Trump Administration Discards Proposed Framework for Preventing Diversion of AI Tech to China

The Trump administration is scrapping its predecessor’s proposed approach to keeping cutting-edge U.S. chips out of Chinese hands but did not propose a clear-cut alternative. On May 13, the Bureau of Industry and Security (BIS) announced plans to rescind the AI Diffusion Rule — a policy that would have restricted countries’ capacities to purchase advanced computing power based on their willingness to prevent its diversion to China.

The rule’s proposed rigidity would have hampered the capacity of U.S. allies and partners to develop their own indigenous AI sectors — though its rescission will likely allow China to leverage preexisting gaps within current export controls to access American technologies.

BIS Issues New Guidance on AI Exports

While the rescission has not yet been finalized, Under Secretary of Commerce for Industry and Security Jeffery Kessler told BIS officials not to enforce the regulation. In announcing the decision, BIS stated that the AI Diffusion Framework would have undermined “diplomatic relations with dozens of countries by downgrading them to second-tier status” — a reference to the policy’s consequences of capping some U.S. allies and partners’ access to American computing power.

Along with rescinding the rule, BIS issued three new guidance documents on AI-related export controls. These included a warning to Americans and foreign firms against using Chinese AI components — particularly Huawei’s Ascend 910 AI chip line, which allegedly violates U.S. export controls — and a notice of best practices to prevent diversion of U.S. AI exports. Likewise, BIS issued guidance that recommends practices that may be contrary to current export controls, including allowing U.S. nationals to provide assistance in training Chinese AI models — a practice that the Biden administration also restricted.

China’s AI Sector Continues to Benefit From Diverted Components

The rescission of the AI Diffusion Framework comes as Beijing continues to build out its indigenous AI sector while expanding the sector’s global reach. Having become one of Beijing’s rising national champions, AI firm DeepSeek has allegedly benefited from smuggling routes to source illicit Nvidia components and gained access to Huawei’s Ascend chip architecture — highlighting a reliance on both domestic Chinese firms and their foreign competitors. Chinese AI firms have also reaped substantial gains from a range of strategic investments within the Gulf region, particularly Saudi Arabia and the United Arab Emirates, expanding access to restricted AI components, data centers, and university laboratories.

In response, the United States instituted a series of export controls, a campaign that culminated in the release of the framework in January 2025. Beginning in 2022, the Biden administration introduced and tightened export controls on advanced AI chips and related manufacturing equipment, along with components used for supercomputing and mass data collection. These efforts sought to stymie Beijing’s AI industry and erode its contributions to the Chinese military — a trend that has steadily increased in tandem with the country’s military build-up.

Washington Should Issue New Rule to Balance Proliferation With Security

Despite its rigidity, the export controls embedded within the AI Diffusion Rule likely would have made significant progress in preventing China from gaining access to leading AI technologies that Beijing will use to bolster its economy and enhance its military prowess.

In considering its replacement, the Trump administration should strive to construct an effective coalition against China’s burgeoning AI sector. To achieve this goal, Washington should transparently condition the sale of U.S. AI components on customers’ willingness to adopt stringent economic security measures, allowing those that comply to purchase components freely. These should include screening foreign investment, harmonizing export controls, investing in research security, and ensuring customers’ alignment with the foreign policy interests of the United States, thereby allowing AI to proliferate freely while preventing the risk of diversion.

Jack Burnham is a research analyst in the China Program at the Foundation for Defense of Democracies (FDD). For more analysis from Jack and FDD, please subscribe HERE. Follow Jack on X @JackBurnham802. Follow FDD on X @FDD. FDD is a Washington, DC-based, nonpartisan research institute focusing on national security and foreign policy.