; for this, we use BLS data of ~$35. We use the population above 18 years old of ~260 million. Source: ARK Investment Management LLC, 2025, based on data from U.S. Department of Agriculture 2025 and Franchisee Resource Center 2021 as of January 10, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Robotics
    • 112. 3D Printing Revenues Could Grow ~40% At An Annual Rate To $180 Billion By 2030 20 40 60 80 100 120 140 160 180 200 2024 2025 2026 2027 2028 2029 2030 $ Billions 3D Printing Revenue Forecast • The 3D printing industry consolidated significantly in 2024, highlighted by Nano Dimension’s acquisition of Desktop Metal and Markforged. • If manufacturers decide to bring the technology in house, as did General Electric, consolidation could continue. • The Trump administration’s efforts to bolster US manufacturing could benefit 3D Printing. • Industries like drone manufacturing and nuclear power, which should scale rapidly, could use 3D printing to accelerate time to market. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. Slide by Tasha Keeney, Director of Investment Analysis & Institutional Strategies 112 ARK’s Big Ideas 2025: Robotics
    • 113. 3D Printing Makes Parts For Nuclear Applications More Efficiently • Westinghouse uses 3D printing for nuclear debris filters. • Those components improved on the measure of resistance to debris by 30%, thanks to the design flexibility offered by 3D printing. • The US Navy engaged Australian additive manufacturing company AML3D to 3D-print tailpiece components for its nuclear submarine program. • AML3D produced those parts in fewer than five weeks, ~15X more quickly than traditional manufacturers. Source: ARK Investment Management LLC, 2025, based on data from Shaikhnag 2024 and Tyrer-Jones 2025 as of January 10, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. Slide by Tasha Keeney, Director of Investment Analysis & Institutional Strategies 113 ARK’s Big Ideas 2025: Robotics
    • 114. Reusable Rockets Sam Korus DIRECTOR OF RESEARCH AUTONOMOUS TECHNOLOGY & ROBOTICS Akaash TK Daniel Maguire, ACA RESEARCH ANALYST AUTONOMOUS TECHNOLOGY & ROBOTICS RESEARCH ASSOCIATE AUTONOMOUS TECHNOLOGY & ROBOTICS 114 ARK’s Big Ideas 2025
    • 115. SpaceX Has A Ten-Year Head Start On Reusable Orbital Rockets “Past experience with launch vehicle reusability has been mixed at best in terms of achieving sustainable cost reductions. So I am a skeptic with regard to many of the claims that have been made for cost reductions.” –David W. Thompson, CEO, Orbital ATK, 2016 “We don’t have any plans to look at reusability at all right now…The reality is that, for the next ten years, other than for Falcon, I don’t think any of us will look at this.” -Phil Slack, President of International Launch Services, 2014 “We have not really changed our assessment over the last couple of years because we have yet to see the other forms of reusability—flyback or propulsive return to Earth—demonstrate economic sustainability on a recurring basis.” –Tory Bruno, CEO, United Launch Alliance, 2020 2015 SpaceX Lands First Falcon 9 2017 SpaceX Launches 1st Flight Proven Booster 2024 No Orbital Reusable Rocket Besides SpaceX 2025+ Blue Origin Rocket Lab China Stoke Space Firefly Relativity 20 40 60 80 100 120 140 160 180 200 2010 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025e Number Falcon Launch Rate New Boosters Reused Boosters Forecast 115 Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Klotz 2020, de Selding 2016, and Faust 2014, as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Reusable Rockets
    • 116. SpaceX Is Refurbishing Rockets In Record Time ARK’s Big Ideas 2025: Reusable Rockets When the Space Shuttle cost ~$1.5 billion per launch, industry experts assumed that an economic reusable rocket would be impossible. SpaceX then flipped the script. According to ARK’s research, the first stage of the Falcon 9 cost <$1 million to refurbish. Now, rocket turnaround time should be proportional to the cost required to refurbish a rocket booster, the key metric in tracking launch cost declines. 50 100 150 200 250 2016 2018 2020 2022 2024 Days Between Reuse Falcon 9 Average Reuse Days Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 252 54 356 27 21 25 14 50 100 150 200 250 300 350 400 Space Shuttle: Average 1972- 2011 Space Shuttle: Fastest 1985 First SpaceX Turnaround Time SpaceX 2021 Fastest Turnaround Time SpaceX 2022 Fastest Turnaround Time SpaceX 2023 Fastest Turnaround Time SpaceX 2024 Fastest Turnaround Time Days Between Reuse Rocket Turnaround Time 116
    • 117. Rocket Reuse And Size Are Driving Rocket Launch Costs Down On A $/kg Basis ARK’s Big Ideas 2025: Reusable Rockets Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. $1 $10 $100 $1,000 $10,000 $100,000 0 1 10 100 1,000 10,000 100,000 1,000,000 $/kg To Low Earth Orbit (2023 $) Cumulative Upmass (Metric Tons) SpaceX Launch Cost Decline Historical 2023 ARK Calculation Forecast 0.1 117
    • 118. SpaceX Continues To Drive Satellite Bandwidth Costs Down In Line With Wright’s Law Anik-F2 (2004) WildBlue-1 (2006) ViaSat-1 (2011) ViaSat-2 (2017) ViaSat-3 (2020) Starlink (2021) (estimated) Starlink (2024) (estimated) Starlink Forecast Starship 2025/2026 Starlink Forecast Rapidly Reusable Starship 1 10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000 1,000,000,000 1 10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000 $/Gbps In Orbit (2024 $) Cumulative Gbps Satellite Bandwidth Cost Decline According to Wright’s Law, satellite bandwidth costs should decline roughly ~45% for every cumulative doubling in gigabits per second (Gbps) in orbit. Since 2004, the cost of satellite bandwidth has dropped 15,000-fold from $300,000,000 to ~$20,000/Gbps. According to ARK’s research, 1 Gbps can serve 200 customers. At a capital cost of ~$1,000/Gbps, SpaceX could recoup its Starship and satellite investment with a one-time charge of $5 per customer. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 118 ARK’s Big Ideas 2025: Reusable Rockets
