The Chinese leadership has put artificial intelligence (AI) front and center in China’s industrial development. As part of this strategy, they have established industry alliances [产业联盟]—collaboration platforms involving local governments, academic institutions, and companies.
This data brief provides a high-level assessment of the role of industry alliances in China’s AI strategy and closely examines one major group: the Artificial Intelligence Industry Alliance (AIIA) [中国人工智能产业发展联盟]. Through the AIIA, the Chinese government aims to foster collaboration among local governments, academic institutions, and companies. In some cases, the Chinese state uses the AIIA to “pick winners,” choosing among favored companies in the AI industry to receive government subsidies.
In addition to its role of advancing China’s industrial policy, the AIIA is notable as a diverse sample of companies active in the Chinese AI industry, ranging from well-known, publicly traded tech giants to obscure startups. This sample helps illuminate aspects of China’s AI industry and provides a fuller picture of the distribution of AI activity in China.
We find that:
- AIIA members come from all over China. Roughly 71 percent of members are clustered in the four wealthiest, most populous first-tier cities, while 22 percent are scattered throughout “emerging new first-tier cities.”
- Most AIIA industry members are diversified, typically large companies; a sizable minority are “pure play,” AI-focused companies, which are predominantly startups and small enterprises.
- The alliance’s AI-focused companies tend to specialize in AI applications for the commercial market. 63 percent of these companies focus on AI-driven business solutions and data storage and processing services.
These conclusions draw on open-source data, collected and annotated by CSET, on the AIIA’s hundreds of members, including their websites, media coverage, and commercial databases. Further details on the dataset are provided in Appendix 1, and the full annotated dataset of AIIA members is available for download at CSET GitHub.