The impact of government-backed venture capital on artificial intelligence startups' productivity: focusing on broker roles

This study investigates the effects of government venture capitals (GVCs) backing on the productivity of startups compared to private venture capitals (PVCs). Based on a sample of 1,149 artificial intelligence startups in South Korea, we find that startups backed by GVCs show better productivity than those backed by PVCs. In examining the underlying mechanism of such higher performance by GVC-backing, we find that GVC-backed startups are more likely to acquire proprietary data and join R&D testbed programmes, thus leading to better productivity compared with PVC-backed startups. We interpret our findings as evidence of the positive abilities of GVCs as intermediaries for their portfolio startups.