Skip to content
Theme
OnDevice Tools

Tech BusinessChips & Semiconductors

Nvidia ends H200 chip sales to China after export rules

Nvidia has stopped selling H200 AI chips in China, according to TechCrunch, as new US rules restrict advanced chips and data centres serving Chinese customers.

Illustration of an Nvidia H200 chip with a China export restriction notice

Key takeaways

  • Nvidia will stop selling H200 chips in China, acting CEO Colette Kress said.
  • New US rules restrict foreign data centres serving Chinese customers remotely.
  • Jefferies forecasts a $44 billion to $62 billion revenue hit, but Nvidia disputes it.
  • TechCrunch reports Chinese AI models show signs of training on restricted Nvidia chips.
In this article
  1. Nvidia stops H200 sales in China
  2. Export rules and revenue forecasts
  3. China’s importance to Nvidia
  4. Overseas data centres face new limits
  5. What it means for Nvidia and buyers

Nvidia stops H200 sales in China

Nvidia H200 China sales have ended, according to TechCrunch, after acting chief executive Colette Kress said on 17 November 2026 that the company will stop selling H200 processors in China. The decision closes a limited route that had allowed Nvidia to supply advanced AI chips despite US export restrictions. The H200 had been central to Nvidia’s effort to keep selling advanced AI hardware in China.

The announcement came during Nvidia’s first earnings call with investors after Jensen Huang stepped down as chief executive last week. TechCrunch also reported that Huang left his role partly because of a personal relationship with the head of a firm under investigation in a chip-smuggling case. Nvidia said it was forming a board committee to review national security policy, export-control compliance and global trade risks.

Export rules and revenue forecasts

Illustration: market trend, update, NVIDIA logo, AI and the labels AI Chips, China, Export Controls

The H200 sales arrangement had been negotiated by Huang and President Donald Trump. It required shipments to pass through the United States, so the government could collect a 25% tariff. The route drew criticism from US officials who wanted to deny Chinese laboratories access to advanced computing power. Huang had argued that American technology firms should compete in China rather than withdraw.

The arrangement had allowed Nvidia to keep selling a high-end AI chip while giving the US government a tariff on each shipment. It also highlighted a policy divide between officials focused on blocking China’s AI progress and those who argued that US companies should remain present in the market.

New US rules issued on Monday go further. They block advanced American chips from foreign data centres that serve Chinese customers remotely. Jefferies forecast that the measures could reduce Nvidia’s data centre revenue by $44 billion to $62 billion next year. Kress rejected that view, saying a backlog of orders for newer chips, including Vera Rubin, means the rules will not reduce total revenue.

China’s importance to Nvidia

Nvidia has faced criticism over its exposure to China. Some critics say up to 20% of its revenue is linked to demand from the country. They point to Nvidia’s Singapore subsidiary, which reported $104 billion in annual revenue in 2025 before the company stopped disclosing that figure.

China had already been reluctant to accept H200 deliveries. Jacob Feldgoise, a researcher at Georgetown University’s Center for Security and Emerging Technology, said the Chinese government prefers to build demand for its domestic chip industry, even though major AI users still want Nvidia equipment. Nvidia may also lose potential customers in Singapore, Malaysia, Thailand and Indonesia over the longer term.

Overseas data centres face new limits

Chinese AI laboratories have been finding ways to use Nvidia chips outside China. TechCrunch reports that leading Chinese AI models show signs of training on Nvidia Grace Blackwell chips, which are prohibited from export to China. That training may have taken place in data centres outside China.

The new rules target those arrangements. Export control attorney Dan Fisher-Owens told TechCrunch that large data centres suited to AI training must confirm they do not serve Chinese customers if they want to buy Nvidia chips. If they cannot provide that confirmation, they may choose other chip suppliers. That could reduce Nvidia’s market share in the AI data centre market.

What it means for Nvidia and buyers

The immediate financial impact is disputed. Nvidia says strong demand for its latest products will offset any loss from China-related sales. Analysts at Jefferies see a possible revenue reduction, while critics warn that Nvidia’s dependence on China-linked demand could fall sharply. The company’s comments suggest it expects newer products to absorb demand. However, losing China-linked customers could still alter its long-term growth path.

For cloud providers, governments and AI developers, the rules may make access to advanced Nvidia hardware more complicated. Data centre operators that serve Chinese customers, even remotely, may need to change compliance processes or source chips elsewhere. Nvidia’s new board committee shows that export controls are now a central governance issue for the company.

Sources

  1. Nvidia is abandoning China after Jensen Huang’s ousterTechCrunch

Tags

Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.