OECD Venture Capital Investments in AI Through 2025: Global Trends & Data Analysis

AI Venture Capital Reaches Record $258.7 Billion in 2025

The global venture capital landscape has undergone a seismic shift over the past decade, and artificial intelligence now sits squarely at its centre. According to the latest OECD report published in February 2026 — drawing on comprehensive Preqin data hosted on the OECD.AI Policy Observatory — global AI venture capital investment reached a staggering USD 258.7 billion in 2025. This figure represents a full recovery from the market correction that brought AI VC down to USD 123.6 billion in 2023, and effectively matches the previous peak of USD 257.3 billion recorded in 2021.

The trajectory tells a compelling story about investor confidence in artificial intelligence. From roughly USD 8.3 billion in 2012, AI venture capital has grown by a factor of 31 over thirteen years. The dataset underpinning the OECD analysis covers more than 33,000 AI firms and approximately 85,000 transactions, making it one of the most comprehensive assessments of AI investment activity available. As the broader tech investment market continues to normalise after its pandemic-era excesses, AI has emerged not just as a survivor but as the primary engine driving venture capital worldwide.

This report matters because it provides policymakers, investors, and technology leaders with empirical evidence about where capital is flowing, which sectors are attracting the most attention, and how the global distribution of AI investment is evolving. For those tracking the competitive dynamics shaping the AI industry, the OECD data offers a rigorous, data-driven foundation for strategic decisions.

How AI Captured 61% of Global Venture Capital

Perhaps the most striking finding in the OECD report is the sheer dominance of AI within the broader venture capital ecosystem. In 2022, artificial intelligence companies attracted approximately 30% of all global VC investment. By 2025, that figure had doubled to 61%. To put this in perspective, total global venture capital across all sectors reached USD 427.1 billion in 2025 — meaning AI companies alone absorbed USD 258.7 billion of that total, while all non-AI ventures combined received just USD 168.4 billion.

This divergence reflects fundamentally different trajectories for AI and non-AI venture capital. The broader VC market peaked at over USD 800 billion in 2021, fuelled by pandemic-era investments in remote work, telemedicine, and digital communications. Since then, total VC has declined approximately 47%, stabilising near pre-pandemic levels. AI venture capital, however, followed a different arc: after dipping to USD 123.6 billion in 2023, it surged 109% over the next two years to reclaim its 2021 highs.

The implication is clear — AI is not merely a hot sector within venture capital; it has become the dominant category, attracting more investment than all other technology sectors combined. As digital economy reports from UNCTAD have also highlighted, this concentration of capital in AI-related ventures is reshaping the entire innovation funding landscape.

Geographic Breakdown: Where AI VC Dollars Flow

The geographic distribution of AI venture capital remains heavily concentrated in a handful of countries, with the United States commanding a dominant position. In 2025, US-based AI firms received approximately USD 194 billion in venture capital — roughly 75% of the global total. The European Union collectively attracted USD 15.8 billion (approximately 6%), while China received USD 13.9 billion (5%) and the United Kingdom captured USD 13.8 billion (5%). The remaining 9% was distributed across 79 additional jurisdictions worldwide.

From the investor side, the picture shifts slightly but the US remains dominant. American investors deployed approximately USD 124 billion in AI venture capital (56% of identified outgoing investment), followed by UK-based investors at USD 20.7 billion (9%), Chinese investors at USD 17.2 billion (8%), and EU27 investors at USD 14.5 billion (7%). Notably, approximately USD 36 billion in AI VC investment had no identified investor country, indicating the complexity of tracking cross-border capital flows in an increasingly globalised market.

The United States sits as a net beneficiary of international AI investment flows — AI firms based in the US received more venture capital than US-based investors deployed outward. The EU27 also emerged as a net beneficiary. Meanwhile, countries including the United States, United Kingdom, Israel, and Canada saw more than 50% of all their venture capital flow specifically to AI firms, underscoring how AI has become the centrepiece of their innovation ecosystems.

China presents an interesting counterpoint. While it remains a major player in absolute terms, China’s AI VC as a share of total domestic VC has shown a flatter trend compared to Western economies. This is partly because Chinese investors have simultaneously increased non-AI investments in sectors such as energy, raw materials, utilities, and autonomous vehicles. When measured as a proportion of GDP, Israel and the United States lead with AI venture capital intensity approaching 0.4% of GDP, followed by the United Kingdom and Canada.

Generative AI Funding Surge From $2.8B to $35.3B

The explosive growth of generative AI — catalysed by the public release of ChatGPT in late November 2022 — has left an unmistakable imprint on venture capital flows. According to the OECD data, generative AI venture capital surged from just USD 2.8 billion in 2022 (approximately 2% of total AI VC) to USD 15.3 billion in 2023 (12%) and continued climbing to USD 35.3 billion in 2025 (14% of all AI VC investment). This represents a roughly 12-fold increase in just three years.

