Rolling Back Export Controls, U.S. Offers China Powerful AI Chips

The United States has instituted a pay-to-play scheme for artificial intelligence (AI) exports. On December 8, President Donald Trump announced that the Commerce Department would allow American AI semiconductor manufacturers, including Nvidia, to sell high-end chips to China in exchange for a 25 percent revenue stake. While Trump claimed he discussed the agreement with Chinese paramount leader Xi Jinping, Chinese authorities will reportedly limit the sale of some approved chips while excluding them from domestic subsidy programs.

The deal, which followed prior reported discussions between Trump and Xi over technology controls, will further enhance China’s growing military while imperiling the future of U.S. export controls.

Proposed Deal Allows U.S. Government Share of Revenue From New China Sales

The agreement allows approved Chinese firms to purchase Nvidia’s H200 chip line, a version nearly six times as powerful as the firm’s current export-compliant H20 chip and far superior to current Chinese-manufactured chips. In his announcement, President Trump claimed that other American chipmakers, including Intel and Advanced Micro Devices (AMD), would also gain expanded access to the Chinese market in exchange for a similar 25 percent share of sales revenue to the United States.

The announcement immediately follows the Department of Justice’s arrest of two businessmen as part of Operation Gatekeeper, a long-running federal probe into a Houston-based smuggling ring that allegedly shipped H200s and other high-end chips to Chinese customers in violation of U.S. export controls. In a press release, U.S. Attorney Nicholas J. Ganjei of the Southern District of Texas described the H200 as “the building blocks of AI superiority” and noted that “the country that controls AI technology will control the future.”

China’s AI Sector Remains Weakened but Undeterred

China’s AI sector continues to struggle under the combined weight of U.S. export controls and internal shortfalls. While Chinese domestic chip manufacturers such as Huawei have produced improved alternatives, partially to satisfy intensifying political demands for self-sufficiency, major Chinese AI firms, such as DeepSeek, report struggling to improve their models due to a lack of advanced computing power. This shortage has forced Chinese firms into complex and costly solutions to compete with American models, including pursuing parallel computing — splitting a computational task into smaller pieces to run on multiple chips — and using more energy to run more chips over a longer period.

Nonetheless, Beijing has maintained its commitment to integrating AI into its military modernization efforts under the guise of its concept of “intelligentized warfare.” Having already incorporated DeepSeek models into its military hospital infrastructure, the People’s Liberation Army (PLA) continues to integrate AI into its growing fleet of autonomous weapons, including a series of “robotic wolves” that debuted during an October amphibious landing exercise. While the PLA seeks domestic supplies for these systems, the growth of China’s military AI program has been partially fueled by components sourced from Nvidia, Intel, and other American firms.

U.S. Should Prioritize Fueling its Domestic AI Sector

While it was a strategic mistake for Washington to offer the H20 to China, it would be an even graver error to offer the more powerful chip H200 while Beijing is flexing its military muscle towards regional U.S. allies and partners. Rather than offer a sensible middle ground between Nvidia’s most and least powerful chips, the new deal is both a short and a long-term concession to Beijing — allowing Chinese firms to immediately purchase superior components while setting a deleterious precedent for future negotiations.

Rather than give China relief from export controls, the United States should strengthen the Bureau of Industry and Security’s capacity to investigate regulatory violations while prioritizing sales to U.S. allies and partners. Moreover, the Trump administration should use its newly launched Genesis Mission as an opportunity to improve U.S.-produced open-source models, a domain where China has a growing advantage.

Jack Burnham is a senior 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.