Limited capital? Explaining transatlantic divergences in artificial intelligence startup financing

Artificial intelligence (AI) is a global general-purpose technology (GPT), yet the financial systems that scale AI ventures differ markedly across regions. This paper compares AI-focused venture capital (VC) dynamics in Europe and the United States from 2000 to 2025 using approximately 8,000 AI-related VC rounds from PitchBook and macroeconomic and institutional data from the World Bank. Exploiting the 2021 generative-AI inflection point as a common technological shock that triggered a structural reallocation of venture capital toward AI ventures, we apply a difference-in-differences (DiD) design to distinguish between startup formation and scale-up financing. Descriptive evidence shows that Europe keeps pace with the US in AI deal counts, suggesting no major pipeline gap in startup formation. However, Europe lags substantially in AI investment volumes and in AI’s share of total VC funding. In the Post-period, AI accounts for around 60% of VC investment volume in the US, compared to roughly 31% in Europe. Difference-in-differences estimates are consistent with parallel trajectories in AI deal activity but point to large and statistically significant transatlantic gaps in AI investment volume and investment intensity. These results are robust to macroeconomic and governance controls, a formal event-study test of pre-treatment trends, and a sensitivity analysis that excludes the United Kingdom (which alone accounts for roughly 31% of European VC volume), a placebo test using a false treatment year, and the exclusion of the COVID-19 period. Interpreting the findings through GPT diffusion and the Varieties of Capitalism framework, we interpret the pattern as being consistent with the US liberal market economy having a structural advantage in mobilizing large-scale risk capital. The results inform current debates on European technological sovereignty, the Capital Markets Union, and AI-related policy initiatives.