President Donald Trump announced on Monday that he will allow Nvidia to sell its advanced H200 AI chips to “approved customers” in China, marking a significant shift in U.S. semiconductor export policy. The decision, shared via a Truth Social post, comes amid growing pressure on the U.S. to balance national security concerns with the economic interests of American tech companies. Trump stated that the U.S. will take a 25% cut of all chip sales to China, a move he framed as a way to benefit American industry while maintaining national security. The approval applies only to the H200 chip, a high-performance processor introduced in 2022 with 144 GB of HBM3 memory, which is well-suited for training large AI models. It does not include Nvidia’s more advanced Blackwell or upcoming Rubin systems, which remain restricted under existing export controls. Trump claimed he informed Chinese President Xi Jinping of the decision, noting a “positive” response from Beijing. He emphasized that sales will proceed under strict conditions to ensure national security, with final details being worked out by the Department of Commerce. The administration also signaled that the same approach—requiring a government revenue share—will apply to other U.S. chipmakers like AMD and Intel, continuing a strategy first proposed in August to incentivize exports while securing federal benefits. Nvidia CEO Jensen Huang has long advocated for loosened export restrictions, arguing that strict controls have backfired by accelerating China’s development of domestic AI chips. Huang met with Trump last week to discuss the issue, underscoring the strategic importance of maintaining Nvidia’s dominance in the Chinese market. The H200, though not the latest generation, still outperforms the H20 chip—specifically designed to comply with U.S. export rules—and remains highly desirable for Chinese AI firms. Despite Huawei’s growing capabilities in AI hardware and software, many Chinese companies continue to favor Nvidia’s chips due to their deep integration with the widely used CUDA ecosystem. The potential approval of H200 exports, reported by Semafor and confirmed by sources familiar with the matter, could strengthen Nvidia’s market position in China and reinforce its software dominance. However, the final outcome depends on whether Chinese authorities allow the chips to enter the country, as Beijing previously blocked imports of the H20 despite its compliance with U.S. restrictions. This move reflects a broader debate over U.S. tech policy: while both Democratic and Republican administrations have restricted advanced AI chip exports to China on national security and economic competition grounds, critics argue such measures push China toward self-reliance and weaken U.S. influence. By allowing controlled sales with revenue-sharing, Trump’s approach aims to strike a balance—boosting American tech firms while limiting strategic risks. Nvidia, now the most valuable publicly traded company in the U.S. with a market cap near $4.5 trillion, has become central to the global AI race. Its chips power much of the world’s AI infrastructure, and access to China remains a critical market. The company welcomed the announcement, calling it a “thoughtful balance that is great for America.” While the H200 is not the most cutting-edge chip, its capabilities make it a powerful tool for AI development. The decision could reshape the AI landscape in Asia and signal a more pragmatic approach to export controls in the age of artificial intelligence.