Nvidia has stopped shipping its Hopper-generation artificial intelligence chips to China despite previously receiving approval from the US government to sell certain processors to the country. The decision marks a significant shift for the world’s leading AI chipmaker, which has spent years trying to maintain access to China’s lucrative semiconductor market amid tightening export restrictions and geopolitical tensions.
The company had shipped $2.4 billion worth of licensed Hopper data centre products to China during the third quarter of fiscal 2027. However, Nvidia’s latest outlook excludes further revenue from these approved chip shipments, signalling that sales to China remain increasingly uncertain even when Washington permits certain products to be exported.
Nvidia Ends Hopper Chip Shipments to China
Nvidia’s interim chief executive Colette Kress confirmed that the company was no longer shipping licensed Hopper data centre products to China during its earnings discussion. The company had previously been permitted to sell certain Hopper-generation processors, including the H200, to approved Chinese customers under US export restrictions.
Despite the approval, Nvidia has now stopped shipments of the licensed products. The company did not publicly provide a detailed explanation for the decision, leaving questions about the role of regulatory uncertainty, customer demand, Chinese import restrictions and changing geopolitical conditions.
The development is notable because Nvidia had been actively pursuing opportunities to sell compliant AI chips to China. The Chinese market remains important for AI infrastructure companies, although restrictions on advanced semiconductor exports have significantly limited the products Nvidia can offer there.
Nvidia Previously Sold $2.4 Billion in Licensed Chips
Nvidia’s China business had generated substantial revenue from approved Hopper-generation products. Colette Kress said the company sold $2.4 billion worth of licensed Hopper data centre products to China during the third quarter.
However, the company’s future revenue guidance no longer includes these products. Nvidia is also not assuming data centre computing revenue from China in its fourth-quarter outlook, highlighting the uncertainty surrounding future shipments.
| Nvidia China Chip Business | Details |
|---|---|
| Chip architecture | Hopper |
| Approved product generation | H200 |
| Q3 licensed Hopper sales | $2.4 billion |
| Future shipment status | No longer shipping licensed Hopper products to China |
| China data centre revenue in Q4 outlook | Not assumed |
The shift means that even previously approved chip sales are no longer a dependable source of near-term revenue for Nvidia.
US Export Controls Continue to Tighten
The decision comes as the US government increases scrutiny of advanced AI chip exports and potential technology diversion to China.
Washington has been concerned that Chinese companies could obtain access to restricted computing power through overseas data centres rather than directly importing advanced processors. This has created additional challenges for export-control enforcement because AI chips can be installed in facilities outside China and accessed remotely.
The US Commerce Department recently introduced new restrictions requiring licences for certain advanced chip shipments to countries including Singapore, Malaysia, Thailand, Indonesia and Brazil. The measures are intended to address concerns about the diversion of advanced AI hardware and remote access to computing resources by Chinese companies.
The regulatory environment is becoming more complicated for semiconductor companies, cloud providers and data centre operators. Businesses must increasingly evaluate not only where chips are shipped but also who ultimately accesses the computing infrastructure.
China Is Also Restricting Nvidia Chip Imports
US restrictions are only one part of the challenge facing Nvidia in China. Beijing has also been cautious about approving large-scale imports of American AI processors as it attempts to strengthen its domestic semiconductor industry.
Chinese technology companies continue to require substantial computing power for AI model training and inference. Nvidia’s processors remain attractive because of their performance and software ecosystem, but Chinese authorities have been encouraging companies to rely more heavily on domestic alternatives.
Earlier reports indicated that large quantities of Nvidia H200 chips were waiting in Hong Kong because Chinese authorities had not approved large-scale imports into mainland China. This highlights the conflicting pressures facing the market: Chinese AI companies want access to advanced computing hardware, while the government is attempting to reduce dependence on foreign semiconductor technology.
The situation could leave Nvidia facing restrictions on both sides of the supply chain.
Nvidia Faces a Changing China Strategy
Nvidia has long viewed China as a major potential market for AI accelerators. However, the company’s ability to sell products there has been affected by successive US export-control changes, China’s domestic technology policies and uncertainty surrounding approvals.
The company has previously attempted to develop or offer products that comply with US restrictions while still providing useful AI computing performance to Chinese customers. But the rules have repeatedly changed, making product planning and long-term sales forecasts more difficult.
Nvidia’s latest decision to stop shipping licensed Hopper chips could signal a more cautious approach to the Chinese market as export restrictions, import approvals and national security considerations become increasingly difficult to navigate.
The company is also facing pressure to manage its global supply chain while maintaining its position in the AI infrastructure market, where demand for high-performance processors remains strong.
The Bigger Picture
The halt in Hopper chip shipments highlights how the AI semiconductor market is increasingly shaped by geopolitics rather than just technology and commercial demand.
Nvidia may have government permission to sell certain processors, but regulatory approval alone does not guarantee that shipments will take place. Customer access, import approvals, supply-chain restrictions, data centre rules and national security concerns can all influence whether products ultimately reach a market.
The uncertainty also creates opportunities for domestic Chinese chipmakers. Companies developing AI accelerators could benefit if restrictions limit the availability of Nvidia products, although matching Nvidia’s hardware performance, software ecosystem and developer adoption remains a significant challenge.
For the global AI industry, the situation demonstrates that access to computing power is becoming a strategic issue for governments, technology companies and businesses building advanced AI systems.
Looking Ahead
Nvidia’s decision to stop shipping Hopper chips to China could reshape expectations for its China revenue in the coming quarters. The company has already excluded licensed Hopper sales from its future guidance, suggesting that investors should not assume the previously approved export channel will resume immediately.
The broader US-China technology relationship will remain a major factor in Nvidia’s international business. Further export-control changes, Chinese import restrictions or diplomatic developments could influence whether the company eventually reopens shipments or adopts a different product strategy for the market.
For now, Nvidia’s China business remains caught between strong demand for AI computing and increasingly complex regulatory barriers, leaving the future of advanced chip exports uncertain.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.



