In a move that could reshape the global artificial intelligence (AI) landscape, the Trump administration is weighing whether to allow Nvidia, the world’s most valuable semiconductor company, to sell its powerful H200 AI chips to China. The possibility, reported by Reuters and confirmed by several sources familiar with the matter, comes just weeks after President Donald Trump and Chinese leader Xi Jinping struck a trade and technology truce during a high-profile meeting in Busan last month.
The H200, unveiled in 2023, is Nvidia’s latest high-performance GPU based on the Hopper architecture. It boasts more high-bandwidth memory and processing muscle than its predecessor, the H100, and is estimated to be twice as powerful as the H20 chip—the most advanced AI semiconductor currently permitted for export to China under existing U.S. regulations. The H200’s capabilities make it a coveted component for companies and governments racing to train and deploy advanced AI models, and its potential entry into the Chinese market is stirring both excitement and anxiety across the tech and political spheres.
According to Reuters, the White House’s consideration of this policy shift marks a significant departure from the Trump administration’s earlier, more hardline stance on semiconductor export controls. These controls, first imposed in 2022 and later tightened, were designed to curtail China’s access to cutting-edge U.S. technology, especially in fields with military applications. The Export Control Reform Act (ECRA), passed in 2018 during Trump’s previous term, specifically banned exports of advanced AI chips like the H200 to foreign adversaries. Yet, the prospect of a new détente with Beijing—coupled with mounting pressure from industry leaders—has prompted Washington insiders to revisit these restrictions.
Nvidia CEO Jensen Huang has been at the forefront of lobbying efforts to ease export controls. Speaking to Cybernews, Huang remarked that denying China access to advanced Nvidia chips “leaves a massive hole in the market for rapidly growing foreign competitors to fill.” Nvidia further stated, “Our foreclosure from the China data centre compute market has no impact on our ability to supply customers in the USA,” adding that current regulations prevent it from offering a competitive product in China and risk ceding the market to rivals.
The stakes are high on both sides of the Pacific. For China, access to Nvidia’s H200 would be a boon for its ambitious goal to become the world leader in AI by 2030. Court documents from a recent U.S. Department of Justice (DOJ) indictment revealed that China is actively seeking advanced U.S. AI technology to modernize its military, including the development of weapons of mass destruction and sophisticated AI surveillance tools. On November 20, the DOJ charged two American citizens and two Chinese nationals with attempting to illegally export Nvidia H200 chips to China, a scheme allegedly financed by Chinese interests. The four men face up to 50 years in prison if convicted.
These revelations have reignited national security concerns among U.S. lawmakers. The House Select Committee on China, chaired by Republican John Moolenaar, issued a stark warning: “The DOJ’s latest indictment makes it clear: the CCP will stop at nothing to obtain advanced US AI chips. We cannot allow American innovation to fuel Beijing’s military ambitions.” The committee is now pushing for swift passage of the bipartisan CHIP Security Act, which would require AI chips to be equipped with location trackers and mandate reporting of potential diversions to prevent sensitive technology from ending up in the wrong hands.
Despite these warnings, the Trump administration’s approach appears to be evolving. Officials are currently reviewing export policy, with the Department of Commerce’s Bureau of Industry and Security stressing that “plans could change.” Some in the administration view selling the H200 to China as a compromise—less provocative than approving exports of Nvidia’s latest Blackwell chips, which remain off the table for now. The Blackwell B200 and Grace Blackwell GB200 Superchip, released in 2024, represent Nvidia’s most advanced AI offerings and are not under consideration for export to China at this time.
Still, the debate is far from settled. Many lawmakers and security officials on both sides of the aisle oppose any relaxation of export controls, arguing that even incremental concessions could undermine U.S. technological leadership and national security. Just five months ago, Treasury Secretary Scott Bessent testified that the administration had “no intent” to increase China’s access to advanced semiconductors, noting that access to the less-powerful H20 chips had already been curbed. Yet, in a twist emblematic of the shifting landscape, the U.S. later approved limited shipments of H20 chips to China in exchange for a 15% revenue cut—a deal whose legality remains murky.
Meanwhile, Beijing has pushed back against Washington’s restrictions, accusing the U.S. of weaponizing economic and trade issues. Chinese companies, though encouraged by their government to develop domestic alternatives like those from Huawei Technologies, still prefer Nvidia’s hardware for its superior performance. Some Chinese authorities have even discouraged or outright forbidden firms from using the H20 and other China-specific Nvidia products, signaling their dissatisfaction with being relegated to second-tier technology.
The broader context is no less complex. As the U.S. weighs whether to loosen its grip on advanced chip exports, it must also contend with China’s own leverage over critical supply chains—most notably, its dominance in rare earth minerals essential for manufacturing high-tech goods. Earlier this year, Trump threatened new restrictions on tech exports to China in response to Beijing’s assertive use of export controls on these minerals, but ultimately rolled back most of those measures as part of the broader trade détente.
Adding to the intrigue, the Commerce Department recently approved shipments of up to 70,000 Nvidia Blackwell chips to Saudi Arabia’s Humain and the United Arab Emirates’ G42, a move that underscores the global demand for Nvidia’s technology and the strategic calculations underpinning export decisions. During Saudi Crown Prince Mohammed bin Salman’s visit to the White House in November, Nvidia CEO Jensen Huang—whom Trump has described as a “great guy”—was among the high-profile guests, highlighting the company’s growing influence in international affairs.
As of now, no final decision has been made on the H200 exports, and any move will require formal export licenses under U.S. law. The issue remains the subject of heated internal debate, with some administration officials framing the choice as one between the Hopper and Blackwell lines of Nvidia products. Others insist that no additional advanced chips should be sent to China under any circumstances, a view echoed by many in Congress.
For now, the only certainty is uncertainty. The outcome of these deliberations will have far-reaching implications—not just for Nvidia and its shareholders, but for the future of U.S.-China tech competition, global AI development, and the delicate balance between economic opportunity and national security.
