If you follow global technology investment, it is nearly impossible to avoid one name—Cathie Wood, whom Chinese investors more familiarly call "Sister Wood." Over the past decade, she and the ARK Invest she founded have been doing something not particularly popular on Wall Street: ignoring short-term noise and betting on long-term, extreme, non-linear technological change. ARK's annual research report, "Big Ideas," has been published for ten consecutive years. It is not merely an industry outlook but more like a "technology map for the next decade." You may disagree with its conclusions, but it is difficult to ignore the questions it raises. This year's "ARK Big Ideas 2026" has a very striking overarching title: The Great Acceleration.
The report focuses on 13 major innovation areas, with the core thesis that five innovation platforms centered on artificial intelligence are accelerating their convergence, which will trigger a step-change in global economic growth by the end of this decade. The real GDP growth rate is projected to reach 7.3% by 2030, 4 percentage points higher than the International Monetary Fund's forecast of 3.1%. The most important assertion in the report is that AI is not just another significant technological advancement but a "Central Dynamo" that is simultaneously driving the acceleration of multiple technology curves. For decades, technological innovation mostly followed a linear structure: one technology → one industry → one capital cycle. ARK believes this paradigm is now obsolete. In the current phase, technologies are no longer parallel but are highly coupled and mutually unlocking:
AI's computational demands are driving revolutions in next-generation cloud computing, energy storage, and data centers; blockchain and digital wallets provide a trusted settlement and execution layer for AI Agents; robotics and autonomous driving push AI from the "digital world" into the "physical world"; multi-omics and programmable biology supply high-dimensional life data to AI, in turn accelerating model capabilities.
ARK uses a metric to describe this state: Convergence Network Strength. By 2025, this indicator increased by 35% year-over-year—signaling that mutual catalysis between different technologies is noticeably speeding up. This is also why ARK labels 2026 as: The Great Acceleration.
ARK research shows that reusable rockets launching AI chips into orbit, multi-omics data driving precision therapy development, and smart contracts enabling AI agents to coordinate real-world resources—these seemingly independent innovations are forming unprecedented synergistic effects. The importance of robotics technology as a catalyst reached an inflection point in 2025, while energy storage and distributed energy systems have become key enablers for next-generation cloud infrastructure development. The report states that the direct impacts of this technological revolution are:
The market share of innovation-related assets is projected to grow from approximately 20% in 2025 to about 50% by 2030, with market capitalization potentially expanding from the current roughly $5 trillion to around $28 trillion. Data center systems investment is forecast to grow from approximately $500 billion in 2025 to about $1.4 trillion by 2030, representing a compound annual growth rate of 30%. Commercialization processes in areas like autonomous ride-hailing, AI drug development, and domestic humanoid robots are accelerating, with some sectors already entering scaled deployment phases.
However, ARK also clearly points out that not all eye-catching technologies are disruptive. Using quantum computing as an example, the report suggests that even under the most aggressive development pace, its practical utility for cryptography decryption might not materialize until the 2040s. Truly disruptive technologies must meet criteria such as急剧的成本下降, unlocking compelling unit economics across multiple industries, and serving as a platform for other technological innovations.
AI Leads the "Great Acceleration" Era The report states that ARK has named this wave of technological change "The Great Acceleration," positing that the interdependence among five major innovation platforms—AI, public blockchain, robotics, energy storage, and multi-omics—is strengthening, where performance gains in one platform unlock new capabilities in another. One of the most striking examples in the report is the combination of reusable rockets and AI computing power. Neural networks' demand for next-generation cloud computing capacity is facing terrestrial scaling limits, and reusable rockets could provide a solution. At competitive costs, space-based AI computing could provide cloud capabilities unconstrained by terrestrial power and cooling limitations. ARK's analysis indicates that AI chip growth could increase demand for reusable rockets by approximately 60 times compared to existing models. Based on projected launch costs, space-based computing costs might be 25% lower than terrestrial computing.
According to the report, this technological convergence is spawning an unprecedented investment cycle. ARK research suggests that capital investment alone could contribute 1.9 percentage points to annualized real GDP growth during this decade. The new capital base—autonomous taxi networks, next-generation data centers, and corporate investment in AI agents—should enhance returns on invested capital. As other innovations begin to impact the growth trajectory, realized growth could exceed consensus expectations by more than 4 percentage points annually. Historically, technological paradigm shifts have repeatedly caused structural changes in GDP growth rates. ARK data shows that the global real GDP growth rate gradually increased from 0.037% around 100,000 BC, through stages like the Agricultural and Industrial Revolutions, to the current approximately 3%. This AI-centric technological revolution could potentially push this growth rate above 7%.
