Government Venture Capital and AI Development in China

Government Venture Capital and AI Development in China [ 5 min read ]

Insights

  • Over the past decade, China’s central and local government VC funds used tax revenues, land sale proceeds, and contributions from state-owned partners to channel $912 billion into strategic industries like AI.
  • Unlike private VCs that cluster in richer coastal regions, government VC funds are geographically dispersed, present even in poorer regions, and often managed by local governments.
  • Government VC investments, compared to private VC investments, tend to support firms with less existing software production. However, these firms exhibit higher growth rates following investment.
  • Most AI firms (71%) with both government and private VC funding received government investment first, serving as a signal for private VCs to follow. But whether the innovation return justifies these sizable investments remains unclear.


Source Publication: Martin Beraja, Wenwei Peng, David Y. Yang, and Noam Yuchtman (2024). “Government as Venture Capitalists in AI.” National Bureau of Economic Research (NBER) working paper.

Read this brief on SUBSTACK

China’s government venture capital (VC) funds allow national and local governments to channel capital into early-stage firms in strategic industries like artificial intelligence (AI). In the past decade, China’s government VC funds have invested $912 billion, with yearly investments comparable to the annual spending on all U.S. government industrial policies during the same period. Twenty-three percent of this investment has been directed to 1.4 million AI-related start-ups across China. What role does this money have in shaping the development and direction of China’s VC industry and AI firms?  

The data. To measure the investment flows of China’s government VC funds (2000–2023), researchers analyzed data from Zero2IPO, China Venture Source, and Tianyancha, which included detailed information on government fund contributors, agencies, capital, and firm-level details. Since government VCs often invest in other VC funds, the researchers traced investments until over 90% could be linked to specific firms. The dataset includes 1,863 government VC funds, 6,514 intermediate funds, and 45,419 firm investors, alongside 331,302 private VC investments for comparison. They further identified over 9,000 AI firms using AI-related keywords in business registration data and tracked AI investments by government and private VCs.

Government VC funds: background. The primary funding for government VC funds comes from government budgets (tax revenues and land sale proceeds) and contributions from state-owned enterprises or banks. From 2013 to 2018, subnational government VC funds surged, with an average of 238 funds created annually. However, growth slowed to 115 funds per year since 2019 due to China’s economic slowdown and stricter regulations on what industries, firm stage, and locations they can invest in. 


Share of AI investment across fund types and industry sectors

3 pie charts show government VC investment in AI


Private VC funds cluster in wealthier regions, government VC funds (and AI firms) appear all over the country. Although private and government VC funds invest in similar sectors, private VC funds are concentrated in wealthy coastal regions, while government VC funds are more evenly distributed across China’s provinces. Private VC funds are overrepresented in richer regions but underrepresented in poorer interior areas, whereas government VC funds are the opposite: they are overrepresented, relative to the share of AI firms, in poorer regions. The more dispersed nature of government VC funds is likely due in part to regulations requiring local government VC funds to invest locally to generate local economic growth, substituting for absent private VC funds in less developed regions.

Government VCs, compared with private VCs, invest in firms with less software production, but these firms exhibit higher growth rates following investment. Researchers found that government VC funds often invest in AI firms with weaker initial software production than firms receiving private VC funding. Both types of investments lead to significant software production growth, with government-funded AI firms achieving 500% growth in software production by 2023, exceeding the growth in software production (but not total software production) of private VC-funded firms. As measured by software growth rates, AI firms backed by local VC funds outperform those supported by non-local VC funds, perhaps due to the importance of local knowledge and relationships in China’s generally low-information business environment. 


Government VC investment trends in the AI sector

Chart showing government and private VC investment in AI


Private VCs follow government investment. Between 2000 and 2023, 4,115 AI firms received investments from both government and private VCs, accounting for 52%  of government VC portfolios and 27% of private VC portfolios. Seventy-one percent of these firms were funded by government VC funds first, suggesting their investments act as positive signals for private VC  funds. Direct investments by hometown government funds were more likely to prompt follow-up private investments, especially for firms with less software production. These findings suggest that government investment induces private VC sector investment by identifying local firms that may be better matched  with the policy environment, especially when signals  on business performance are lacking.


Investment in AI firms: government VC funds vs. private VC funds

Chart showing government AI investment in China


Role of government VC funds reflect distinctive features of China’s business environment. On one hand, China’s government VC funds act like private VC funds by investing in early-stage firms. On the other hand, government VC investments convey distinct information that is otherwise unavailable, thereby influencing other market actors. In China’s low-information environment, government VC funds with access to exclusive information may guide private market investment and address informational asymmetries between state and private actors. However, whether the innovation return of government VC funds justifies the sizable investments remains unclear.