Nvidia halts H200 AI chip shipments to China, interim CEO says
Nvidia is no longer shipping licensed Hopper data-center products, including the H200, to China, interim CEO Colette Kress said on the company’s third-quarter fiscal 2027 earnings call. Kress said Nvidia’s guidance for the quarter contains no licensed Hopper data-center products and gave no reason for the halt. The stop comes despite approval from the White House last December that let Nvidia ship the H200 to China, a right the company had lobbied for under then-CEO Jensen Huang.
Verified and unconfirmed
- Verified: Interim CEO Colette Kress said on Nvidia’s third-quarter fiscal 2027 earnings call that its guidance includes no licensed Hopper data center products and that the company is “no longer shipping licensed Hopper data center products to China,” according to Bloomberg and CNBC-TV18. Kress also said licensed Hopper data center sales to China totaled $2.4 billion in the quarter, Yahoo Finance reported.
- Not yet confirmed: Nvidia has not said why it stopped the shipments, when they ceased, or whether the halt covers only the US-licensed Hopper line or all advanced AI chips bound for China; the report offers no independent confirmation of the figures, and the December US approval referenced by The Verge is described only in the outlet’s own earlier coverage.
Background
Nvidia’s H200 is part of its Hopper data-center line, and The Verge reported that the White House approved shipments of the chip to China in December last year — a permission the company had lobbied for under then-CEO Jensen Huang. Two days before the earnings call, AI Daily Radar’s November 16 digest reported that the Commerce Department had issued an interim rule requiring export licenses for AI chips that may provide compute to China, extending controls beyond physical shipments to remote access from countries including Singapore, Malaysia and Thailand. AI Daily Radar’s November 17 digest also noted that Nvidia had privately warned customers about legal risk under the proposed Remote Access Security Act, which is not yet law; Kress did not say which factor, or anything else, prompted the halt.
Why it matters
Chinese cloud and AI developers that had counted on licensed H200 capacity now have no route to new Hopper-class compute from Nvidia and must fall back on existing inventory or domestic accelerators. The revenue hit looks limited for Nvidia: Kress said guidance already excludes licensed Hopper data center products, and the backlog for its newer Vera Rubin chips means the export rules would not affect top-line revenue, according to reports on the call. One account put the preceding Hopper shipments to China at roughly $2.4 billion a quarter, and noted that Chinese buyers had already largely stopped accepting H200 deliveries before Nvidia halted them.
What to watch
Nvidia’s next quarterly earnings report will show whether China Hopper revenue stays at zero, and any statement from Nvidia or the US Commerce Department should clarify whether last December’s export approval is formally withdrawn or simply left unused.
Further reading
- 2026-11-16 — US Commerce tightens AI chip export rules to cover remote compute access
- 2026-11-17 — Nvidia reportedly warns buyers of legal risk under proposed Remote Access Security Act
- Nvidia no longer selling US-approved Hopper chips to China …
- Interim Nvidia CEO Says It’s No Longer Sending AI Chips to …
- Nvidia says no longer shipping ‘Hopper’ AI chips to China
- Nvidia stops licensed H 200 shipments to China — Data Phoenix
- Nvidia is abandoning China after Jensen Huang’s ouster
Comments
Comments use a GitHub account and are moderated. Comment rules