Surge in AI Infrastructure Investment The growth rate of data center systems investment is accelerating. Since the launch of ChatGPT, the annualized growth rate for such investment jumped from a previous 5% to 29%.
In 2025, global data center systems investment reached approximately $500 billion, nearly 2.5 times the average level from 2012 to 2023. ARK predicts this investment scale could grow to about $1.4 trillion by 2030.
The core driver behind this investment surge is the explosive growth in AI demand. Inference costs have dropped by over 99% in the past year, prompting exponential growth in AI usage by developers, enterprises, and consumers. For example, on the OpenRouter platform, computational demand for large language models has grown approximately 25-fold since December 2024. However, compared to the dot-com bubble era, current valuations in the tech sector are far more rational. Although capital expenditure as a percentage of GDP for the Information Technology and Communication Services sectors has reached its highest level since 1998, the price-to-earnings ratios of tech stocks are significantly lower than the peaks seen during the dot-com bubble. The average P/E ratio for the six companies—NVIDIA, Alphabet (Google's parent), Apple, Amazon, Meta, and Microsoft—is only a fraction of their historical highs, indicating the current investment boom is more based on practical application demand than speculative frenzy. The competitive landscape is also shifting. NVIDIA's early investments in AI chip design, software, and networking gave it an 85% share in GPU sales with gross margins as high as 75%. However, competitors like AMD and Google have caught up in certain areas, particularly in small language model inference. ARK data shows that AMD's MI355X can process approximately 38 million tokens per dollar of Total Cost of Ownership (TCO) for small models, surpassing NVIDIA's B200. Nevertheless, NVIDIA's Grace Blackwell rack-scale system still leads in large model inference, powering the most advanced foundation models. AI Consumer Operating Systems Reshaping Business Models AI models are converging to form new consumer operating systems, fundamentally altering how people interact with the digital world. Consumer adoption of AI is far outpacing the initial spread of the internet—AI chatbots reached a penetration rate of about 25% among smartphone users within 7 years of launch, whereas the internet took much longer to achieve the same penetration among PC users. This shift is compressing the shopping funnel. Completing a purchase took about 1 hour in the pre-internet era, shortened to minutes in the mobile era, and is now being compressed to around 90 seconds in the AI agent era. AI shopping agents are transforming the purchase funnel with unprecedented personalization and speed; today, 95% of the consumer journey occurs before purchase, making personalization not an option but a moat.
Underpinning this transformation are new protocol standards. Anthropic's open-source Model Context Protocol (MCP) enables agents to seamlessly access real-time information across the entire internet, while OpenAI's Agent Commerce Protocol (ACP) can protect end-to-end transactions. These protocols are simplifying and powering transactions in the AI era.
The scale of the market opportunity is staggering. ARK predicts that global online consumer spending facilitated by AI agents will grow from about 2% of online sales in 2025 to approximately 25% by 2030, potentially exceeding $8 trillion in scale by then.
AI's share of search traffic is projected to increase from 10% in 2025 to 65% by 2030, with AI-related search ad spending growing at an annual rate of about 50%.
By 2030, AI agents could generate approximately $900 billion in commerce and advertising revenue, where lead generation and advertising are the dominant growth factors, far outweighing contributions from consumer subscription revenue.
Robotics: A Severely Underestimated GDP Engine If AI is the central dynamo of the digital world, then robotics is its most crucial "physical embodiment." The report emphasizes that rapid advancements in AI are transforming robots from specialized devices for fixed tasks into relatively open general-purpose platforms, which is key to unlocking their potential in industrial and domestic markets. ARK estimates the global revenue opportunity for the robotics market is approximately $26 trillion, divided into two main segments: manufacturing and domestic services.
In the manufacturing sector, with global manufacturing GDP projected to reach $32 trillion by 2030, robotics technology could create a revenue opportunity of about $13 trillion, assuming a 100% labor productivity increase and a 35% service provider share. In the domestic services sector, approximately 2.8 billion global workers engage in 2.3 hours of unpaid household labor daily. Valued at the global average hourly wage of $12 and a 50% time value conversion, this also corresponds to a market space of around $13 trillion.
