Governmental Venture Capital and Startup AI Innovation: Evidence from China

This study examines whether governmental venture capital (GVC), relative to private venture capital (PVC), more effectively steers startup innovation toward strategic, high-uncertainty technologies, using the rise of artificial intelligence (AI) in China as our empirical setting. Our analysis is based on a comprehensive sample of Chinese startups (operating outside the core AI industry) that received venture capital between 2010 and 2019. We leverage the staggered rollout of GVC investments across Chinese startups as a quasi-natural experiment and employ a difference-in-differences (DID) design combined with propensity score matching (PSM). Our baseline results show that GVC-backed startups generate significantly more AI patents than PVC-backed counterparts. We also find that the innovation-promoting effect of GVC is amplified in regions with greater state-owned sector dominance but attenuated in regions with higher corruption. Our findings are robust to a series of tests, including parallel trends validation, placebo analyses, an instrumental variable (IV) strategy using regional GVC fund density and alternative specifications. Further heterogeneity analyses indicate that this effect is stronger for startups backed by VCs with less specialization or lower IPO experience, and that it becomes significantly more pronounced in the latter half of the 2010s as China's AI policy ecosystem matured. Overall, this study demonstrates how GVC shapes corporate innovation strategy in emerging technological waves, with direct implications for the design of public venture programs and innovation policy in technology-driven economies.