The generative AI boom has driven investment into two primary areas: large language model development and the compute infrastructure required to train and deploy these models. Companies like Anthropic, Mistral AI, and X.AI have attracted billions in funding as investors race to back the next generation of foundation models. Mistral AI’s EUR 1.7 billion Series C in 2025 — led by semiconductor giant ASML — illustrates how generative AI funding is drawing in non-traditional investors from across the technology stack.

Despite the rapid growth, generative AI still represents only 14% of total AI venture capital, suggesting that the broader AI ecosystem — spanning enterprise applications, autonomous systems, healthcare AI, and more — continues to attract the lion’s share of investment. The generative AI wave has amplified overall AI investment rather than simply redirecting existing capital, which explains why total AI VC recovered to record levels even as broader VC contracted.

Mega Deals and the Concentration of AI Capital

One of the most consequential trends identified in the OECD report is the growing concentration of AI venture capital in mega deals — transactions exceeding USD 100 million. In 2025, these mega deals comprised approximately 73% of total AI VC investment value. Even more striking, deals exceeding USD 1 billion represented almost half of all venture capital investment value across the year. The top five mega deals alone totalled roughly USD 63 billion, accounting for about 25% of the entire global AI VC market.

The average AI VC deal size has more than tripled over the past decade, rising from approximately USD 11.2 million in 2014 to USD 35.8 million in 2025. However, the median deal size of just USD 5 million reveals the extreme skew in deal value distribution. A relatively small number of blockbuster transactions — such as Databricks’ USD 1 billion Series K at a valuation exceeding USD 100 billion, and Anysphere’s USD 900 million Series C at a USD 9.9 billion valuation — are disproportionately driving headline figures.

This concentration raises important questions about the health of the AI startup ecosystem. While headline numbers suggest a booming market, the reality is more nuanced. The proliferation of mega deals may come at the expense of early-stage companies that struggle to compete for investor attention and capital. Among mega deals with a specified investment stage, only 14% were classified as early-stage in 2025 — the vast majority went to established companies with proven business models and significant existing traction.

AI Investment by Industry Sector and Use Case

The sectoral composition of AI venture capital has shifted dramatically in recent years. The OECD report reveals that IT infrastructure and hosting — a category encompassing both compute infrastructure companies and model developers — has become the dominant industry for AI VC since 2023. In 2025, this sector attracted a remarkable USD 109.3 billion in investment, nearly matching all other industries combined (USD 149.4 billion). Its cumulative total from 2012 to 2025 reached USD 256.1 billion.

This sectoral dominance reflects the capital-intensive nature of building and scaling large AI models. Training frontier models requires massive investments in GPU clusters, data centres, and specialised infrastructure. Companies classified under this category include both pure infrastructure plays like Databricks and model developers such as Anthropic and X.AI, blurring traditional distinctions between infrastructure and application layers.

For much of the previous decade, mobility and autonomous vehicles held the top position for AI VC investment, accumulating USD 147.4 billion between 2012 and 2022 — with 94% of that capital flowing to US and Chinese companies. However, investment in this sector has softened in recent years as the initial hype around self-driving vehicles has given way to a more measured assessment of deployment timelines and technical challenges.

Healthcare, drugs, and biotechnology experienced its own investment cycle, more than tripling from approximately USD 8 billion in 2019 to USD 25 billion in 2021 — a surge driven by COVID-19 pandemic needs. Investment tapered to USD 12 billion in 2023 as pandemic urgency waned but recovered to USD 20 billion by 2025, reflecting sustained interest in AI applications for drug discovery, diagnostics, and clinical decision support. For a broader perspective on how technology investments are reshaping economies, the BIS Annual Economic Report 2025 provides complementary analysis of capital allocation trends.

Early-Stage vs Late-Stage AI Startup Funding

The OECD data reveals a growing divergence between early-stage and late-stage AI venture capital that has significant implications for the future of AI innovation. By deal count, early-stage investments reached a historic high in 2025, comprising just over 75% of all AI VC transactions. This suggests that entrepreneurial activity in AI remains vibrant, with new startups continuing to attract initial rounds of funding at an accelerating pace.

However, the picture changes dramatically when measured by investment value. Early-stage deals have consistently accounted for approximately 25% of total AI VC deal value over the past decade — a remarkably stable proportion. The average early-stage deal in 2025 was approximately USD 11.8 million, compared to USD 131 million for later-stage deals. This eleven-fold difference in average deal size explains how early-stage companies can dominate by deal count while representing a modest share of total capital deployed.

Since 2023, AI firms have attracted a declining share of early-stage VC relative to all funding rounds, as capital increasingly concentrates in mega deals over USD 100 million. This trend has important implications for the pipeline of AI innovation. While established AI companies are attracting record sums, smaller startups working on novel approaches, niche applications, or emerging use cases may find it increasingly difficult to secure the capital needed to scale. The OECD notes that this creates a potential vulnerability in the innovation ecosystem — one that policymakers should monitor and address through targeted interventions such as co-investment schemes and matching grants.