    • 119. Lower Satellite Launch Costs Should Enable Continuous Global Coverage With Low Latency ARK’s Big Ideas 2025: Reusable Rockets While latency precluded geostationary orbit (GEO) satellites from offering a compelling broadband internet solution, now thousands of low-cost, low earth orbit (LEO) satellites can provide service with low latency, continuous global coverage, and direct-to-mobile device connectivity. GEO ~22,000 miles 700 ms latency LEO ~300 miles <40 ms latency Debris will fall back to earth within ~5 years Debris will fall back to earth within 1,000+ years Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. 119
    • 120. Starship Will Help The Starlink Constellation Achieve Its Potential ARK’s Big Ideas 2025: Reusable Rockets 500 1,000 1,500 2,000 2,500 2023 2024e Metric Tons Spacecraft Upmass SpaceX Other Providers Starship’s payload capacity to LEO is ~5x that of the Falcon 9. While impressive, given the five-year life of its satellites, Starship still will have to fly every 2.3 days to maintain its target constellation of 42,000 Starlink satellites. As of January 2025, SpaceX has a constellation of over 7,000 satellites. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 3.8 (2023 Cadence) 2.7 (2024 Cadence) 2.1 (2025 Goal) 1.0 (Modeled) 0.5 (Modeled) Metric Tons Falcon 9 Annual Upmass Capability By Launch Cadence Upmass necessary to maintain a 42,000 satellite constellation 2.3 (Modeled) Starship Annual Upmass Capability By Launch Cadence Days Between Launches 120
    • 121. Despite A Proliferation In Company Creation, Small Launch Providers May Not Be The Winners ARK’s Big Ideas 2025: Reusable Rockets After capital spending booms, industries tend to consolidate. In the space industry, while launch capability is critical, the larger opportunity could be in the services enabled by low launch costs. Today, only 17 of the 194 small launch providers created since 1996 are operational. Source: ARK Investment Management LLC, 2025, based on data from NewSpace Index as of January 10, 2025. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 5 10 15 20 25 30 19961997 1998 1999 2001 2002 2003 2004 20052006 2008 200920102011 20122013 20142015 20162017 2018 2019 20202021 20222023 2024 Number Of New Small Launch Providers Founded 10 27 10 47 56 27 17 50 100 150 200 Stage Of Small Launch Providers Created Since 1996 Operational Development <5 years Development >5 years Dormant Concept Cancelled Retired 121
    • 122. SpaceX Has Converted The Addressable Market Into Revenue With Increased Starlink Capacity ARK’s Big Ideas 2025: Reusable Rockets Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including SpaceX 2024, as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ~ 20 24 96 1,000 200 400 600 800 1,000 1,200 V1 V1.5 V2 Mini V3 Gbps SpaceX Bandwidth Per Satellite 100 200 300 400 1 2 3 4 5 2021 2022 2023 2024 2025 Terabits Per Second (Tbps) Millions Starlink Subscribers And Bandwidth Capacity Starlink Subscribers (lhs) Cumulative Bandwidth 122
    • 123. Thanks to declining rocket launch costs, telecom operators could provide 100% of cellular subscribers satellite connectivity anywhere in the world. Satellite Connectivity Revenues Could Exceed $130 Billion Per Year 0.001% 0.010% 0.100% 1.000% 10.000% 100.000% 1997 2007 2017 2027 Log Scale Growth Of Satellite Subscribers Relative To Cellular Subscriber RVs 11 Million $1,620 ~$18 Billion 25 Thousand $225,000 ~$6 Billion Commercial Aircraft Fleet 100 Thousand $60,000 ~$6 Billion Cruise Ships, Warships, Commercial Ships Recreational Boats 8.5 Million $1,620 ~$14 Billion Global Households Without Access To Broadband 600 Million $60 ~$40 Billion Addressable Subscribers* Annual Revenue* Direct-To-Device 8 Billion $6 ~$48 Billion Annual Addressable Market* Total: ~$132 Billion *Forecasts. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. SpaceX Direct To Device With T-Mobile, At Scale 123 ARK’s Big Ideas 2025: Reusable Rockets
    • 124. By 2030, Hypersonic Flight Could Be A ~$35 Billion Market, Ready To Scale To ~$350 Billion Longer Term ARK’s Big Ideas 2025: Reusable Rockets According to the US Department of Transportation, leisure travelers are willing to spend 60%-90% of their estimated hourly household income to save one hour.* Compared to conventional flights that can take 28 hours roundtrip, ARK estimates that hypersonic flights could take just 6 hours, saving each traveler ~22 hours. Given the typical cost and potential time savings, ARK’s research suggests that a firstclass passenger should be willing to spend $44,000 roundtrip for a hypersonic flight. If launch costs decline in line with ARK’s expectations, early adopters of hypersonic flight could generate $35 billion revenue by 2030. 110kg $200/kg to LEO $44,000 Starship Price 2 Roundtrip Total Number Of Airline Passengers Worldwide: 6.7 Billion 5% Of Flights Are Long-Haul 5% Of Passengers Are First-Class Number Of Passengers On Long-Haul Flights: ~335 Million Number Of Passengers Flying First-Class: ~16 Million 50% Adoption At Maturity Number Of Passengers Flying Hypersonic: ~8 Million $44,000 Roundtrip Ticket Annual Addressable Market: ~$350 Billion Building Blocks Of Addressable Market Forecast *Forecasts. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of January 10, 2025, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 124
    • 125. Multiomics Operationalizing Data With AI To Transform Diagnostics, Drug Discovery, And Therapies Nemo Despot Marjanovic, PhD Rong Guo, PhD Brett Winton ANALYST, MULTIOMICS RESEARCH ASSOCIATE, MULTIOMICS CHIEF FUTURIST 125 ARK’s Big Ideas 2025
    • 126. Multiomics Performance Should Improve By Orders Of Magnitude By 2030* 10x increase in cancer monitoring market 20x increase in screening productivity 1.6x faster commercialization 4x decrease in cost to develop 5x higher return on R&D 1,000x more performant at same cost 20x more valuable than standard of care drugs 2.4x more valuable than best-in-class precision drugs Generate Data Run Tests Infer And Inform Decisions Run Experiments Infers And Informs Development Address Small Disease Populations Economically Identify Patients And Diseases More Precisely • Intellia Therapeutics • CRISPR Therapeutics • Beam Therapeutics • Prime Medicine • Natera • Guardant Health • Recursion Pharmaceuticals • Absci • Isomorphic Labs Biological Insights Generates Molecular Diagnostics • Adaptive Biotechnologies • Veracyte AI Developed Drugs Cures AI Multiomic Tools 100x decrease in cost to read DNA et al. 1,000x decrease in cost to write DNA et al. • Twist Biosciences • Tempus AI • 10X Genomics • Illumina 126 *The multiomics performance statistics provided on this slide represent ARK’s research-based forecasts for 2030, which may not be realized. The companies listed represent companies that are currently working toward achieving the forecasted results, but the list does not include all companies that may be pursuing the same goals, and which may do so more successfully. The companies listed may or may not be held in ARK portfolios. