Washington’s China Crackdown Could Hand America’s AI Giants Exactly What They Want

The Trump administration is considering a series of measures that could make it significantly harder for American companies to use Chinese artificial intelligence models, according to Axios, reviving a policy debate that critics say could strengthen the position of the country’s largest AI companies.

Officials have discussed using existing government authorities, including Entity List restrictions, procurement rules and liability requirements, to discourage businesses from deploying Chinese open-source models rather than imposing an outright ban, Axios reported, citing people familiar with the discussions.

The renewed effort comes as Chinese models such as Kimi gain traction among developers and enterprises seeking lower-cost alternatives to leading American systems, intensifying a debate that now stretches beyond national security into competition within the AI industry.

Officials Explore New Ways to Restrict Chinese AI

According to Axios, the Commerce Department last year considered adding several Chinese AI laboratories to its Entity List, a move that would effectively restrict U.S. access without a government licence.

The National Security Agency and the White House Office of the National Cyber Director also examined issuing guidance warning companies about potential security risks associated with Chinese AI models, while the White House discussed requiring businesses to assume liability if they hosted Chinese models that later created security problems, the report said.

Commerce officials also circulated draft rules that would have used supply chain authorities to target Chinese open-source AI models, according to another person familiar with the discussions.

Those proposals were ultimately abandoned after administration officials argued they could slow innovation. Axios reported that the discussions have since regained momentum as cybersecurity concerns have grown and Chinese AI capabilities have advanced.

Neither the White House nor the Commerce Department responded to Axios’ requests for comment.

Chinese Open Models Gain Ground

The latest discussions come as Chinese developers continue releasing increasingly capable open-weight models that businesses can run on their own infrastructure.

Axios reported that U.S. companies are adopting Chinese models because they cost less while delivering performance that approaches leading American systems.

The latest catalyst has been Moonshot AI’s Kimi, which attracted widespread attention last week after demonstrating capabilities that many developers viewed as competitive with frontier U.S. models.

Unlike proprietary AI systems that require ongoing API subscriptions, open-weight models allow companies to deploy and customise the software on their own hardware, making them attractive for organisations seeking lower operating costs.

David Sacks Warns Against Limiting Open-Source Competition

The proposals have drawn criticism from David Sacks, an outside White House AI adviser, who argued that restricting Chinese models could also reduce competition within the U.S. AI market.

“We are at a critical inflection point in AI policy. The leading closed labs, already a duopoly in terms of AI model revenue, want the government to eliminate their open-source competition,” Sacks wrote on X on Sunday.

Sacks has previously argued that regulation should not entrench dominant AI companies or discourage open-source development.

The report also cited a source familiar with the discussions who said representatives of leading AI companies or their allies approach the administration every three to five months with proposals aimed at restricting open-source models.

OpenAI and Anthropic have publicly supported AI safety frameworks, including licensing proposals in some cases, although neither company has called publicly for a ban on Chinese open-source models.

Pressure May Come Without a Formal Ban

People familiar with the discussions told Axios the administration may not need to prohibit Chinese AI outright to reduce its use in the United States.

Instead, officials could rely on procurement policies, Entity List designations, government advisories and liability requirements to discourage businesses from deploying Chinese models.

Another source described the approach as an effort to highlight potential security risks, including possible backdoors and governance concerns, while encouraging wider adoption of American open-source alternatives.

AI Policy Is Shifting Beyond Chips

The discussions suggest Washington’s AI strategy is expanding beyond export controls on advanced semiconductors toward the software itself.

Until recently, U.S. policy largely focused on restricting China’s access to advanced chips needed to train frontier AI systems.

The latest proposals target the deployment of completed AI models instead, reflecting growing concern inside Washington over the adoption of Chinese artificial intelligence technology by American businesses.

Whether those discussions ultimately result in formal restrictions remains unclear. But according to reports that momentum behind tougher measures has increased as Chinese models become more capable and gain wider commercial adoption.