Cotton to Lutnick: Combat Communist China’s AI Diffusion Strategy

FOR IMMEDIATE RELEASE
Contact: Tatum Wallace or Hannah McCarthy
July 30, 2026

Cotton to Lutnick: Combat Communist China's AI Diffusion Strategy

WASHINGTON - Senator Tom Cotton (R-Arkansas) sent a letter to Commerce Secretary Howard Lutnick encouraging a proactive, coordinated, multipronged approach to combat Communist China's AI diffusion strategy and ensure lasting U.S. AI dominance, both domestically and globally. The letter follows recent concerns about increased American reliance on Chinese AI models.

In part, Senator Cotton wrote:

"I write to you concerned about American companies increasingly relying on Chinese AI models. While the U.S. is the leader in the most cutting edge, closed models, recent progress in Chinese AI models highlights alarming gaps. To stay ahead in the AI race, the U.S. must fully capture AI's economic and geopolitical returns, while continuing to create the most advanced frontier models. This means ensuring that as many U.S. firms, researchers, and institutions as possible are able to leverage AI for productivity and scientific gains. It also means ensuring that global customers adopt and build on our models, where we can set standards, norms, and dependencies."

Full text of the letter may be found here and below.

July 29, 2026

The Honorable Howard Lutnick
Secretary of Commerce
Department of Commerce
1401 Constitution Avenue, NW
Washington, D.C. 20230

Dear Secretary Lutnick,

I write to you concerned about American companies increasingly relying on Chinese AI models. While the U.S. is the leader in the most cutting edge, closed models, recent progress in Chinese AI models highlights alarming gaps. To stay ahead in the AI race, the U.S. must fully capture AI's economic and geopolitical returns, while continuing to create the most advanced frontier models. This means ensuring that as many U.S. firms, researchers, and institutions as possible are able to leverage AI for productivity and scientific gains. It also means ensuring that global customers adopt and build on our models, where we can set standards, norms, and dependencies.

Communist China is pursuing an aggressive AI diffusion strategy, and Chinese open weight models have become dominant players in the AI space. Foreign and domestic customers seeking cost savings and more flexibility are increasingly moving to Chinese models as a "good enough" alternative to proprietary U.S. models. One report estimates that nearly 80% of U.S. startups build applications on Chinese models. Larger U.S. companies are following suit, with applications such as Airbnb's Qwen-powered chatbot and Cursor's Kimi-based coding agent. The latter is especially concerning, as defense primes use Cursor's coding tools to build software for the Department of War (DOW), even though Chinese models generate lower quality, more vulnerable code when the user is American. Microsoft, one of the largest federal IT vendors, is also considering DeepSeek as an alternative for its Copilot Cowork. Backdoors may be embedded into open weight Chinese models, and auditing produces limited assurances. American companies, particularly those that do business with the U.S. government, should never jeopardize our security for higher profits.

I encourage you to pursue a proactive strategy to address these issues, with a coordinated, multi-pronged approach. First, we should continue to deny Communist China the frontier and limit its ability to serve customers on a global scale by enacting and enforcing strict export controls on advanced compute and semiconductor manufacturing equipment. Second, I would urge that we guard against deepening dependence on Chinese AI models. DOW already bars agency and contractor use of DeepSeek, and other Federal agencies have also blocked access internally. I propose we extend these prohibitions to additional Chinese models and enact a government-wide ban on contractor use. American businesses must understand that the U.S. government will not accept the risks of Chinese AI, and such a prohibition also sends an unmistakable demand signal for American alternatives. Finally, I recommend that we actively nurture an open AI ecosystem in the U.S. to rival Communist China's. Customers at home and abroad will prefer an American option, and that is how we can take the lead in global AI diffusion.

America wins the AI race when we offer models that provide optionality, from the state-of-the-art frontier for high-assurance, mission-critical applications in defense, critical infrastructure, and advanced scientific research, to cheaper, more efficient models for various commercial and consumer workloads where price may be more important than raw capability. I commend the administration's current efforts and believe that if you take prompt action, America can keep the frontier advantage and unleash a diffusion strategy that ensures U.S. AI dominance, both domestically and globally.

Thank you for your attention to this matter.

Sincerely,

Tom Cotton
United States Senator

###