Acting CEO Colette Kress announced Tuesday that NVIDIA will no longer sell its H200 chips in China, ending the company’s participation in a controversial loophole in the U.S. ban on sales of advanced chips to the country.
China’s access to a major AI accelerator was a key topic in the company’s first earnings call since CEO and co-founder Jensen Huang resigned last week. The disclosure came on the same day that Nvidia announced it would create a board-level committee to oversee national security policy, export control compliance, and global trade risks.
NVIDIA’s deal to sell H200 chips in China was brokered by Huang and President Donald Trump. The agreement required the chips to be shipped through the United States and allowed the government to collect a 25% tariff.
This route has angered Chinese hawks who deny the China Institute’s computing capabilities and want to maintain America’s lead in AI, but Huang has long argued that it is more important for the United States to compete in China. In April, he told podcaster Dwarkesh Patel that “allowing a market for the U.S. technology industry would be a disservice to our country.”
Huang’s dismissal, in part due to an affair with the CEO of a company under investigation for smuggling chips into China, coincided with new rules issued on Monday banning the import of advanced U.S. chips from overseas data centers that remotely serve Chinese customers.
Investment bank Jefferies estimates these rules could cut Nvidia’s data center revenue by $44 billion to $62 billion next year. Cress cited the new export restrictions and said the unmet demand for advanced chips like Vera Rubin meant the restrictions would not impact the company’s sales.
In the long term, the company could face the loss of potential customers in Singapore, Malaysia, Thailand and Indonesia. Critics of Nvidia’s dependence on China claim that as much as 20% of its revenue comes from China-related demand (in 2025, before Nvidia stopped disclosing the numbers, the Singapore subsidiary’s annual revenue was $104 billion), but things are clearly changing.
Even before Nvidia decided to stop shipping the H200, China had largely refused to accept H200 deliveries. Jacob Feldgois, a researcher at Georgetown’s Center for Security and Emerging Technologies, said the government wants to create demand for a domestic chip industry, even if major AI computing users still want better Nvidia gear.
Meanwhile, China’s Frontier Research Institute is finding ways to access Nvidia chips elsewhere. All major Chinese AI models show evidence of being trained on Nvidia Grace Blackwell chips, which are banned from export to China.
That training may have taken place in data centers outside China, but the new rules would prohibit sales to such facilities. If large data centers suitable for training AI models want to buy Nvidia chips, they will need to prove they don’t serve customers in China, export control attorney Dan Fisher Owens told TechCrunch.
If it can’t provide that certification, new rules could force other chip options and further erode Nvidia’s market share.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
