Nvidia H200 China sales at zero despite US export clearance
Synopsis
Nvidia has earned zero dollars from H200 chip sales to China despite holding valid US export licences — a dual-gatekeeper impasse laid bare by CFO Colette Kress even as the company posted a record US$81.6 billion quarter.
Key Takeaways
Nvidia confirmed zero revenue from H200 chip sales to China as of 21 May 2026 , despite Washington having approved export licences.
CFO Colette Kress said the company is 'uncertain whether any imports will be allowed into the country,' and excluded China data centre revenue from its current-quarter outlook.
Quarterly revenue reached a record US$81.6 billion for the period ended April 26 , up 85 per cent year on year, beating the US$79.2 billion consensus.
Data centre revenue hit US$75.2 billion , up 92 per cent year on year, led by Blackwell GB300 product ramp-ups.
CEO Jensen Huang identified a growing 'second category' of AI infrastructure customers — including sovereign AI projects and enterprises — contributing US$37.4 billion in data centre revenue.
US President Donald Trump stated that Beijing had not approved H200 purchases despite existing US export clearance, highlighting a dual-approval bottleneck.
Nvidia has confirmed it has generated zero revenue from H200 chip sales to China, even after Washington approved export licences for the product, highlighting how geopolitical friction continues to block the world's leading AI chipmaker from one of its most coveted markets. The disclosure came on Wednesday, 21 May 2026, during a post-earnings call following a record-breaking quarterly report.
CFO flags deep uncertainty on China imports
Colette Kress, Nvidia's Executive Vice-President and Chief Financial Officer, said the company had 'yet to generate any revenue, and we are uncertain whether any imports will be allowed into the country.' Consistent with the prior quarter, Nvidia excluded all China data centre compute revenue from its current-quarter outlook, signalling no near-term expectation of a breakthrough.
The remarks follow a high-profile visit to China by Nvidia CEO Jensen Huang alongside US President Donald Trump. Trump subsequently stated that Beijing had not approved purchases of the H200, despite existing US export clearance — underscoring the dual-gatekeeper dynamic now shaping advanced chip trade.
Record earnings overshadow market access concerns
Nvidia's financial results for the quarter ended April 26 nonetheless beat analyst expectations across the board. The Santa Clara-based company posted revenue of US$81.6 billion, up 85 per cent year on year and 20 per cent sequentially, surpassing the market consensus of US$79.2 billion. Non-GAAP earnings per share came in at US$1.87, above the US$1.77 analysts had forecast.
Data centre revenue hit a record US$75.2 billion, a 92 per cent year-on-year surge, driven by the ramp-up of Blackwell GB300 products alongside strong demand for InfiniBand, Spectrum-X Ethernet, and NVLink solutions.
Hyperscalers and a rising second tier of AI customers
Under Nvidia's updated reporting framework, hyperscale customers contributed US$37.9 billion — roughly half of data centre revenue — while AI clouds, industrial firms, and enterprises accounted for US$37.4 billion. Huang described the latter as a fast-growing 'second category' of AI infrastructure buyers, encompassing AI-native clouds, sovereign AI projects, and companies building purpose-built AI factories outside public cloud environments.
He said the company's growth was 'no longer tied to a handful of hyperscale cloud providers,' pointing to a more fragmented but rapidly expanding customer base as a structural shift in how AI compute is procured globally.
The competitive backdrop
China's continued exclusion from Nvidia's revenue base gives domestic rivals — and alternative suppliers — room to deepen their foothold in the world's second-largest economy. Beijing's push to develop indigenous semiconductor alternatives remains a stated policy priority, and the H200 impasse effectively accelerates that timeline by keeping leading-edge US chips out of reach for Chinese AI developers.
What's next
All eyes will be on whether Beijing moves to formally approve or block H200 imports in the coming months, a decision that could materially shift Nvidia's addressable market. Any diplomatic progress — or further deterioration — in US-China tech relations will be the primary variable to watch for the company's next-quarter guidance.
Point of View
But the more consequential signal is that Beijing now wields an effective veto over US chip exports — a leverage point that accelerates China's domestic semiconductor ambitions by keeping Nvidia's best hardware perpetually out of reach. With hyperscalers already contributing roughly half of Nvidia's data centre revenue, the company's diversification into sovereign AI and enterprise 'AI factories' reads less like a growth story and more like a hedge against both China exclusion and hyperscaler concentration risk. The next inflection point is not an earnings call — it is a diplomatic one.
NationPress
5 Jul 2026
Frequently Asked Questions
Has Nvidia sold any H200 chips to China?
No. Nvidia has generated zero revenue from H200 sales to China as of 21 May 2026 . CFO Colette Kress confirmed on a post-earnings call that the company remains 'uncertain whether any imports will be allowed into the country,' even though Washington has approved export licences.
Why can't Nvidia sell H200 chips in China if the US approved licences?
Beijing has not approved the import of H200 chips despite US export clearance, creating a dual-gatekeeper bottleneck. US President Donald Trump confirmed that Beijing had not approved purchases after accompanying Jensen Huang on a visit to China , illustrating that both governments must act for trade to proceed.
How did Nvidia perform financially in Q1 2026?
Nvidia reported record quarterly revenue of US$81.6 billion for the period ended April 26 , up 85 per cent year on year and ahead of the US$79.2 billion analyst consensus. Non-GAAP earnings per share of US$1.87 also beat the US$1.77 forecast.
What drove Nvidia's record data centre revenue?
Nvidia 's data centre segment generated a record US$75.2 billion , up 92 per cent year on year. Growth was driven by the ramp-up of Blackwell GB300 products and strong demand for InfiniBand , Spectrum-X Ethernet , and NVLink networking solutions.
Who are Nvidia's new AI infrastructure customers beyond hyperscalers?
Jensen Huang described a fast-growing 'second category' of customers that includes AI-native clouds, enterprises, industrial companies, and sovereign AI projects. This segment contributed US$37.4 billion in data centre revenue, nearly matching the US$37.9 billion from traditional hyperscale clients.