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment in any of the companies listed was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 127. Biological Information Is Encoded In Three Layers Sequence Layer The static genetic blueprint within DNA, RNA, and proteins. STATIC Illumina Oxford Nanopore PacBio QuantumSi Bruker DeepMind Isomorphic Labs ThermoFisher 10X Genomics Akoya ScaleBio Vizgen Biological Information Measuring Tools* DYNAMIC The linear arrangement of nucleotides in DNA or amino acids in proteins Examples: • DNA sequences store genetic information, determining an organism's primary traits. • Protein sequences define structure and function in cellular processes. Transcription Translation DNA mRNA Protein 3' Tyr Ser Gly Ser 5' 5' 3' 3' Codons 5' Amino acids G A C T A G T C T G Sequence The three-dimensional organization of biomolecules like proteins and nucleic acids Examples • Protein structures determine binding sites and catalytic activity. • Nucleic acid structures affect stability, replication, and transcription. Agglomeration of molecules Examples • Gene regulation: mRNA and protein abundance vary, controlling gene expression. Structure Systems t-SNE2 -20 -10 0 10 20 -30 -30 -20 -10 0 10 20 30 t-SNE1 30 Label Label Label Label Label Label Label Label Single Cell Genomics AlphaFold AI DNA/Protein Sequencing Advances in sequencing, AI, and single-cell technologies enable us to read, model, and predict complexity. Biological information is organized by three interconnected layers that drive life processes. Systems Layer The dynamic abundance of biomolecules responding to cellular and environmental cues. Structure Layer The 3D folding of biomolecules, defining their function. 127 *The companies listed are examples of companies that employ the measuring tools described, but the list does not include all companies that may be utilizing those tools, and which may do so more successfully. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies listed was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 128. $1 $1,000 $1,000,000 $1,000 M $1,000 B 1990 1995 2000 2005 2010 2015 2020 2025 2030 Cost Per 1B Base Pairs (Log Scale) DNA Sequencing Cost Per Whole Human Genome Actual Projected Moore's Law Projection $0 $1 $10 $100 $1,000 $10,000 $100,000 $1 M $10 M $100 M $1 B $10 B 1990 2000 2010 2020 2030 DNA Synthesis Cost Per 1B Base Pairs Actual Projected Moore's Law Projection The Costs Of DNA Sequencing And Synthesis Are Falling Precipitously Illumina’s introduction of high-density flow cell DNA sequencing and synthesis costs are declining faster than Moore's Law* predicted. The cost of reading biology through multiomics tools like DNA sequencing has plummeted by 10 billion-fold (10¹⁰) in its 35-year history. DNA synthesis costs have dropped 100,000-fold (10⁵) and could continue to fall 10 million-fold (10⁷) by 2030. Twist Bioscience’s silicon-based DNA synthesis, combined with miniaturization and parallelization, has reduced costs by increasing efficiency, reducing reagent usage, and scaling production. Pressure From Competitors Illumina's acquisition of Solexa Twist chip miniaturization for DNA storage $1 M $1 K $100 K $10 K $1 K 128 *Moore’s Law predicts that the number of transistors on a chip doubles every 18 months to two years. See Winton 2019. Note: The total price of sequencing the first human genome was estimated at ~2.7B, which included all the costs of developing and improving sequencing tools, actual sequencing, and all analysis. In the above graph, we are calculating only actual sequencing cost per base pair. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including BusinessWire 2024 and Carlson 2022 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 129. The Productivity Of Genomics Analysis Has Skyrocketed Since 2001, the compute time required to analyze a human genome has plummeted from 180 days to 10 minutes. Today we can analyze ~1.4 billion genomes for the cost of analyzing a single genome in 2001. Alongside the plummeting costs of sequencing and synthesis, as noted on the previous slide, genome analysis is commoditizing, driving a revolution in biological research and applications, from personalized medicine to global-scale genomics. ~180 Days ~70 Days ~15 Days ~15 Hours ~60 Min ~34 Min <10 Min 1 10 100 1,000 10,000 100,000 1,000,000 2001 2006 2010 2012 2015 2020 2024 Minutes Compute Time For Analysis Of One Human Genome 1 ~1.4 Billion 1 10 100 1,000 10,000 100,000 1,000,000 10,000,000 100,000,000 1,000,000,000 10,000,000,000 2001 2024 Number Genomes Analyzed For $450,000 Budget 129 Note: 2001 marks the publication of the first draft of the human genome, covering approximately 90% of the genome. In 2003, a near-complete sequence (99.99%) was published, while the fully complete genome was finalized in 2022. Although sequencing the first human genome took about 13 years in total, this slide focuses solely on the time required for the computational analysis, excluding the experimental phase, which was a significantly larger bottleneck. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Saunders et al. 2021, Miller et al. 2015, Gorzynski et al. 2022, and Clifford 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 130. 15 408 2,880 112,000 800,000 1 10 100 1,000 10,000 100,000 1,000,000 Used to Train LLaMa 405b Available in UK biobank Largest publicly available whole genome datasource Generated Annually if we sequenced every newborn in the US Generated annually if we sequenced every newborn globally Devote 1% of global healthcare budget to sequence a billion people* Trillion Tokens (Log) Biological Data Is Big Data LLaMa 405b LLaMa 405b UK Biobank US Newborns Sequenced Per Year Forecast Global Newborns Sequenced Per Year Forecast A Billion People Sequenced Forecast Multiomics tools create volumes of data. The largest publicly-available genomics database, the UK Biobank, includes a half-million patients with 27x more data than the 15 trillion tokens of text powering large AI language models. If newborn sequencing were to become the standard globally, as we expect, data volumes would explode by 1,000x or more. Data Volumes Should Explode As Measurement Costs Decline Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources , including UK Biobank as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 130 ARK’s Big Ideas 2025: Multiomics
