The United States still leads the global AI race, with greater investment, infrastructure and superior AI models, but the gap is closing fast. Chinese models like DeepSeek-R1 and the recent Kimi K2 Thinking model have an edge over American models in cost efficiency and certain analytical functions.

The United States’ status as the leading technology and innovation superpower has never been more uncertain. With the future of innovation in medicine, cybersecurity and defense hinging on the victor of the global AI race, U.S. policymakers must do all they can to ensure America does not squander its lead.

Beijing undoubtedly wants to see China become the global AI superpower. The Chinese government has offered significant state-backed incentives to companies developing AI models and boosted energy subsidies for data centers. Similarly, funding is being poured into data-center projects, with a planned $37 billion investment in the 760-acre “data island” in Wuhu.

After AI development was prioritized in China’s 15th Five-Year Plan, Goldman Sachs forecast that Chinese cloud service providers will increase capital expenditures by 65% in 2025, with $70 billion being invested to support development.

Dividends have already been paid, with Moonshot AI’s language model, Kimi K2-thinking, outperforming OpenAI’s GPT-5 and Anthropic’s Claude Sonnet 4.5 in key tests. Worse still for the U.S. AI industry, these models are much cheaper to develop in China than they would have been in America.

Beijing also intends to place AI systems at the core of its future national security and defense strategies. Last year, researchers at Shanghai Jiao Tong University won a government contract after researching how artificial intelligence could be used to deploy weapons in automated “kill webs” that adjust in real time during combat at sea.

To maintain its lead in the global AI race,  the U.S. must create an attractive environment for AI companies, where low regulation and government support enable innovation to flourish. This is already happening, with OpenAI developing a $500 billion data center in Texas that dwarfs the investment in the Wuhu center. Congress needs to pass the 10-year ban on state-level AI regulation that was removed from the Tax Cut and Spending Bill in July. Over-regulation stifles American growth and aids China in this crucial race for global strength.

Home to the world’s most powerful AI chip, the Nvidia Blackwell, the U.S. must continue restricting access to its technology. Without the most powerful chips, China may struggle to overtake the U.S. in AI innovation and power. DeepSeek has already faced this problem. Not having access to Nvidia’s strongest chips delayed the release of one of its models, as the domestic chips did not perform as well as the American ones.

The AI race is fundamentally about the future of global power, in arenas from defense and security to healthcare. The U.S. is still in the lead, but will only maintain it if policymakers continue to act.

Samuel Field is a policy fellow at the Pinsker Centre, a British-based foreign policy think tank focusing on the Middle East and wider international affairs./InsideSources