ARK particularly highlights the macroeconomic significance of Humanoid Robots. A frequently overlooked fact is that a significant amount of household maintenance, care, cleaning, and management labor is not counted in GDP today. ARK's calculations show: A single domestic humanoid robot → could convert about $62,000 worth of隐性劳动 into explicit GDP annually; If 80% of US households adopt them within 5 years → annual GDP growth could leap from 2–3% to 5–6%. The report argues this is not a story of "job replacement" but rather transforming non-market activities into market activities, liberating time as productivity. Autonomous Driving Reaches an Inflection Point ARK assesses that the complexity of humanoid robots is roughly 200,000 times higher than that of autonomous vehicles. This complexity ratio defines the theoretical capability required for full autonomy. Nevertheless, by mapping the relationship between the computational power required for Tesla's Full Self-Driving (FSD) and its performance improvements, ARK predicts that, under conditions of continued AI compute expansion and hardware progress, the Optimus humanoid robot could achieve human-level task execution capability around 2028. Autonomous ride-hailing is beginning to erode the market share of traditional ride-hailing services. In San Francisco's operational areas, Waymo's market share is already putting pressure on Uber and Lyft. Companies like Waymo, Baidu's Apollo Go, and Pony.ai have accumulated billions of miles in autonomous driving, with daily unmanned mileage growing rapidly. Cost reduction will be key to driving demand. ARK forecasts that by 2035, the price per mile for global autonomous ride-hailing could drop to $0.25, significantly lower than the $2.80 for human-driven ride-hailing and $0.80 for private car ownership in the US in 2025. In the early commercialization phase, vehicle costs will dominate unit economics, while vehicle utilization will drive down the cost per mile at scale.
The potential market value is enormous. ARK estimates that by 2030, autonomous ride-hailing could create approximately $34 trillion in enterprise value, with autonomous technology providers capturing about 98% of EBIT and enterprise value, while automakers and fleet operators hold relatively smaller shares. A key risk for this prediction is whether automakers other than Tesla can scale their autonomous ride-hailing fleets fast enough.
Autonomous logistics also holds broad prospects. Fully automated last-mile delivery—whether by drone or ground robot—already exceeds 4 million annualized deliveries globally. Autonomous long-haul trucking has commenced in the US, with operators planning rapid route expansion. ARK predicts that by 2030, global autonomous delivery revenue could reach $480 billion, with regulation and automation of backend loading operations being important limiting factors.
Multi-omics and AI Drive Biological Breakthroughs Multi-omics—encompassing genomics, epigenomics, transcriptomics, proteomics, and metabolomics—combined with AI is creating a flywheel effect for biological innovation. This flywheel includes: generating richer, lower-cost biological data, enabling more accurate tests, yielding better biological insights, developing AI-driven drugs, and ultimately achieving disease cures. Data generation costs are plummeting. The cost of whole-genome sequencing could drop to $10 by 2030, a roughly 10-fold decrease from 2015.
This will drive a surge in sequencing demand. The number of next-generation molecular diagnostic tests is projected to grow from less than a million in 2020 to about 7 million by 2030. The annual volume of tokenized data generated could reach approximately 200 billion tokens, surpassing the 150 trillion tokens used to train leading language models like those from OpenAI, Gemini, Anthropic, and xAI.
AI-enabled diagnostic capabilities are reaching an inflection point. Following ChatGPT's launch, the success rate of FDA-approved AI-driven tests and devices showed an inflection point from single-digit percentage levels. ARK's best-fit model suggests that the proportion of AI-driven diagnostics and devices could expand to about 30% by 2030 and eventually approach 100%. Drug development economics are being reshaped. AI-driven drug development could shorten time-to-market by about 40%, from 13 years to 8 years, while reducing total drug costs by approximately 4 times, from $2.4 billion to $700 million. Combining AI acceleration and disease cure potential, the value of an AI-designed drug in Phase I clinical trials could exceed $2 billion, whereas traditional drug assets often only recoup capital costs. The market potential for biological cures is particularly staggering. ARK research indicates that the average price for curing a rare disease could currently exceed $1 million, nearly 15 times the lifetime prescription cost for managing the disease. Curative drugs can capture revenue from most of the patient population before patent expiration, potentially making them 20 times more valuable than typical drugs and 2.4 times more valuable than prescription drugs for chronic diseases.
