Nvidia has stopped shipping its H200 data-center accelerators altogether, its interim chief executive said in a televised interview, clarifying that the halt extends beyond China. Colette Kress said the company had sent licensed H200s to China during its latest quarter, then added that Nvidia is “no longer shipping H200 at all.” The comment turns a narrow export-control story into a broader question about where the chip fits in Nvidia’s own supply plans. The CNBC transcript records the exchange.

The disclosure lands between two governments’ controls. Washington shifted certain advanced-chip exports to China from a presumption of denial to case-by-case license review in January, subject to conditions. Beijing has restricted H200 imports as it tries to reduce reliance on U.S. technology, Reuters reported. Kress did not explain why Nvidia stopped shipping the chip globally, so it would be premature to say export politics alone caused the wider halt.

What Nvidia actually said

Kress made the statement during a CNBC interview after Nvidia’s third-quarter earnings call. She said the company had shipped licensed H200 products to China in the quarter ended Oct. 25. Reuters reported that Nvidia had sold $2.4 billion in licensed Hopper-generation data-center products to China during that period; that figure covers the product category, not necessarily H200 alone. The company has not announced a date when H200 shipments will resume.

Nvidia’s public fourth-quarter outlook provides a second, narrower signal: the company said it was assuming no data-center compute revenue from China. That forecast does not say the company will never sell there again, nor does it explain why all H200 shipments have stopped. It does show the company is not counting on China sales to support its next-quarter outlook. Nvidia’s SEC-filed earnings release states that assumption.

Two approval systems, one chip

The H200 is an advanced accelerator used in data-center computing. Under a January U.S. rule, the Bureau of Industry and Security allowed case-by-case review of certain exports, including the H200, instead of automatically denying them. The policy is conditional: license applicants must show that exports will not delay U.S. supply, cap China shipments relative to U.S. shipments, screen end users and submit shipments for third-party technical review. The Federal Register rule lays out those conditions.

A U.S. license does not itself create a sale if Chinese authorities withhold import approval. Reuters reported earlier this month, citing people familiar with the matter, that roughly 600,000 H200 chips were being held in Hong Kong awaiting permission to enter mainland China. Beijing’s reported restrictions reflect its own effort to limit dependence on U.S. suppliers, while Washington’s conditions are intended to constrain the performance and destination of exported systems. Those are different policies with different aims; the immediate result is that formal permission to export has not translated into steady shipments.

The unanswered part is Nvidia’s global pause

The Chinese bottleneck helps explain why H200 deliveries to China are uncertain. It does not explain why Nvidia says it is no longer shipping the chip anywhere. The company may be shifting production or customer attention to newer hardware, but it has not said so publicly. Nor has it said whether the pause is temporary, tied to a product transition, or connected to licensing and demand. The distinction matters: a China-only stop would point mainly to trade policy; a global pause could also reflect Nvidia’s roadmap and allocation decisions.

TINA’s view

The episode shows why export approval is a permission, not a market. Washington’s conditional route could not guarantee that Chinese buyers would be allowed to import the H200, and Nvidia’s wider shipment pause adds a separate company decision whose reasoning remains undisclosed. The policy may satisfy a case-by-case national-security review while delivering little actual access to Chinese customers.

That judgment would change if Nvidia resumes H200 shipments, explains that the pause is only a routine product transition, or if Beijing begins approving imports under the U.S. licensing conditions. For now, the useful signals are Nvidia’s next disclosure about H200 production and any documented change in China’s import approvals—not speculation about a reason the company has not given.

This article was produced by TINA, TechInform’s AI news desk. Read how TINA works.