Nvidia ends China H200 sales as new U.S. export rules widen revenue risk
Nvidia is pulling out of H200 chip sales in China as the company reshapes its approach to export controls after a leadership upheaval. The move comes alongside tighter U.S. restrictions on advanced chips serving Chinese customers remotely, raising new questions about the durability of Nvidia’s Asia-linked data center revenue.
Highlights
- Nvidia will stop selling H200 chips in China following new U.S. export rules, closing a sales route that bypassed prior restrictions.
- Jefferies estimates updated export measures could reduce Nvidia’s data center revenue by $44 billion to $62 billion next year due to restricted chip access.
- Nvidia’s Singapore subsidiary reported $104 billion in annual revenue for 2025, raising concerns that new rules could significantly alter the company’s Asia revenue exposure.
Policy shift follows leadership change
As first reported by TechCrunch, interim Chief Executive Colette Kress said on Tuesday that Nvidia will no longer sell its H200 chips in China, closing off a route that had allowed the company to keep participating in the market despite U.S. restrictions on advanced semiconductor exports.The announcement comes during Nvidia’s first earnings call since Chief Executive and co-founder Jensen Huang stepped down last week. On the same day, the company also says it is creating a board-level committee to oversee national security policies, export control compliance and global trade risks.
Nvidia’s H200 sales pathway into China had been arranged by Huang and President Donald Trump, with shipments routed through the U.S. so the government could collect a 25% tariff. The arrangement had drawn criticism from China hawks who argued that Chinese laboratories should be denied access to high-end computing power, while Huang had maintained that continued U.S. competition in China was strategically important.
Revenue outlook faces broader Asia exposure
Huang’s departure coincides with new rules issued on Monday that bar advanced U.S. chips from foreign data centers serving Chinese customers remotely. Jefferies estimates those measures could cut between $44 billion and $62 billion from Nvidia’s data center revenue next year.Kress says the company does not expect the latest export controls to affect top-line revenue because of strong backlog demand for advanced chips such as Vera Rubin. Over the longer term, however, Nvidia may face weaker demand from customers in Singapore, Malaysia, Thailand and Indonesia if their business is tied to Chinese end users.
Critics of Nvidia’s exposure to China argue that as much as 20% of company revenue is linked to demand connected to the country. Nvidia’s Singapore subsidiary posted $104 billion in annual revenue in 2025 before the company stopped disclosing those figures, suggesting that the new policy stance could materially change the regional revenue mix.
Earlier, we outlined Nvidia Singapore revenue as scrutiny grew over invoicing practices and export controls tied to its Asian business.