Notable examples of early-stage deals mentioned in the report include Deep Algorithms Solutions in India (~USD 1.3 million seed for AI fraud defence), Antescofo SAS in France (EUR 4 million seed for AI-powered music tools), and FileAI in Singapore (USD 14 million Series A for AI data pipelines). These deals illustrate the global breadth of early-stage AI activity, even as the majority of capital gravitates toward later-stage US-based companies.

OECD Policy Recommendations for AI Investment

The OECD report goes beyond data analysis to offer five substantive policy recommendations aimed at strengthening and diversifying the global AI investment ecosystem. These recommendations reflect the organisation’s dual mandate of promoting economic growth while ensuring that AI development proceeds in a trustworthy and inclusive manner.

1. Promote AI Literacy for Investors and Investment Literacy for Policymakers

The report calls for tailored AI literacy programmes that help venture capital investors better understand AI risks, opportunities, and technical capabilities. Simultaneously, it recommends that policymakers develop investment literacy to understand the VC ecosystem’s dynamics, incentive structures, and limitations. This bidirectional knowledge transfer is essential for crafting effective policies that support scaling domestic AI firms across all maturity levels.

2. Provide Policy Incentives for Scaling AI Firms

Recognising that market forces alone may not adequately support AI firms with longer-term societal benefits, the OECD recommends tax credits, grants, matching public financing, and hybrid public-private co-investment platforms. These mechanisms can help bridge the gap between early-stage innovation and commercially viable scale, particularly for companies working in areas of strategic national interest. For analysis of how global trade policy intersects with technology investment, the WTO’s latest outlook provides valuable context.

3. Foster Enabling Investment Environments

Clear, predictable regulatory frameworks — including regulatory sandboxes and specialised guidance agencies — are critical for attracting both domestic and international AI investment. The report emphasises understanding gaps in local investment ecosystems and taking steps to catalyse local capital while creating conditions that attract foreign investment. This is particularly relevant for emerging economies seeking to develop competitive AI sectors.

4. Build Resilience Through Diversification Across the AI Stack

With investment heavily concentrated in IT infrastructure and model development, the OECD encourages diversification across the full AI technology stack — including semiconductors, energy-efficient hardware, edge computing, and AI security. This diversification would create more resilient ecosystems capable of sustaining innovation during market corrections and reduce dependence on any single layer of the technology stack.

5. Explore Drivers of Investment in Trustworthy AI

The final recommendation calls for systematic research into what drives investment decisions around trustworthy AI — including safety, fairness, transparency, and robustness. The OECD suggests studying investor criteria, assessing the extent to which trustworthy AI considerations factor into investment strategies, and potentially developing standardised terminology to facilitate alignment with the OECD Recommendation on AI. For those interested in the safety dimension, the technical analysis of AGI safety and security challenges offers a deeper exploration of these considerations.

What the Data Means for AI Investors and Policymakers

The OECD’s comprehensive analysis of venture capital investments in AI through 2025 paints a picture of an industry at an inflection point. On one hand, the recovery to USD 258.7 billion and AI’s capture of 61% of global VC signals extraordinary investor confidence in the transformative potential of artificial intelligence. On the other hand, the extreme concentration of capital — geographically, sectorally, and by deal size — raises legitimate questions about the sustainability and inclusiveness of the current investment trajectory.

Several key tensions emerge from the data. The first is between growth and concentration: while total AI investment has recovered to record levels, this recovery has been disproportionately driven by a handful of mega deals in a single country (the United States) and a single sector (IT infrastructure). The second tension is between early-stage dynamism and late-stage dominance: the record number of early-stage deals in 2025 suggests robust entrepreneurial activity, but the capital flowing to these deals remains a fraction of what established players attract.

The OECD is careful to note that investment markets are inherently cyclical and that past performance does not guarantee future returns. AI has experienced several waves of heightened attention and funding throughout its history, each followed by periods of correction. While long-term signals point to continued advances in AI capabilities and applications, short- to medium-term investment outcomes remain uncertain.

For investors, the data suggests a need for careful attention to portfolio diversification across AI subsectors, geographies, and investment stages. The dominance of IT infrastructure investment may create opportunities in underserved areas such as AI safety, application-layer companies, and emerging market AI ecosystems. For policymakers, the report provides a clear roadmap: invest in AI literacy, create enabling regulatory environments, provide targeted incentives for scaling AI firms, and build resilient, diversified innovation ecosystems that can weather the inevitable cycles of the venture capital market.

The OECD report ultimately underscores a fundamental reality: venture capital represents just one dimension of AI investment. Using System of National Accounts data, the report estimates that AI-related capital formation within the EU27 alone potentially reached EUR 294 billion in 2023 — approximately 2.4 times higher than total AI-related VC for that year. This highlights the crucial role of internal corporate investments, public sector funding, and other capital sources in shaping the AI landscape. Understanding the full picture of AI investment will require continued expansion of data collection, analysis, and international cooperation — exactly the type of work the OECD is uniquely positioned to lead.