At the World Economic Forum in Davos earlier this year, Nvidia CEO Jensen Huang outlined a five-layer framework for artificial intelligence (AI) infrastructure, encompassing energy, chips, cloud, models, and applications. He emphasized the concept of "sovereign AI," where nations maintain control over the production of their own intelligence technologies. This notion, initially aimed at smaller countries pursuing digital self-determination, has become a guiding policy for the United States and China, the world’s two largest digital economies.

In a move reflecting this shift, President Donald Trump announced that the U.S. government intends to acquire equity stakes in leading AI firms such as OpenAI and xAI. This approach aims to position the American public as a partner in the emerging AI economy. Meanwhile, China has made similar strategic investments. DeepSeek, a Chinese AI-only startup valued at over $50 billion following its first external funding round, received significant backing from China’s National Artificial Intelligence Industry Investment Fund. This state fund notably secured direct corporate ownership and voting rights in DeepSeek.

Both the U.S. and China appear to share the belief that AI is too critical a sector to be left exclusively to private markets. This view complements a broader trend toward “sovereign AI” encompassing the entire technology stack, from hardware manufacturing to advanced AI models. The Trump administration’s earlier 10% equity acquisition in Intel highlighted a determination to safeguard domestic semiconductor production, and the recent steps regarding AI firms represent a similar strategy applied to higher layers of AI development.

China, benefiting from a decade’s lead in this realm, has long supported its semiconductor industry through the China Integrated Circuit Industry Investment Fund and now extends state capital into AI model development via its dedicated AI fund. Efforts to supply AI systems with domestically produced chips, such as Huawei Ascend clusters, have been prioritized, with recent U.S.-China summit negotiations highlighting tensions around chip sales. Although the U.S. partially authorized the sale of Nvidia’s H200 chips to select Chinese companies, Beijing has reportedly blocked these transactions, favoring domestic technology.

This transition from regulation to direct government participation marks a notable evolution in how states approach AI. Unlike conventional industrial subsidies or protections, these investments position governments as shareholders in frontier AI laboratories, elevating the companies beyond traditional software startups to strategic national infrastructure. The resulting ownership stakes grant governments a degree of influence that raises complex governance issues, including how to reconcile national security interests with shareholder returns, manage tensions between commercial and strategic objectives, and balance global market engagement with national priorities.

While OpenAI and DeepSeek are widely seen as competitors in the AI sector, their rise underscores a shared global dynamic where leading economies are channeling state capital into increasingly central elements of AI technology. The initial phase concentrated on semiconductor chips, followed by advances in AI models, with future discussions likely to center on the ownership and governance of AI intelligence itself.