Nvidia loses China AI-chip share to Huawei

In the first half of 2026, Nvidia’s share of China’s data-centre AI-chip market declined to 34%, reflecting a significant shift in the country’s accelerator landscape. The company shipped 678,000 AI accelerators to mainland China between January and June, a 44% drop compared with the same period a year earlier. The data highlights how Nvidia’s position in one of the world’s largest AI markets has weakened within a relatively short time frame.

The contraction in Nvidia’s shipments occurred against a backdrop of restrictions on the range of its products that can be sold in China. These limits have reduced the availability of some of Nvidia’s most sought-after data-centre AI chips, affecting its ability to meet demand from Chinese customers. The resulting gap has created more room for domestic suppliers to expand their presence.

Huawei substantially narrowed the distance to Nvidia during the same six-month period. It shipped 572,000 AI accelerators in China, up from more than 300,000 units a year earlier, giving it a 28% share of the overall data-centre AI-chip market. This strong year-on-year growth positions Huawei as the closest rival to Nvidia in China’s accelerator segment.

Beyond Huawei, domestic chipmakers as a group increased their footprint. IDC found that nine of the 11 Chinese chipmakers it tracks boosted shipments year-on-year in the first half of 2026. The broad-based increase underscores the growing role of local suppliers in meeting China’s demand for AI computing hardware.

China’s AI-accelerator mix also changed notably in the first half of 2026. Total AI-accelerator shipments reached 2.008 million units, but growth was concentrated in non-GPU products. Shipments of GPUs, which are widely used for training AI models, fell 26% to 991,000 units over the period.

In contrast, shipments of other types of AI accelerators rose nearly 80% to 1.016 million units, allowing them to surpass GPUs by volume. This marks a significant milestone in China’s data-centre market, indicating rising adoption of alternative accelerator architectures and solutions.

Policy measures on both sides of the Pacific influenced these market dynamics. U.S. export controls have limited the range of Nvidia data-centre AI chips that can be supplied to China, constraining the company’s local product lineup. These controls have encouraged Chinese companies to seek substitutes where possible.

At the same time, Beijing has promoted domestic alternatives to foreign AI chips and required companies to seek approval before importing compliant Nvidia products. Together, these measures have supported increased shipments from local chipmakers and accelerated the shift toward domestic and non-GPU accelerators in China’s data-centre AI market.