Introduction

Nvidia is pulling its H200 AI accelerator from the Chinese market, according to acting chief executive Colette Kress. The move shuts down a politically sensitive arrangement that let the company ship advanced chips to China through the United States while the federal government collected a 25 percent tariff. The announcement arrived during Nvidia’s first earnings call since co-founder Jensen Huang stepped down, and on the same day the chipmaker stood up a new board committee to manage national security, export compliance, and international trade risk.

What Happened

Kress confirmed Tuesday that Nvidia will stop selling H200 processors in China. The sales channel had been negotiated by Huang alongside President Donald Trump and relied on routing hardware through U.S. soil. Critics in Washington viewed the setup as a loophole that undermined export restrictions meant to keep cutting-edge computing power out of Chinese hands.

The timing is significant. Huang’s departure follows reports linking him to a personal relationship with the chief executive of a firm under investigation for chip smuggling into China. Separately, the U.S. government issued new rules on Monday that prevent advanced American chips from being installed in foreign data centers that provide remote computing access to Chinese customers.

Why This Matters

The decision removes one of Nvidia’s most visible workarounds in a market that has long been central to its growth. Investment bank Jefferies projects the fresh export rules could cut between $44 billion and $62 billion from next year’s data center revenue. Kress pushed back on that outlook, pointing to strong backlog demand for upcoming products such as Vera Rubin and arguing the restrictions will not dent overall revenue.

Beyond the immediate financial hit, Nvidia may lose future customers across Southeast Asia, including Singapore, Malaysia, Thailand, and Indonesia. Observers have estimated that as much as 20 percent of the company’s revenue has been tied to China-related demand, though Nvidia stopped publishing Singapore subsidiary figures after they reached $104 billion in 2025.

Key Takeaways

  • Nvidia will no longer sell H200 chips in China, ending a tariff-based shipping arrangement.
  • The halt was disclosed by acting CEO Colette Kress during the first earnings call since Jensen Huang’s departure.
  • A new board-level committee will oversee export controls, national security policy, and global trade exposure.
  • New U.S. rules block advanced chips from foreign data centers serving Chinese customers remotely.
  • Jefferies estimates the rules could remove $44 billion to $62 billion from next year’s data center revenue.
  • Kress said demand for Vera Rubin and other future chips should protect total revenue despite the curbs.
  • Chinese AI labs have reportedly trained leading models on prohibited Grace Blackwell hardware outside China.

Conclusion

Nvidia’s exit from H200 sales in China marks a sharp turn for a company that once argued the U.S. could not afford to cede the Chinese market. With Huang gone and Washington tightening the screws on cross-border AI infrastructure, the chipmaker is betting that demand for its next-generation processors will outweigh the loss of one of the world’s largest technology markets. How that bet pays off will depend on whether foreign data centers can satisfy new certification requirements and whether Chinese buyers can continue accessing Nvidia silicon through indirect channels.