The Trump administration is quietly building a case to restrict advanced Chinese AI models from U.S. markets, a shift that would cement market control for OpenAI and Anthropic while reshaping the global technology landscape.
Administration officials have explored multiple enforcement mechanisms to accomplish this goal without formally announcing a ban. The Commerce Department last year considered placing Chinese AI laboratories on its Entity List, a designation that would effectively require a government license for any U.S. access. The National Security Agency and White House Office of the National Cyber Director both explored issuing warnings about Chinese AI threats. An earlier draft executive order would have forced U.S. companies hosting Chinese models to guarantee security or face liability for breaches.
Commerce also circulated draft rules designed to leverage domestic supply chain authorities specifically to target open-source Chinese models. These proposals died internally, blocked by administration voices prioritizing innovation over restriction. But with key moderates like former White House adviser Sriram Krishnan departing and national security hawks gaining influence, momentum is rebuilding.
The timing matters. Kimi, a Chinese AI model, recently demonstrated capabilities rivaling U.S. competitors. American companies have increasingly adopted Chinese open-source models because they cost less and perform comparably to domestic alternatives. A policy shift would upend that calculation.
Rather than a direct ban, sources close to government discussions describe a slower, more durable strategy. Procurement rules, Entity List threats, and coordinated public pressure campaigns targeting companies that use Chinese models could achieve similar effects. Administration thinking also includes highlighting potential security vulnerabilities and governance gaps in Chinese platforms while encouraging investment in U.S. open-source alternatives.
David Sacks, an outside White House AI adviser, warned on social media Sunday that dominant U.S. labs want government to eliminate open-source competitors. He noted that leading AI companies approach the administration every three to five months with new proposals to restrict open-source technology.
The restriction strategy faces a fundamental problem: major U.S. open-source alternatives remain limited compared to the cheaper, more practical Chinese models currently available. Any meaningful shift would leave American developers with fewer viable options, not more.
Sacks called on the broader technology industry to resist what he described as regulatory capture, urging Silicon Valley to openly oppose measures that would benefit the largest labs while suppressing competition. OpenAI and Anthropic have long advocated for licensing regimes and safety regulations that would effectively serve the same gatekeeping function.
Neither the White House nor the Commerce Department responded to requests for comment on the administration's deliberations or current posture toward Chinese AI access.
Author James Rodriguez: "This looks less like principled tech policy and more like a replay of the oldest Silicon Valley playbook: use government to lock out competitors while claiming it is about national security."
Comments