Huawei Gains Ground as China’s Demand for Nvidia Chips Stalls

Export restrictions imposed by the U.S. government have significantly altered the landscape of China’s artificial intelligence sector, leading to a surge in domestic chip adoption [1]. While the U.S. recently approved the sale of Nvidia’s H200 GPUs to Chinese firms, the anticipated rush to purchase these chips has failed to materialize [1].

Key takeaways

• China’s self-sufficiency ratio for AI chips reached 41% this year, a fourfold increase over the past five years [1].
• Huawei’s Ascend AI chips are increasingly powering domestic workloads as Chinese firms adapt their software to local hardware [1].
• Morgan Stanley projects that China’s domestic AI chip self-sufficiency could reach 85% by 2028 [1].
• Despite the stall in China, Nvidia continues to see global growth driven by demand from U.S. hyperscalers like Microsoft, Amazon, and Meta [1].

The Shift Toward Domestic Semiconductor Independence

For years, China served as a critical growth engine for Nvidia, at one point accounting for roughly 20% of the company's data center revenue [1]. However, beginning in 2022, the Biden administration implemented export controls on advanced semiconductors, including the A100 and H100 GPUs, to slow China’s technological progress [1]. These restrictions eventually expanded to include chips specifically modified for the Chinese market, such as the H800 [1]. By early 2025, Nvidia’s revenue from China had fallen to low single digits as a percentage of its total sales [1].

Following a meeting between President Trump and Chinese President Xi Jinping, the U.S. government reportedly approved the shipment of Nvidia’s H200 GPUs to China [1]. While major Chinese technology companies including Alibaba, JD.com, Lenovo, and ByteDance expressed interest in the hardware, the market has not reopened as expected [1]. Beijing’s strategy of pushing domestic companies toward local suppliers during the period of U.S. restrictions appears to have been effective [1]. Chinese cloud providers and data center operators have spent years building their software stacks around domestic hardware, reducing their reliance on Nvidia’s CUDA ecosystem [1].

Why it matters

The shift in China’s semiconductor market suggests that the U.S. effort to curb China’s AI ambitions may have inadvertently accelerated the country’s push for technological independence [1]. While Nvidia remains a dominant global force with strong margins and significant free cash flow, its negotiating power in China has weakened [1]. If domestic suppliers continue to meet a larger percentage of local demand, China may become a permanently smaller market for American chipmakers than previously anticipated [1]. For investors, the situation highlights that while global AI spending remains robust, the Chinese market is no longer the guaranteed growth opportunity it once was [1].