    • 131. 0.001 0.01 0.1 1 10 100 1000 $ (Log Scale) Number Of Cells Profiled (Approximations) Cost Per Single Cell 1 2 100 1K 3K 23K 123K 33M 277M 1B 2.5B 4.2B 7B 13B 20B 27B 2.6T 9T 11T 10^5 10^6 10^7 10^8 10^9 2024 2028 $ (Log Scale) Cellular Foundation Model Training $240 Million $2 Million Share of Cost: Compute Sequencing Single Cell Virtual Cell Foundation Models Are Collapsing The Cost Of Single-Cell Genomics Single-cell genomics costs are dropping exponentially, in line with Wright’s Law.* According to ARK’s research, trillion-cell experiments are feasible now that single-cell sequencing costs have dropped to $0.01 per cell. Virtual cell foundation models are likely to become the killer AI application in drug discovery and precision medicine. By 2028, virtual cell costs could drop ~120x. Companies like 10x Genomics and Scale Bio are leading the way. 2024 ~$0.01 2024 ~$0.05 2019 ~$0.2 2014 ~$9 2009 ~$400 120x cost reduction 42% 42% 16% 58% 20% 22% 131 Note: The companies mentioned are examples of companies developing virtual cell models, but the list does not include all companies that may be developing these models, and which may do so more successfully. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be profitable. *Wright’s Law states that for every cumulative doubling of units produced, costs will fall by a constant percentage. See Winton 2019. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Tang et al. 2009, Jovic et al. 2022, Ding et al. 2020, and Rood et al. 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 132. AlphaFold Is Redefining The Prediction Of Molecular Structures AlphaFold has transformed the prediction of protein structures, surpassing traditional methods like X-ray crystallography and Cryo-EM in both productivity* and scope.** Each version of AlphaFold has delivered exponential gains, not only improving efficiency, but also unlocking entirely new structural predictions. AlphaFold1, for example, predicted single-chain proteins (∼60M structures), but AlphaFold3 can predict more than 1 billion structures, including multi-chain complexes, protein-ligand, protein-DNA/RNA interactions, and post-translational modifications. 1.1 289 508 0 100 200 300 400 500 600 Productivity Gains: Alphafold VS. Traditional Methods 60 140 1000 0 200 400 600 800 1000 1200 Millions Structures Predicted AlphaFold 2018 AlphaFold2 2020 AlphaFold3 2024 AlphaFold 2018 AlphaFold2 2020 AlphaFold3 2024 Unit Of Productivity Gains 132 *Productivity gains are defined as the ratio of AlphaFold's Efficiency Metric—accuracy per cost per speed—to the average efficiency of traditional methods. **Scope is calculated as the Efficiency Metric × Number of Structures Predicted. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Jumper et al. 2021, CASP13 2019, and Liu et al. 2024, as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 133. Minimal residual disease (MRD) testing can detect the recurrence of cancer up to 20 months earlier than traditional imaging. Yet, more than eight million suitable cancer patients in the US currently are not reimbursed for MRD testing. Now, at only 10% penetration, MRD testing should become the standard of care for every cancer patient, with multiple tests per year and follow-ups for five years. ARK’s research suggests that MRD testing could produce 700x more data than the largest genomics project to date, the UK Biobank, enabling advanced predictive and prescriptive analytics that could revolutionize how we understand, monitor, and manage cancer. MRD Testing Is Catalyzing The Generation Of Genomics Data 0.3 0.7 8.1 Number (Millions) 0 1 2 3 4 5 6 7 8 9 10 MRD Patient Population 30 4320 21600 0 5000 10000 15000 20000 25000 PetaBytes MRD Testing Data MRD Tested Patients 2024 Current Patient TAM Future Patient TAM UK Biobank MRD TAM Annual MRD TAM Five-Year Forecast 5-fold 144-fold 133 Note: “TAM”: Total addressable market. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Wan et al. 2017, Luskin et al. 2017, and Borfitz 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 134. 100 240 $0 $50 $100 $150 $200 $250 $300 Traditional Cancer Screening Universal MultiCancer Screening Market Potential For Screening Methods ($ Billions) 0.82 3.15 0 0.5 1 1.5 2 2.5 3 3.5 Lives Saved (Millions) $1,500 $300 $0 $200 $400 $600 $800 $1,000 $1,200 $1,400 $1,600 Cost Efficiency Of Screening Methods ($ Billions) Multi-Cancer Screening Could Save Lives And Cut Costs Powered by liquid biopsy technology, multi-cancer screening is a revolutionary approach to early cancer detection. Unlike traditional methods, a single blood test can screen for multiple cancers simultaneously, streamlining screening and expanding its scope. The FDA’s recent approval of Guardant Health’s “Shield” colorectal cancer screening test suggests that the technology is proving its potential. Pending resolution of regulatory and reimbursement paths, universal multi-cancer screening could more than double the Total Addressable Market (TAM) in the US. ~2x increase ~5x Reduction ~4x Increase Traditional Cancer Screening Universal MultiCancer Screening Traditional Cancer Screening Universal MultiCancer Screening 134 Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Philipson et al. 2023, Goddard et al. 2024, and Mannucci and Goel 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 135. Target Hypothesis TargetTo-Hit Hit-ToLead Lead Optimization Preclinical Phase I Phase II Phase III Submission To Launch ~1 year ~1.5 years ~1.5 years ~1 year ~1.5 years ~2.5 years ~2.5 years DISCOVERY DEVELOPMENT ~1.5 years The “virtual cell” is a cutting-edge achievement that integrates artificial intelligence and single-cell genomics. By leveraging multimodal, multi-scale data, virtual cells use foundation models