Chinese semiconductor companies captured nearly 41 percent of the country’s artificial intelligence accelerator server market in 2025, significantly narrowing U.S. tech giant Nvidia’s once‑dominant position, industry data showed Wednesday.
The figures, compiled by market research firm IDC, reflect a rapid rise in locally produced graphics processing units (GPUs) and AI chips as Beijing steps up efforts to reduce dependence on foreign technology amid ongoing U.S. export restrictions on advanced semiconductors.
Nvidia, long the dominant supplier of AI accelerator cards, remained the market leader with a 55 percent share, shipping around 2.2 million units across China last year. But its control of the market has noticeably weakened from earlier years, as domestic vendors gain traction.
Together, Chinese firms shipped about 1.65 million AI accelerator cards in 2025 — roughly 41 percent of the total — underscoring how aggressively local players have expanded their presence in one of the world’s most important technology markets.
Among the Chinese vendors, Huawei emerged as the clear leader, with approximately 812,000 units shipped, accounting for nearly half of domestic brand shipments. Alibaba’s T‑Head unit followed with about 265,000 units, while Baidu’s Kunlunxin and Cambricon each shipped roughly 116,000 cards. Smaller players such as Hygon, GPU startup MetaX and Iluvatar CoreX also contributed to the surge in domestically produced AI chips.
More Read
The growth of China’s homegrown AI chip industry has been driven in part by government‑backed investments in AI infrastructure and implicit policies that favour domestic technologies. Local authorities across provinces have accelerated plans for intelligent computing centres, many of which carry directives to “buy Chinese,” according to sources familiar with the sector.
Analysts say the shift reflects broader geopolitical tensions and technology competition between Beijing and Washington, especially after a series of U.S. export controls curtailed deliveries of Nvidia’s most advanced products to China. Those restrictions have cut off access to high‑end foreign chips used in data centres and AI workloads, prompting Chinese firms to accelerate development of local alternatives.
While Nvidia still leads the market overall, its reduced share highlights how China’s semiconductor ecosystem is evolving rapidly. Local companies have been focusing on improving performance and compatibility with native AI frameworks to narrow the gap with global rivals, analysts say.
More Read
“The Chinese market is one of the largest in the world for AI computing,” said a Beijing‑based semiconductor analyst. “Capturing a significant share of this space shows how quickly domestic vendors can scale when supported by policy and investment.”
Industry observers note, however, that despite the gains, Chinese chips generally still lag behind the most advanced AI accelerators in raw performance and power efficiency. The challenge of closing that gap — especially in high‑end data‑centre chips — remains significant, analysts say.
The shift in China’s AI chip market comes as global demand for artificial intelligence computing power continues to surge, driven by rapid growth in generative AI applications, data‑intensive computing tasks and large language models. At the same time, geopolitical factors such as export restrictions and technology decoupling have reshaped supply chains and competitive dynamics.
For Nvidia, the shrinking share in China underscores both the limits of foreign reliance in a politically sensitive sector and the opportunities for domestic chipmakers to take a larger role in their home market. The U.S. firm has faced challenges in securing broad access to China’s AI infrastructure business amid policy uncertainties and national security concerns, even as it remains a dominant supplier in many other global markets.
Despite these headwinds, Nvidia continues to invest in research and development and explores ways to navigate regulatory and market shifts. Meanwhile, Chinese chipmakers are expected to continue growing their capabilities, driven by strong domestic demand, government backing and expanding ecosystem support.
