Technology WWC APAC Desk

Nvidia China Chip Sales Face a New Test—Can Exports Resume?

Nvidia China Chip Sales Face a New Test—Can Exports Resume?

SANTA CLARA — Nvidia’s latest earnings disclosures show that its Data Center Hopper products generated US$2.4 billion in shipments to China during the quarter ended Oct. 25, 2026, while the company assumes no Chinese data-centre computing revenue in its forecast for the next quarter. The Nvidia China chip sales outlook therefore presents an important distinction between business already recorded, products authorised for certain customers and sales the company currently expects to make.

The figures follow a leadership change at the US chipmaker and a new American regulation addressing overseas access to advanced computing. They do not establish that Nvidia has stopped shipping every type of graphics chip to China, or that an export licence automatically produces a sale.

Nvidia China Chip Sales: What the Latest Filing Shows

Nvidia’s Nov. 17 chief financial officer commentary reported US$115.8 billion in total revenue for its third quarter of fiscal 2027, up 103% from a year earlier. Data Center revenue reached US$107.6 billion. In that same commentary, the company recorded US$2.4 billion of Data Center Hopper product shipments to China during the quarter.

That last figure matters because it prevents a misleading reconstruction of recent events. The company had recorded shipments in a completed reporting period; its forward-looking outlook is a separate statement about expected activity in the period ahead.

Nvidia’s fourth-quarter revenue outlook is US$130 billion, plus or minus 3%, and assumes no Data Center compute revenue from China. A forecast assumption is not a retrospective amendment to its reported third-quarter figures.

Analysis: Investors should avoid turning a zero-revenue forecast for one product segment into a claim that all company exports or sales in China have ceased. The categories and time periods are different.

Why Export Approval Has Not Guaranteed Sales

The US Commerce Department revised its licensing policy on Jan. 13, 2026, to review applications involving Nvidia H200 processors, AMD MI325X products and comparable chips on a case-by-case basis. Its announcement described conditions involving purchaser screening, independent US testing and protection of semiconductor supply capacity.

Nvidia’s previously filed quarterly report said US authorities had issued licences for limited H200 shipments to specified China-based customers beginning in February. It also said Chinese government restrictions prevented the company from selling all the products covered by those licences.

The same filing described a US$400 million charge in the first half of fiscal 2027 relating to H200 excess inventory and purchase obligations. It identified costs associated with the inspection process, including a 25% tariff when products were brought into the United States for inspection, and said the company had not been able to pass that cost to customers.

WWC previously examined why Nvidia’s export licences do not necessarily lead to completed Chinese sales. The key commercial test is whether regulatory permission, buyer approval, price and actual demand align.

Leadership Changed Before the Earnings Release

An SEC current report filed on Nov. 12 confirms that Nvidia’s board accepted Jen-Hsun Huang’s resignation as president, chief executive and director on Nov. 11. The board appointed Colette M. Kress interim chief executive while she continued as chief financial officer, and named Harvey C. Jones board chair.

Those are documented governance changes. The filing itself does not establish that the CEO transition caused the company’s China shipment outlook or that new leadership ordered a permanent withdrawal from that market.

Analysis: A management transition may affect how investors scrutinise international risk, but causal claims require explicit company records. Export-control exposure already appeared in Nvidia filings before the change of leadership.

New Rules Reach Beyond Physical Chip Shipments

On Nov. 16, the Commerce Department’s Bureau of Industry and Security announced an interim final rule addressing remote access to advanced US computing through data centres outside restricted jurisdictions. The agency said it was responding to circumstances in which customers could use powerful computing resources without importing processors directly.

The rule is distinct from the January H200 licensing policy. Physical shipment, resale and remote computing access can trigger different questions about a product’s destination, the customer and the ultimate user of the computing capacity.

WWC’s report on the November restrictions covering overseas computing access describes the broader regulatory context. Nothing in the official announcement proves that a specific Nvidia customer violated the new requirements.

Analysis: For data-centre operators, compliance can involve understanding who ultimately controls a customer and what computing service is being delivered. That is more involved than checking only where a server is installed.

A China Revenue Gap Does Not Mean Global Demand Has Vanished

Nvidia’s record quarterly Data Center revenue shows that demand from other customers continued to grow even with uncertainty around Chinese sales. Its financial commentary attributes growth to a broad range of computing platforms and customers, not a confirmed recovery in the Chinese market.

Earlier, the company said it could still ship uncontrolled products to China, including some gaming and workstation GPUs, while being effectively shut out of wide distribution in China’s data-centre computing market. The statement illustrates why product categories matter: restrictions on advanced data-centre processors are not automatically a complete ban on consumer graphics hardware.

WWC has also reviewed how new US AI chip controls can affect Southeast Asian data centres. For Philippine businesses buying cloud services, the immediate concern is whether providers can demonstrate lawful access and dependable capacity; no official record reviewed establishes an automatic ban on Philippine customers.

What Happens Next

Future Nvidia filings can clarify whether the zero-China-compute-revenue assumption becomes the actual result, whether additional licences produce deliveries, and how the new remote-access requirements affect business. Government licensing announcements and formal regulatory text will matter more than broad interpretations of a single executive comment.

The reliable picture today is specific: Nvidia recorded Hopper shipments to China in its latest completed quarter and excluded China Data Center compute revenue from its next-quarter forecast. Whether that gap persists or changes is still uncertain.

Frequently Asked Questions

Has Nvidia stopped selling all chips to China?

The official releases reviewed do not establish a halt to every product category. Nvidia distinguishes controlled data-centre processors from other products, including gaming and workstation GPUs.

Did Nvidia ship AI data-centre chips to China in its latest quarter?

Yes. Its Nov. 17 financial commentary records US$2.4 billion of Data Center Hopper product shipments to China in the quarter ended Oct. 25, 2026.

Does Nvidia expect China Data Center compute revenue next quarter?

No. Its fourth-quarter fiscal 2027 outlook assumes none, although the assumption is forward-looking.

Who is Nvidia’s interim chief executive?

Colette M. Kress was appointed on Nov. 11, 2026, and remains chief financial officer during the interim appointment.

Are US licences enough to ship H200 processors to China?

No. Licences are subject to conditions, and Nvidia’s own filing describes restrictions and commercial obstacles that limited shipments despite approval.

Get our stories first on Google

More in Usa