At the World Economic Forum in Davos earlier this year, Nvidia CEO Jensen Huang outlined a framework for artificial intelligence comprising five layers: energy, chips, cloud, models, and applications. He emphasized the importance of “sovereign AI,” arguing that nations must maintain control over the production of their own AI technologies. What started as an idea intended for smaller nations pursuing digital self-reliance has evolved into a guiding policy for the United States and China, the world’s two largest digital economies.

In recent developments, former U.S. President Donald Trump indicated that the U.S. government might take equity stakes in leading AI companies such as OpenAI and xAI, proposing that the American public become “partners” in the AI economy. Meanwhile, in China, the AI start-up DeepSeek reportedly completed its first external funding round, reaching a valuation exceeding $50 billion. China’s National Artificial Intelligence Industry Investment Fund—a state-controlled entity used to channel investments into priority AI ventures—was a major contributor and notably the only investor to obtain direct corporate ownership and voting rights.

These moves reflect a shared conclusion: AI is strategically critical enough to warrant direct government involvement beyond traditional regulatory frameworks or market competition. Analysts have pointed to earlier government interventions, such as the Trump administration’s acquisition of a 10 percent equity stake in semiconductor company Intel, which aimed to safeguard domestic semiconductor capacity. The current proposals to take stakes in AI model developers represent an extension of that strategy, moving up Huang’s stacked model of AI infrastructure.

China’s approach has similarly progressed over the past decade, initially focusing on semiconductor supply through the China Integrated Circuit Industry Investment Fund. The country’s more recent efforts, including state investment in DeepSeek, extend this strategy to AI software and intelligence layers, underpinned by domestic hardware such as Huawei Ascend clusters. Notably, despite a partial lifting of U.S. export controls on Nvidia’s second-tier H200 chips following the U.S.-China summit in Beijing, Chinese authorities have blocked their acquisition, focusing instead on supporting indigenous alternatives like DeepSeek.

This trend marks a shift from government regulation to active ownership in frontier AI enterprises. While states have historically subsidized or protected strategic industries, their new approach involves becoming direct shareholders in companies that develop cutting-edge AI technologies. This shift carries significant implications for governance, raising complex questions about how to balance national security interests with commercial objectives, manage potential conflicts on company boards, and reconcile serving global markets with advancing national priorities.

Although OpenAI and DeepSeek remain competitors, they illustrate a broader structural movement in which both the U.S. and China are deploying state capital across increasingly strategic layers of AI development. The initial phase concentrated on physical infrastructure like chips, followed by investments in AI models. The evolving focus may soon encompass the ownership and governance of AI intelligence itself, signaling a new dimension in the geopolitical stakes surrounding artificial intelligence.