Nvidia's sales of artificial intelligence chips in China have stalled, and domestic chipmakers, led by Huawei, now hold the lead in the country's AI accelerator market, AP News reports. The shift marks one of the clearest reversals yet in the long-running contest between US export controls and Beijing's push for semiconductor self-sufficiency.

The report centers on a straightforward commercial fact: Nvidia, which once dominated China's market for AI training chips, is no longer converting that position into sales. Chinese buyers are turning instead to homegrown suppliers, with Huawei named as the front-runner. Huawei's Ascend product family has become the principal domestic alternative for Chinese firms building AI infrastructure.

Why have Nvidia's sales stopped growing?

The mechanism is political as much as commercial. Washington has progressively tightened export restrictions on advanced chips sold to China, pushing Nvidia's most capable data center products out of the market entirely. Nvidia responded with reduced-specification parts designed to comply with the rules, but those chips offer less performance than the full versions, and Chinese customers have grown reluctant to build AI systems around them.

At the same time, Beijing has actively encouraged domestic firms to buy local. Government procurement guidance and informal pressure have both pointed in the same direction, toward Huawei and other Chinese designers. The result is a market where Nvidia's compliant chips must compete not only on performance but against a policy environment that favors local suppliers. In plain terms, Nvidia is fighting export rules on one side and state-backed rivals on the other.

What is Huawei's position?

Huawei has spent years building a domestic alternative stack: it designs its own AI processors and has secured fabrication capacity inside China. That vertical position matters, because it insulates the company from the same export controls that constrain Nvidia. Chinese cloud providers and AI developers now treat Huawei's accelerators as the default high-end option when US chips are unavailable or unwelcome, according to the report.

The competitive picture is therefore asymmetric. Nvidia retains the technological lead in raw chip performance, but it cannot freely sell its best products in China, and its compliant offerings face a market increasingly structured to exclude them. Huawei's advantage is not superior silicon; it is access.

What does this mean for the AI supply chain?

China is one of the largest markets for AI compute, and its detachment from the US supplier base carries consequences in both directions. Nvidia loses revenue that once flowed from Chinese cloud and internet companies, a segment that accounted for a meaningful share of its data center business before the restrictions tightened. Chinese AI firms, meanwhile, must build their systems on domestic chips that generally trail the frontier in performance and availability.

The broader supply chain is splitting into two tracks. One track runs through Nvidia and the leading-edge foundries outside China; the other runs through Huawei and domestic manufacturing. Each side now has a captive customer base, and the overlap between them is shrinking.

What comes next?

The AP report does not suggest a near-term recovery for Nvidia in China. As long as US export controls stay in place and Huawei continues to expand its installed base among Chinese buyers, the local suppliers' lead is likely to widen rather than narrow, making China's AI chip market a structurally separate arena for the foreseeable future.