Nvidia will stop selling its H200 chips in China, acting CEO Colette Kress said Tuesday. The decision ends the company’s use of a disputed route for selling advanced chips to the country despite US export restrictions.
Kress made the announcement on Nvidia’s first earnings call since CEO and co-founder Jensen Huang stepped down last week. On the same day, Nvidia said it was forming a board-level committee to oversee national security policies, export control compliance and global trade risks.
The H200 deal allowed sales through the US
Huang and President Donald Trump brokered the H200 arrangement. It required chips bound for China to pass through the US, where the government could collect a 25% tariff. Critics argued that the route gave Chinese labs computing power that could narrow the US lead in AI. Huang argued that US companies should compete in China. In April, he told podcaster Dwarkesh Patel that “to concede that market for the United States technology industry is a disservice to our country.”
Huang’s ouster came in part over an affair with the CEO of a company under investigation for smuggling chips into China. It coincided with rules issued Monday that bar advanced US chips from foreign data centers serving Chinese customers remotely.
Investment bank Jefferies forecast that the new rules could reduce Nvidia’s data center revenue next year by $44 billion to $62 billion. Kress said a backlog of demand for advanced chips such as Vera Rubin means the export controls will have no effect on Nvidia’s top-line revenue.

China and overseas data centers face restrictions
Nvidia could lose potential customers in Singapore, Malaysia, Thailand and Indonesia over the longer term. Critics of its reliance on China say as much as 20% of its revenue comes from demand linked to the country. Annual revenue at Nvidia’s Singapore subsidiary reached $104 billion in 2025 before the company stopped disclosing those figures.
China had largely declined H200 deliveries even before Nvidia ended sales. Jacob Feldgoise, a researcher at Georgetown’s Center for Security and Emerging Technology, said the government prefers to build demand for domestic chips, even though major AI computing users want Nvidia’s still-superior equipment.
Chinese frontier labs have also found access to Nvidia chips elsewhere. All of China’s leading AI models show evidence of training on Nvidia Grace Blackwell chips, which cannot be exported to China. That training may have happened at data centers abroad. Under the new rules, large data centers seeking Nvidia chips for AI training will have to certify that they serve no Chinese customers, export control attorney Dan Fisher-Owens told TechCrunch. Facilities unable to do so may turn to other chips, further reducing Nvidia’s market share.