to simulate cellular functions and predict responses to perturbations across various biological states. A transformative leap in drug discovery—particularly target identification and validation—virtual cells enable in-silico experiments to test hypotheses, simulate perturbations, and identify viable targets efficiently and cost-effectively. They also predict how specific mutations or drugs influence cellular pathways, potentially accelerating the pipeline from target discovery to validation and reducing reliance on costly and time-consuming wet-lab experiments. We expect the same timeline reductions as AlphaFold achieved in the prediction of molecular structure. Companies like 10X Genomics, Illumina’s FluentBio, ScaleBio, and ParseBio are providing technologies that build Virtual Cells. Single-Cell Genomics And AI Are Creating Virtual Cells And Transforming Drug Discovery Patient Cell Single Cell Genomics +AI Virtual Cell Academic Studies That Can Take Decades 135 Note: The companies mentioned are examples of companies developing virtual cell models, but those examples do not include all companies that may be developing these models, and which may do so more successfully. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Bunne et al. 2024 and Rood et al. 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics Lead Candidate IND Hit
    • 136. A self-driving lab (SDL) combines not only multiomics tools that read biology at scale, but also automation that enables high-throughput experimentation and large language models (LLMs) that analyze data and design the next set of experiments. The SDL’s integrated approach should enable the transition from automated to autonomous drug discovery and research, potentially accelerating the cycle of discovery and reducing costs/time while increasing efficiency. According to our research, SDLs could save up to two years and hundreds of millions of dollars in the discovery process. Currently achieving ~200x productivity gains, Recursion Pharmaceuticals is pioneering this approach. Self-Driving Labs Are Using Scalable Biology And AI To Revolutionize Drug Discovery Self-Driving Lab Target Hypothesis Lead Optimization Preclinical Phase I Phase II Phase III Submission To Launch ~1 year ~1.5 years ~1.5 years ~1 year ~1.5 years ~2.5 years ~2.5 years DISCOVERY DEVELOPMENT ~1.5 years Lead Candidate IND Hit 136 Academic Studies That Can Take Decades TargetTo-HitHit-ToLead Note: Recursion Pharmaceuticals is the company currently pioneering the use of SDLs, but there may be other companies that utilize SDLs and which may ultimately do so more successfully. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment in the company mentioned was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Recursion 2024, Abolhasani and Kumacheva 2023, and Martin et al. 2023 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 137. Often called the "valley of death" in drug discovery, preclinical studies are prone to high failure rates that stem from predictive models like 2D cell systems and animal testing. Thanks to advances in microfluidics, 3D bioprinting, and organoids, next-generation preclinical models like organs-on-chips are evolving. The new systems are more physiological, scalable, and capable of supporting higher-throughput testing. Integrating Gen 2 models—grown from patient-derived biopsies— with AI-powered analyses of vast datasets could lead to precise, patient-specific predictions that improve drug discovery outcomes dramatically. Large Language Models are beginning to impact and accelerate clinical trials by optimizing regulatory filings, patient selection, trial designs, and data analysis. Organs-On-Chips And AI Are Rescuing Drug Discovery From The Valley Of Death Target Hypothesis Lead Optimization Preclinical Phase I Phase II Phase III Submission To Launch ~1 year ~1.5 years ~1.5 years ~1 year ~1.5 years ~2.5 years ~2.5 years DISCOVERY ~1.5 years DEVELOPMENT Lead Candidate IND Hit Percentage responding(%) 25 50 75 100 ED5 0 TD5 0 Therapeutic effect Toxic effect Therapeutic index Gen 1 2D Human Systems Animal Models Organs On Chip 3D Human Systems Human Biopsy Gen 2 Gen 3 3D Printing/ Microfluidics AI Valley of Drug Discovery Death Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources, including Zhou et al. 2023, Tong et al. 2024, and Thng et al. 2024 as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 137 Academic Studies That Can Take Decades TargetTo-HitHit-ToLead ARK’s Big Ideas 2025: Multiomics
    • 138. Biotech valuations include little-to-nothing for the preclinical and Phase 1 stages of development, but increasingly they are likely to incorporate the higher probability of longer patent lives. AI-driven drug development could reduce time-to-market nearly 40%, from 13 years to 8 years, while reducing total drug costs 4-fold, from $2.4 billion to $0.6 billion. 0 100 200 300 400 500 600 13 12 11 10 9 8 7 6 5 4 3 2 1 11 10 9 8 7 6 5 4 3 2 1 8 7 6 5 4 3 2 1 $ Millions Per Phase Years To Commercialization Drug Development Average CostHit-to-Lead Lead Optimization Pre-clinical Safety Phase I Efficacy Phase II Phase III Confirmation Target-to-Lead AI Is Transforming The Economics Of Drug Development Time Cost Of Money Failed Candidates Successful Candidates Licensing INDUSTRY AVERAGE 2024 • $2.4 billion total • $1.1 billion pre-clinical • 7.5 human trial failures • 13 years time-to-market INITIAL AI-DRIVEN DRUG EFFORTS • $1.5 billion total • $0.6 billion pre-clinical • 4.3 human trial failures • 11 years time-to-market FUTURE AI-DRIVEN DRUG DESIGN • $0.6 billion total • $0.13 billion pre-clinical • 1.4 human trial failures • 8 years time-to-market ARK’s Big Ideas 2025: Multiomics 138 Note: 10% discount rate, 2024 dollars. For initial AI-driven drug efforts, assumes 70+% failure reduction in phase I that moderates to 20% in phase II and phase III. For future AI-driven drug design, assumes demonstrated 70+% reduction in phase I failure rates that moderates to a 50% failure reduction in phase II and 25% failure reduction in phase III. Assumes that pre-clinical efficiency is similar to what Absci has indicated is achievable. Also assumes that licensing timeframe reduces to 12 months from 18 months for future drugs. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon.