A broader perspective involves the extension of healthspan. If the US population could live in perfect health to the theoretical maximum lifespan of 120 years, but accident mortality risks remain, this would yield a gain of 119 billion Quality-Adjusted Life Years (QALYs). Valuing each healthy life year at $100,000, the potential lifespan gain market opportunity is approximately $1.2 quadrillion. The current global biotechnology market represents only about 0.1% of this potential market.
Reusable Rockets Open the Space Economy SpaceX's reusable rocket technology is pushing the economy into the space age. In 2025, the annual mass launched into orbit reached a historical high, with SpaceX dominating. The company has over 9,000 active Starlink satellites, accounting for about 66% of all active satellites in Earth's orbit.
Launch costs continue to fall. According to Wright's Law, launch costs should decrease by about 17% with each doubling of cumulative launch mass. Over 17 years since 2008, leveraging the partial reusability of the Falcon 9, SpaceX has reduced costs by approximately 95%, from about $15,600 per kilogram to under $1,000 per kilogram. ARK research suggests Starship could extend this trajectory to below $100 per kilogram at scale.
Satellite bandwidth costs are also declining. Per Wright's Law, satellite bandwidth cost per cumulative orbital gigabit per second (Gbps) should fall by about 44% with each doubling, enabling satellite connectivity to complement cellular towers and provide ubiquitous mobile coverage across the entire US.
A comparison shows that in 2001, US consumer mobile connectivity cost about $90 per month (in 2025 dollars) for just 0.001GB of data, covering about 1% of US land area; in 2025, it costs about $100 per month for unlimited high-speed internet, covering about 86% of the land area; by 2030, full coverage is expected at the same price. The market opportunity is substantial. Thanks to falling costs and improving performance, scaled satellite connectivity could generate over $160 billion in annual revenue, representing about 15% of ARK's forecast for global communications revenue. This prediction is based on the relationship between constellation bandwidth capacity and revenue opportunity, showing exponential growth potential.
Distributed Energy Supports AI Compute Demand Energy is increasingly efficiently powering economic growth. Despite concerns about energy intensity during the internet boom, economies actually became more energy-efficient, and the AI era might replicate this dynamic. The energy intensity (kWh required per dollar of GDP) of major economies like China, the US, Japan, India, and Germany has consistently declined over the past three decades. Multi-omics data costs are plummeting. Solar and battery costs continue to decline following Wright's Law. Nuclear energy cost declines were interrupted in the 1970s due to regulatory changes, but recent US executive orders should push nuclear power back onto its previous cost-decline trajectory. Historically, solar and nuclear costs (per megawatt) and battery costs (per megawatt-hour) have fallen significantly with each doubling of cumulative capacity. Electricity prices are expected to resume their downward trend. According to Wright's Law, ARK research shows that, except during World War II, US electricity prices steadily declined from the late 19th century until 1974, after which they plateaued due to increased regulation raising nuclear construction costs. ARK studies indicate that without heightened regulation, today's electricity prices might be about 40% lower than actual levels. As low-cost generation scales up to serve power-hungry AI data centers, retail electricity prices should start declining again after 50 years of stagnation.
Investment requirements are massive. Given ARK's rapid GDP growth forecast, cumulative capital expenditure in global power generation must expand by about 2 times to approximately $10 trillion by 2030 to meet global electricity demand. Consequently, stationary energy storage deployment needs to expand by another ~19 times. Between 2026 and 2030, data centers are expected to account for about 5% of total power generation investment.
Digital Asset Market Shows Evolving Trends Influenced by potential regulatory frameworks from legislation like the GENIUS Act, stablecoin activity saw significant growth in 2025. Several companies and institutions announced stablecoin-related plans, with BlackRock disclosing preparations for an internal tokenization platform. Stablecoin issuers and fintech companies like Tether, Circle, and Stripe launched or supported Layer 1 blockchains optimized for stablecoins. Data shows the market value of tokenized real-world assets (RWA) grew approximately 208% in 2025, reaching about $18.9 billion. BlackRock's BUIDL money market fund reached approximately $1.7 billion, reportedly accounting for 20% of the ~$9 billion US Treasury tokenization market. Tether's XAUT and Paxos's PAXG held market sizes of about $1.8 billion and $1.6 billion respectively in the tokenized commodities market. ARK predicts that by 2030, the scale of tokenized assets could grow from $19 billion to around $11 trillion, although this forecast carries significant uncertainty. While sovereign debt currently dominates the tokenized market share, the future development path remains to be seen.