President Donald Trump announced a significant policy shift on Monday, allowing Nvidia to sell an advanced type of computer chip crucial for artificial intelligence development to “approved customers” in China. This decision, which Trump communicated to China’s leader Xi Jinping who reportedly responded positively, aims to support American jobs, strengthen U.S. manufacturing, and benefit American taxpayers, while also fostering the development of the AI ecosystem with U.S. companies like Nvidia.

The chip in question, known as the H200, is more advanced than the H20 chips Nvidia specifically designed for the Chinese market but is not the company’s most cutting-edge product; its top-tier chips, Blackwell and the upcoming Rubin, remain excluded from this approval. The Department of Commerce will oversee the vetting of commercial customers for these H200 sales, a measure Nvidia stated would “strike a thoughtful balance” on economic and national security priorities. The U.S. is also set to take a 25% cut of these sales, with Nvidia only able to send H200s that are approximately 18 months old.

Despite the administration's stated benefits, the decision has drawn sharp criticism from a group of Democratic senators, including Chris Coons, Jeanne Shaheen, Jack Reed, Elizabeth Warren, Brian Schatz, Andy Kim, Michael Bennet, and Elissa Slotkin. These senators voiced serious concerns that access to such chips would provide China’s military with “transformational technology,” potentially making its weapons more lethal, enhancing cyberattacks against American businesses and critical infrastructure, and strengthening China’s economic and manufacturing sector. They highlighted a recent statement from Chinese AI company DeepSeek, which identified the lack of access to advanced American-designed chips as their biggest challenge in competing with U.S. AI leaders like OpenAI, Google, and Microsoft.

The policy announcement comes amidst a complex history of U.S. restrictions and negotiations regarding advanced chip exports to China. The Biden administration previously sought to limit these exports due to national security concerns. President Trump himself had a fluctuating stance, initially imposing licensing requirements on chip companies like Nvidia in April, rescinding a Biden administration diffusion rule in May, and later signaling that companies could send chips to China for a 15% revenue cut as a trade bargaining tool. Congressional efforts, such as the Secure and Feasible Exports Act (SAFE) Chips Act introduced by Republican Senator Pete Ricketts and Democratic Senator Chris Coons, aimed to block advanced AI chip exports to China for over two years, requiring the Commerce Department to deny licenses for 30 months. The fate of this bill remains uncertain following Trump’s green light for H200 sales.

Adding to the complexity, China’s Cyberspace Administration banned domestic companies from buying Nvidia’s chips in September, pushing them to rely on less advanced domestic alternatives from Alibaba and Huawei. Trump’s decision also reflects the increasing influence of Nvidia’s founder and CEO, Jensen Huang, within the administration. Following Trump’s announcement, Nvidia, which boasts a market capitalization of $4.5 trillion, saw a slight uptick in its stock in after-hours trading. The Commerce Department is reportedly finalizing details for other chipmakers, like AMD and Intel, to sell their technologies abroad as well, indicating a broader shift in export policy.