On Wednesday, July 23, 2025, the White House released the “America’s AI Action Plan” (the “AI Action Plan”). The AI Action Plan is the result of the Trump Administration’s Request for Information, which describes policy actions related to artificial intelligence as it seeks to replace the Biden Administration’s “AI Diffusion Rule.” The AI Action Plan offers a broader strategy to utilize export controls to address the evolving AI landscape, promote the U.S. as the dominant player on the world stage, and constrain foreign adversaries of the United States, namely China.
The AI Action Plan reexamines the role U.S. export controls play to achieving these ends. Centered within Pillar III of the AI Action Plan, exports and export controls are highlights in four key policy principles:
- Export American AI to Allies and Partners: The goal is to export the full AI technology stack (hardware, models, software, applications and standards) to those countries willing to join “America’s AI alliance,” which is not specifically defined in the AI Action Plan. To achieve this goal, Commerce will coordinate with other relevant U.S. government agencies to facilitate deals that “meet U.S.-approved security requirements and standards,” likely akin to those highlighted in President Trump’s recent trip to the Middle East. These security requirements also are likely targeted to achieve the same goals described in the May 13, 2025 policy statements BIS issued, namely, to promote the adoption of American AI while protecting U.S. national security interests.
- Strengthen AI Compute Export Control Enforcement: This goal is to ensure that the United States has appropriate controls in place to deny foreign adversaries and countries of concern access to the most advanced chips. The AI Action Plan tasks Commerce (with the White House and industry) to identify location verification features to impose on advanced chips used for AI purposes and use those to ensure the chips are not sent to prohibited countries. Similarly, it tasks Commerce to collaborate with the intelligence community to monitor emerging technology developments in order to identify diversion points, using this information to expand and increase end-use monitoring in countries with high diversion risk.
- Plug Loopholes in Existing Semiconductor Manufacturing Export Controls: Here, to prevent adversaries from receiving this same technology, the AI Action Plan discusses both the need to address gaps in current semiconductor manufacturing export controls and to enhance enforcement in this space. Here, Commerce is meant to “develop new export controls on semiconductor manufacturing subsystems.”
- Align Protection Measures Globally: The last export controls-focused priority discusses the importance of encouraging partners and allies to follow U.S. export controls and preventing those countries from backfilling orders, if the United States imposes unilateral controls, by threatening imposition of the foreign direct product rule or the use of “secondary tariffs.” The AI Action Plan discusses five recommended policy actions. These include: (i) Commerce and State (with other agencies) developing, implementing, and sharing information on technology protection measures (including in the basic research and higher education space); (ii) developing a whole-of-government strategy to induce “key allies” (likely Japan, the Netherlands, and South Korea) to adopt similar controls across the full supply chain; (iii) expanding new initiatives for plurilateral controls for the AI tech stack, without relying on current multilateral treaty bodies; (iv) coordinating to ensure other countries adopt U.S. controls and continue to work with the U.S. to develop new controls as needed; and (v) “prohibit U.S. adversaries from supplying their defense-industrial base or acquiring controlling stakes in defense suppliers.”
On the whole, these principles remain under development, with the AI Action Plan charging various agencies with implementing the direction therein. As with other strategic-level documents, much remains to be seen in the agency-level rollout of these directives.
Open questions notwithstanding, the overall direction seems clear: moving forward, the U.S. will include additional export controls on AI related materials and information, particularly in critical locations in the AI technology supply chain, but will also utilize additional enforcement mechanisms and add pressure via “secondary tariffs” when other nations fail to cooperate.
Next steps:
- Watch for continued opportunities to engage in the rollout of these principles, e.g., Commerce will be engaging a “consortia” that will inform its “deal” proposals.
- Multi-jurisdictional export compliance will become more important as other countries begin to adopt complementary controls to the United States, such as in Malaysia
- BIS will impose new export controls, building upon those controls from October 2022 and 2023 on the semiconductor supply chain, particularly focused on subsystems. Companies should watch for these controls and begin to prepare for them by analyzing their technologies, supply chains, and customer bases.
- Though BIS will impose new controls, the AI Action Plan focuses heavily on preventing diversion of these items. Companies should ensure they conduct appropriate “know your customer” diligence, to minimize risk of illegal and unauthorized diversion.