    • 139. Reducing Development Timelines Should Increase The Value Of Patents And Commercialization Industry Average Current AI methods accelerate time-to-market by 2 to 3 years, increasing the value of intellectual property by 30% to 50% Emerging AI methods could accelerate time-to-market by 4 to 5 years, increasing the value of intellectual property by 70% to 80% 0.5x 1.0x 1.5x 2.0x 3 4 5 6 7 8 9 10 Patent Life Years Remaining At Commercialization Value Of A Drug At Commercialization By Patent Life Remaining Relative To Industry Average Faster commercialization extends patent-protected revenue, allowing companies to generate higher returns for longer. Current AI efforts shorten development timelines by 2-3 years, increasing a drug's lifetime value by 30-50% compared to the industry average of five years. Emerging AI designs could accelerate time-tomarket by 4-5 years and boost the value by 70-80%. Note: 10% discount rate. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 139 ARK’s Big Ideas 2025: Multiomics
    • 140. Industry Average Current AI methods accelerate time to market by 2 to 3 years, increasing IP value by 30% to 50% Emerging AI methods could accelerate time to market by 4 to 5 years, increasing IP value by 70% to 80% 0.5x 1.0x 1.5x 2.0x 3 4 5 6 7 8 9 10 Patent Life Years Remaining At Commercialization Value Of A Drug At Commercialization by Patent Life Remaining Relative To Industry Average -$1,500 -$500 $500 $1,500 $2,500 $3,500 $4,500 Years From Start Of Development Drug Development Cumulative Cashflow ($ Millions) Drug Development Cost Efficiencies Could Boost Cash Flow Significantly AI Designed Drugs Forecast The value of AI-driven drug development should compound thanks to lower costs, an acceleration in time-to-market, and a longer timeline of patent-protected revenue. Over a 30-year period, the cumulative cashflow for the average AI-designed drug could reach $4 billion, compared to less than $1 billion for traditional drug models. Even before a traditional drug breaks even, an AI-developed drug could generate $2.5 billion in cashflow. Traditional Drug 5 10 15 20 25 30 140 Note: 10% discount rate, 2024 dollars. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 141. Cures for diseases can command much higher prices than traditional treatments. The average price of a cure today exceeds $1 million, nearly 15x the the lifetime prescription cost that would be necessary to manage the disease. Capturing revenue upfront and addressing much of the patient population well before patent expiration suggests that cures could become more valuable—potentially 20x more valuable—than a typical drug and 2.4x more valuable than a chronic prescription that abates the disease. Biological Cures Could Become Quite Valuable $8,000 $75,000 $390,000 $620,000 $1,100,000 $0 $200,000 $400,000 $600,000 $800,000 $1,000,000 $1,200,000 Price Of A Cure For Rare Disease (Per Patient, On Average) Typical Prescription Cost Per Year Lifetime Prescription Cost Lifetime Non-Drug Health Service Costs Saved Life Years Gained Value Of Cure $0 $500 $1,000 $1,500 $2,000 $2,500 $3,000 $3,500 Value Of A Cure For Rare Disease ($ Millions, On Average) Cures avoid competition, with lifetime value captured on the first dose Cures frontload cash in a drug’s commercial lifecycle Typical Prescription Medicine Drug That Displaces Health Service Costs Chronic Prescription That Abates The Disease One Time Cure Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 141 ARK’s Big Ideas 2025: Multiomics
    • 142. 0% 15% 30% 45% Drug Development Return On Invested Capital Cures And Accelerated Development Should Transform The Economics Of Drug Development At inception Preclinical Phase I Phase II Phase III Commercial (300) (200) 0 200 500 1,000 200 600 1,000 1,400 2,000 3,000 300 500 600 800 1,200 1,800 600 700 900 1,000 1,300 1,800 1,500 1,800 2,300 2,500 3,400 4,400 1,800 2,000 2,500 2,700 3,500 4,400 Value Of Assets at Different Clinical Phases ($ Millions) The combination of AI-accelerated drug development and the value of cures for disease could boost returns on research and development (R&D) and improve the value of pre-clinical pipeline assets materially. According to our research, an AI-accelerated cure in the first phase of human testing could be worth more than $2 billion. Traditional assets typically pay back only the cost of capital for Phase 1 pipelines. Representative Companies* Industry Average Beam, Crispr Tx, Editas, Intellia, Prime Abcellera , Absci, Generate Bio, Isomorphic Lab, Lantern Pharma, Recursion, Relation, Roivant Sciences, Xtalpi In Development Traditional Drug, Traditional Development Cure, Traditional Development Traditional Drug, AI Speedup, Traditional Fail Rate Traditional Drug, AI Speedup And AI Success Rate Cure, AI Speedup, Traditional Fail Rate Cure, AI Speedup And AI Success Rate 142 Note: 10% discount rate. *Companies pursuing this strategy for drug development, though this does not imply that every asset the company has at that particular stage is worth that amount or that company return on R&D dollars will meet the modeled value. The information provided should not be used as the basis for any investment decision, and it should not be assumed that an investment any of the companies mentioned was or will be profitable. Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. ARK’s Big Ideas 2025: Multiomics
    • 143. AI-Driven Drug Development And Cures For Disease Could Reverse A Multi-Decade Decay In Pharma/Biotech Returns -10% 0% 10% 20% 30% 40% 50% 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 As AI and other technologies Return On Drug Development R&D impact the drug development pipeline, industry-wide returns should turn around. Meanwhile, many development projects in place today could become sunk costs as more nimble AI drug development platforms beat them to market with cures that eliminate disease. A bifurcation could grow between the returns generated by AI driven, cure-seeking firms and those traditional pharma/biotech companies that do not adapt. 14% 22% 30% 47% Cure, Traditional Development Traditional Drug, AI Speedup, Traditional Fail Rate Traditional Drug, AI Speedup and AI Success Rate Cure, AI Speedup, Traditional Fail Rate Cure AI Speedup And AI Success Rate Source: ARK Investment Management LLC, 2025. This ARK analysis draws on a range of external data sources as of December 31, 2024, which may be provided upon request. For informational purposes only and should not be considered investment advice or a recommendation to buy, sell, or hold any particular security. Past performance is not indicative of future results. Forecasts are inherently limited and cannot be relied upon. 143 ARK’s Big Ideas 2025: Multiomics
