Nvidia’s dominance in China’s data-centre AI chip market has declined, falling to a 34% share in the first half of 2026. This drop comes amid US export controls that have restricted supplies of Nvidia’s products, allowing domestic rivals like Huawei to gain ground. According to an IDC report seen by Reuters, Nvidia shipped 678,000 AI accelerators in China during this period, a 44% decrease from the previous year. AI accelerators are specialized chips used to train and run AI models.
Washington’s restrictions have limited China’s access to Nvidia’s most powerful AI chips, such as the B200 and B300, which are banned from export but have reportedly been smuggled into the country. Nvidia CEO Jensen Huang has criticized these export curbs, stating they have forced a US company to cede China’s $50 billion AI chip market to Huawei. The IDC report did not specify which Nvidia products contributed to its 34% market share or whether it included cutting-edge chips.
Huawei, meanwhile, shipped 572,000 AI accelerators, up from more than 300,000 a year earlier, giving it a 28% share of the overall market. Other Chinese chipmakers, including T-Head (7%), Cambricon (5%), Kunlunxin (4%), and Biren (3%), also increased their shipments. Total AI-accelerator shipments reached 2.008 million units in the first half of the year.
Shipments of graphics processing units (GPUs) fell 26% to 991,000 units due to limited supplies eligible for sale in mainland China. However, shipments of other AI accelerators rose nearly 80% to 1.016 million units, overtaking GPUs by volume. This category includes chips designed for specific AI tasks, such as running trained models. Huawei, T-Head, Kunlunxin, and Cambricon together accounted for 95% of the non-GPU segment, which is being adopted by government agencies, internet companies, and telecoms providers.
IDC expects total accelerator shipments in China to grow 19% this year, even as GPU shipments fall 9%. The competition is shifting from the performance of a single chip to system-level capabilities, including how well chips work in a larger computing cluster.