TRENDING

Nvidia’s Chips Have Long Been a Bargaining Chip in the U.S.-China Trade War—Until Now

The problem isn’t that Nvidia opposes the Chinese trade embargo. It’s that China no longer wants to rely on its GPUs for AI development.

China wants to de-Americanize its AI
No comments Twitter Flipboard E-mail
javier-pastor

Javier Pastor

Senior Writer
  • Adapted by:

  • Karen Alfaro

javier-pastor

Javier Pastor

Senior Writer

Computer scientist turned tech journalist. I've written about almost everything related to technology, but I specialize in hardware, operating systems and cryptocurrencies. I like writing about tech so much that I do it both for Xataka and Incognitosis, my personal blog.

263 publications by Javier Pastor
karen-alfaro

Karen Alfaro

Writer

Communications professional with a decade of experience as a copywriter, proofreader, and editor. As a travel and science journalist, I've collaborated with several print and digital outlets around the world. I'm passionate about culture, music, food, history, and innovative technologies.

538 publications by Karen Alfaro

Nvidia engineers have nearly finished developing a GPU with Blackwell microarchitecture for one purpose: replacing the H20 chip they can’t sell in China. The U.S. Department of Commerce banned such exports, so Jensen Huang’s company moved to create a chip tailored to government requirements. TSMC will begin manufacturing the GPU later this month, but the effort may be futile. Nvidia isn’t facing mistrust from U.S. lawmakers. Instead, China is turning its back on the company.

According to the South China Morning Post, the chip for the Chinese market isn’t ready. At last week’s investor conference, Huang said, “The key is to understand the limits and see if we can come up with interesting products that could continue to serve the Chinese market. We don’t have anything [to announce] at the moment, but we’re considering it.”

Chinese President Xi Jinping Wants to Eliminate Reliance on U.S. Software and Hardware—And He’s Making Progress

Huang made a surprise trip to China in mid-April after the U.S. announced the export veto, meeting with senior Chinese officials to show his commitment to the market. The chip Nvidia is preparing for China will be significantly cheaper than the H20, though Huang clarified it won’t be based on the same microarchitecture.

This effort makes sense for Nvidia. During the last fiscal year, which ended Jan. 26, 2025, China accounted for about 13% of Nvidia’s global revenue—roughly $17 billion. China ranks as Nvidia’s third-largest customer, behind the U.S. and Taiwan. However, U.S. sanctions pose a serious threat to Nvidia’s business in China. A second, more troubling threat has emerged.

During a session with the Chinese Politburo, President Xi made it clear that China aims to “de-Americanize” its AI technology stack.

Meanwhile, China is working hard to cut its dependence on the U.S. in this sector. Dozens of companies are designing GPUs for gaming and AI, and the Chinese government is pouring subsidies into the effort. These companies include Huawei, MetaX, Biren Technology, Moore Threads, Innosilicon, Zhaoxin, Iluvatar, oreX, DenglinAI, Lisuan Technology, and VastAI Tech.

The problem for Huang and Nvidia is that President Xi emphasized China’s goal to “de-Americanize” its AI stack. All AI development—software and hardware—must rely on domestic solutions. No matter how Nvidia tailors its chips for the Chinese market, China’s top priority is self-reliance in AI development. And that spells trouble for Nvidia.

Image | Taiwan Presidential Office

Related | Samsung Wants to Overtake TSMC. Its Best Shot Is Convincing Nvidia and Qualcomm That Its 2 nm Tech Is Superior

Home o Index