According to a recent statement by President Donald Trump, the U.S. government has officially granted Nvidia authorization to begin exporting its H200 artificial intelligence GPU processors to a carefully selected group of “approved customers” in China, as well as in certain other international markets. This approval, issued directly by the White House, represents a significant policy development at the intersection of advanced computing technology and global trade. In a post shared on Truth Social on Monday, President Trump explained that this new export arrangement is not without restrictions: it is accompanied by specific national security conditions designed to ensure that the United States maintains a solid defensive and technological advantage. Furthermore, as part of the agreement, the U.S. government will claim a 25 percent share of Nvidia’s GPU sales revenues under the export program—an increase over the 15 percent rate originally proposed back in August, reflecting Washington’s intention to capture greater economic value from technology exports deemed strategically sensitive.
The H200 GPU chips themselves represent a substantial technical advance compared to Nvidia’s H20 processors, which the company had already been permitted to sell in China under existing export limitations. Although the H200 line provides enhanced performance and efficiency, it still falls short of the computational capabilities of Nvidia’s newest flagship architecture, known as the Blackwell GPUs. This distinction is crucial, since U.S. policymakers have been closely regulating the export of the most cutting-edge chips to ensure that America’s technological leadership in artificial intelligence is not compromised. President Trump emphasized the broader goals of the decision, stating that the policy would bolster American employment, fortify domestic manufacturing, and channel additional revenue to taxpayers. He further noted that Nvidia’s U.S.-based clients are already advancing their operations using the company’s ultra-sophisticated Blackwell and upcoming Rubin chips—both of which remain outside the bounds of this China-focused export authorization.
This policy announcement follows earlier steps taken by the Trump administration, which in May opted to ease some of the strict export controls that had previously constrained the AI semiconductor industry. Still, the issue continues to generate significant political debate. A bipartisan group of senators recently urged the President to sustain restrictions on the sale of Nvidia’s most powerful processors to China, arguing that unrestrained access could erode the United States’ dominant position in the global artificial intelligence sector. Despite these warnings, the new round of approvals signals an acknowledgment of Nvidia CEO Jensen Huang’s persistent engagement with U.S. policymakers and his efforts to align the company’s interests with national priorities. However, the practical success of these exports will depend largely on China’s willingness and ability to purchase such chips—an outcome complicated by Beijing’s own prohibitions on acquiring advanced Nvidia hardware in favor of promoting indigenous semiconductor development through an expanding network of domestic manufacturers. In this complex environment, Nvidia’s newly authorized exports constitute not only a commercial opportunity but also a strategic gesture amid the ongoing global contest for AI and technological leadership.
Sourse: https://www.theverge.com/news/840557/nvidia-h200-ai-chips-china-exports-us