    • 144. Works Cited 144 ARK’s Big Ideas 2025
    • 145. Abolhasani, M. and E. Kumacheva. 2023. “The Rise of Self-driving Labs in Chemical and Materials Sciences.” Nature Synthesis. American Automobile Association. 2050. “Your Driving Costs.” American Automobile Association. 2024. “How Much Does it Really Cost to Own a New Car?” Artemis Terminal. 2024. Bitcoin Treasuries. 2024. Borfitz, D. 2024. “UK Biobank Pivoting to Platform-Only Model For Big Data Sharing.” Bio IT World. Bunne, C. et al. 2024. “How to Build the Virtual Cell with Artificial Intelligence: Priorities and opportunities.” Cell. BusinessWire. 2024. “Ultima Announces UG 100™ and Reveals Disruptive Cost and Accuracy Profile to Enable the Era of the $100 Genome and Beyond.” Calma, J. 2024. “Google Orders Small Nuclear Reactors for its Data Centers.” Financial Times. Carlson, R. 2022. “DNA Cost and Productivity Data, aka ‘Carlson Curves.’” Synthesis. Capital IQ Pro. 2024. CASP13. 2019. “13th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction.” Circle. 2024. Cleveland, C. 2023. “United States Electricity History in Four Charts.” Boston University Institute for Global sustainability. Clifford, J. 2024. “Boost Multi-Omics Analysis with GPU-Acceleration and Generative AI.” Nvidia Developer. CoinGecko. 2024. Dediu, H. 2017. “Share of United States Households Using Specific Technologies.” Our World in Data. DeLong, B. 1998. "Estimating World GDP, One Million B.C. — Present." Brad De Long’s homepage. Department of Economics, University of California, Berkeley. de Selding, P.B. “Orbital ATK Believes in Satellite Servicing, but Not Rocket Reusability.” Space News. Ding, J. et al. 2020. “Systematic Comparison of Single-cell and Single-nucleus RNA-sequencing methods.” Nature Biotechnology. Dune Analytics. 2024. Electric Capital. 2024. “Developer Report 2023.” EV Volumes. 2025. Foust, J. 2014. “Reusability and Other Issues Facing the Launch Industry.” The Space Review. Franchisee Resource Center. 2021. “Fast Casual Industry Analysis 2020 - Cost & Trends.” Franchise Help. Glassnode. 2024. Goddard, K. et al. 2024. “Estimation of Cancer Deaths Averted From Prevention, Screening, and Treatment Efforts, 1975-2020.” JAMA Oncology. Goddard, T. 2024. “AlphaFold Performance: Molecule Size, Speed, Memory, and GPU.” UCSF Resource for Biocomputing, Visualization, and Informatics. Gorzynski, J. et al. 2022. “Ultrarapid Nanopore Genome Sequencing in a Critical Care Setting.” The New England Journal of Medicine. Gumber, A. et al. 2024. “A Global Analysis of Renewable Energy Project Commissioning Timelines.” Applied Energy. Hartig, K. 1998. “TUNING IN: Communications Technologies Historically Have Had Broad Appeal for Consumers,” The Wall Street Journal. Jiminez, C. et al. 2024. “SWE-bench: Can Language Models Resolve Real-World GitHub Issues?” arXiv. Jovic, D. et al. 2022. “Single-cell RNA Sequencing Technologies and Applications: A Brief Overview.” Clinical and Translational Medicine. Jumper, J. et al. 2021. “Highly Accurate Protein Structure Prediction with AlphaFold.” Nature. Jurvetson, Steve 2024. “The Moore’s Law Update.” X. Klotz, I. 2020. “Economics Of Rocket Reuse Still Up In The Air.” Aviation Week. Kurzweil, R. 2006. “The Singularity Is Near: When Humans Transcend Biology.” Duckworth. Lawrence Berkeley National Laboratory. 2024. “Energy Markets & Policy.” Liu, L. et al. 2024. “Technical Report of HelixFold3 for Biomolecular Structure Prediction.” arXiv. Lovering, J. et al. 2016. “Historical Construction Costs of Global Nuclear Power Reactors.” Energy Policy. Luskin, M. et al. 2018. “Targeting Minimal Residual Disease: A Path to Cure?” Nature Reviews Cancer. Mannucci, A. and A. Goel. 2024. “Stool and Blood Biomarkers for Colorectal Cancer Management: An Update on Screening and Disease monitoring.” Molecular Cancer. ARK’s Big Ideas 2025: Works Cited
    • 146. Martin, H.G. et al. 2023. “Perspectives for Self-driving Labs in Synthetic Biology.” Current Opinion in Biotechnology. McCormick, M. and J. Smith. 2025. “Microsoft in Deal for Three Mile Island Nuclear Power to Meet AI Demand.” Financial Times. Miller, N.A. et al. 2015. “A 26-hour System of Highly Sensitive Whole Genome Sequencing for Emergency Management of Genetic Diseases.” Genome Medicine. Model T Ford Forum. 2010. “How Many Miles Can a T Drive Down, Before it Needs a Rebuild?” Morningstar Direct. 2024. National Renewable Energy Laboratory (NREL). 2024. “Solar Resource Data.” NewSpace Index. 2025. “Small Satellite Launchers.” OpenAI. 2024. OpenPhilanthropy. 2025. Philipson, T. et al. 2023. “The Aggregate Value of Cancer Screenings in the United States: Full Potential Value and Value Considering Adherence.” BMC Health Services Research. Pitchbook. 2024. Recursion. 2024. “Recursion Unveils LOWE Drug Discovery Software at the J.P. Morgan Healthcare Conference.” Rood, J. et al. 2024. “Toward a Foundation Model of Causal Cell and Tissue Biology with a Perturbation Cell and Tissue Atlas.” Cell. Rwa.xyz. 2024 Saunders, C.J. et al. 2021. “Rapid Whole-genome Sequencing for Genetic Disease Diagnosis in Neonatal Intensive Care Units.” Science and Translational Medicine. Shaikhnag, A. 2024. “Westinghouse Facilitates Improved Nuclear Power with Its 3D Printed Nozzles.” 3D printing Industry. Sidoti, O. et al. 2024. “Mobile Fact Sheet.” Pew Research Center. Small Modular Reactor (SMR) Global Tracker. 2024. Smil, V. 2000. “Energy in the Twentieth Century: Resources, Conversions, Costs, Uses, and Consequences.” Annual Review of Energy Environment. Smyth, J. 2024 “Amazon Buys Stake in Nuclear Energy Developer in Push to Power Data Centres.” Financial Times. SpaceX. 2024. “STARSHIPS ARE MEANT TO FLY—2024 PROGRESS REPORT.” Tang, F. et al. 2009. “mRNA-Seq Whole-transcriptome Analysis of a Single Cell.” Nature Methods. Tether. 2024. The Block. 