Back and Forth in Semiconductor Policy USA Allows Nvidia Export With a 25-Percent Levy

From Manuel Christa | Translated by AI 2 min Reading Time

Related Vendors

The US government plans to allow the export of Nvidia's H200 accelerators to China, but ties the approval to certain conditions. A key point is a 25 percent levy on the revenues. Additionally, only "approved customers" are to be supplied. The details are being worked out by the US Department of Commerce.

China export relaxed again: Nvidia CEO Jensen Huang is likely pleased about this.(Image: Nvidia)
China export relaxed again: Nvidia CEO Jensen Huang is likely pleased about this.
(Image: Nvidia)

The measure is intended to secure the economic interests of the USA without jeopardizing national security. According to Trump's statements, the model applies not only to Nvidia but also prospectively to other US chip manufacturers like AMD and Intel.

For Nvidia, the move is politically and economically significant, which is why massive lobbying efforts were undertaken in Washington. China is one of the most important markets for AI accelerators but has recently been largely cut off due to export restrictions. The company stated that the offering for approved commercial customers vetted by the Department of Commerce creates a "well-considered balance" and is "great for America."

A Compromise Below the Performance Limit

The most powerful generation is not being approved. The H200 is still based on the Hopper architecture and remains below the current Blackwell chips, which are still excluded from export. Nevertheless, the H200 significantly surpasses the trimmed-down H20 models, whose export to China was allowed earlier this year. At that time, Trump demanded a 15 percent government revenue share, which has now increased to 25 percent.

US government officials view the decision as a compromise. A complete export ban, they argue, would strengthen Chinese providers like Huawei, which are developing their own AI accelerators. At the same time, the US wants to prevent the most advanced hardware from being used directly for military or security-critical applications.

Criticism comes from security circles and Congress. They argue that even H200 chips are powerful enough to support military AI projects or surveillance systems. Whether planned customer vetting will effectively limit such risks remains unclear.

Adding to this is the geopolitical context: the U.S. remains heavily reliant on China for critical raw materials, such as rare earths. Observers see the new regulation as a political trade-off, aiming to reduce economic pressure without abandoning the fundamental stance on technology restrictions. For cloud providers and data center operators in China, this would ease access to more powerful U.S. hardware, albeit at higher costs. This shifts the boundaries of what is considered acceptable in trade policy. (mc)

Subscribe to the newsletter now

Don't Miss out on Our Best Content

By clicking on „Subscribe to Newsletter“ I agree to the processing and use of my data according to the consent form (please expand for details) and accept the Terms of Use. For more information, please see our Privacy Policy. The consent declaration relates, among other things, to the sending of editorial newsletters by email and to data matching for marketing purposes with selected advertising partners (e.g., LinkedIn, Google, Meta)

Unfold for details of your consent