Nvidia's share of China's data-center AI accelerator market fell to 34% in the first half of 2026, while Huawei reached 28%, according to IDC data reviewed by Reuters. The six-percentage-point gap shows how export controls are reshaping a market Nvidia once dominated.
The shipment data make the shift concrete
Reuters reported that Nvidia shipped 678,000 AI accelerators in China from January through June, down 44% from a year earlier. Huawei shipped 572,000 units, up from more than 300,000 in the prior-year period.
These are IDC estimates, not audited company disclosures. Even so, they provide a consistent measure of the competitive direction. Nvidia still led the overall market in the period, but domestic suppliers gained ground as U.S. rules limited access to Nvidia's most capable processors and Chinese policy supported local alternatives.
The mechanism is policy as much as product
Export restrictions reduce the range and volume of Nvidia products available to Chinese customers. That creates room for Huawei and other domestic suppliers to improve software, packaging and system design around locally available chips.
Nvidia's advantage is not only raw processor performance. Its CUDA software ecosystem, networking products and developer base make switching costly. Huawei can narrow the gap by offering complete systems and by linking many processors into large clusters, even if an individual chip is less capable.
Immediate guidance already assumes little direct benefit
Nvidia reported fiscal second-quarter 2027 revenue of $96.2 billion, up 18% sequentially and 106% year over year. Importantly, the company's outlook assumed no data-center compute revenue from China.
That reduces the risk that the reported market-share loss will create a sudden miss against guidance built on Chinese accelerator sales. It does not remove the strategic cost. A durable local ecosystem can reduce Nvidia's future opportunity in a large market and create competitors that later expand outside China.
Separate market share from AI financing fears
The same trading session included concern about the cost of financing global AI infrastructure. That is a different risk. China share loss affects competitive access, while financing constraints affect how quickly customers everywhere can build data centers.
Investors should track licensed product approvals, Nvidia's disclosed China revenue, Huawei shipment growth and adoption of domestic software tools. BTI's conclusion is that the near-term earnings hit is partly contained by Nvidia's existing guidance, but the longer-term competitive signal is negative because policy is helping a domestic ecosystem scale.