2024. Thng, D. et al. 2024. “A Functional Personalised Oncology Approach Against Metastatic Colorectal Cancer in Matched Patient Derived Organoids.” NPJ Precision Oncology. TicData Treasury. 2024. “Table 5: Major Foreign Holders of Treasury Securities.” Token Terminal. 2024. Tong, L. et al. 2024. “Patient-derived Organoids in Precision Cancer Medicine.” MED. TradingView. 2024. Tyrer-Jones, A. 2025. “3D Printer Nuclear Submarine Components from AML3D Delivered to US Navy.” 3D PRINTING INDUSTRY. Ulvog, J. 2012. “Travel Cost by Stagecoach in 1870s – Part One.” Outrun Change. UK Biobank. 2024. U.S. Department of Agriculture. 2025. “Food Expenditure Series.” U.S. Department of Agriculture. 2025. “Farm Labor.” U.S. CENSUS BUREAU. 2023. “2021 SUSB Annual Data Tables by Establishment Industry.” Visa OnChain Analytics. 2024. Wallstreet Zen. 2024. Wan, J. et al. 2017. “Liquid Biopsies Come of Age: Towards Implementation of Circulating Tumour DNA.” Nature Reviews Cancer. Winton, B. 2019. “Moore’s Law Isn’t Dead: It’s Wrong—Long Live Wright’s Law.” ARK Investment Management LLC. World Nuclear Association. 2024. “Small Modular Reactor Global Tracker.” Worldpay Inc. 2018. “2018 Global Payments Report.” Worldpay Inc. 2019. “2019 Global Payments Report.” Worldpay Inc. 2020. “2020 Global Payments Report.” Worldpay Inc. 2021. “2021 Global Payments Report.” Worldpay Inc. 2022. “2022 Global Payments Report.” ARK’s Big Ideas 2025: Works Cited
    • 147. Worldpay Inc. 2023. “2023 Global Payments Report.” World Gold Council. 2024. “Gold ETFs, holdings and flows.” World Nuclear Association. 2024. “United State of America Vallecitos Permanent Shutdown.” Zhou, Z. et al. 2023. “Harnessing 3D In Vitro Systems to Model Immune Responses to Solid Tumours: A Step Towards Improving and Creating Personalized Immunotherapies.” Nature Reviews Immunology. ZipRecruiter. 2024. ARK’s Big Ideas 2025: Works Cited
    • 148. ARK’s Big Ideas 2025: Disclosure For more research on disruptive innovation visit www.ark-invest.com ©2021-2026, ARK Investment Management LLC. No part of this material may be reproduced in any form, or referred to in any other publication, without the express written permission of ARK Investment Management LLC (“ARK”). Please note, companies that ARK believes are capitalizing on disruptive innovation and developing technologies to displace older technologies or create new markets may not in fact do so and/or may face political or legal attacks from competitors, industry groups, or local and national governments. ARK aims to educate investors and to size the potential opportunity of Disruptive Innovation, noting that risks and uncertainties may impact our projections and research models. Investors should use the content presented for informational purposes only, and be aware of market risk, disruptive innovation risk, regulatory risk, and risks related to Deep Learning, Digital Wallets, Battery Technology, Autonomous Technologies, Drones, DNA Sequencing, CRISPR, Robotics, 3D Printing, Bitcoin, Blockchain Technology, etc. Cryptocurrency Risk. Cryptocurrencies (also referred to as “virtual currencies” and “digital currencies”) are digital assets designed to act as a medium of exchange. Cryptocurrency is an emerging asset class. There are thousands of cryptocurrencies, the most well-known of which is bitcoin. Cryptocurrency generally operates without central authority (such as a bank) and is not backed by any government. Cryptocurrency is not legal tender. Federal, state and/or foreign governments may restrict the use and exchange of cryptocurrency, and regulation in the U.S. is still developing. The market price of bitcoin and other cryptocurrencies have been subject to extreme fluctuations. Similar to fiat currencies (i.e., a currency that is backed by a central bank or a national, supra-national or quasi-national organization), cryptocurrencies are susceptible to theft, loss and destruction. Cryptocurrency exchanges and other trading venues on which cryptocurrencies trade are relatively new and, in most cases, largely unregulated and may therefore be more exposed to fraud and failure than established, regulated exchanges for securities, derivatives and other currencies. Cryptocurrency exchanges may stop operating or permanently shut down due to fraud, technical glitches, hackers or malware, which may also affect the price of cryptocurrencies. Cryptocurrency Tax Risk. Many significant aspects of the U.S. federal income tax treatment of investments in bitcoin and other cryptocurrencies are uncertain and still evolving. The content of this presentation is for informational purposes only and is subject to change without notice. This presentation does not constitute, either explicitly or implicitly, any provision of services or products by ARK and investors are encouraged to consult counsel and/or other investment professionals as to whether a particular investment management service is suitable for their investment needs. All statements made regarding companies or securities are strictly beliefs and points of view held by ARK and are not endorsements by ARK of any company or security or recommendations by ARK to buy, sell or hold any security. Historical results are not indications of future results. Certain of the statements contained in this presentation may be statements of future expectations and other forward-looking statements that are based on ARK's current views and assumptions and involve known and unknown risks and uncertainties that could cause actual results, performance or events to differ materially from those expressed or implied in such statements. The matters discussed in this presentation may also involve risks and uncertainties described from time to time in ARK's filings with the U.S. Securities and Exchange Commission. ARK assumes no obligation to update any forward-looking information contained in this presentation. Certain information was obtained from sources that ARK believes to be reliable; however, ARK does not guarantee the accuracy or completeness of any information obtained from any third party. ARK and its clients as well as its related persons may (but do not necessarily) have financial interests in securities or issuers that are discussed. ARK Investment Management LLC


    • Previous
    • Next
    • f Fullscreen
    • esc Exit Fullscreen