NVIDIA CEO Jensen Huang says the company has effectively stepped back from China’s advanced AI chip market, giving local competitors such as Huawei a major opening to grow.
The comments come after NVIDIA Chief Financial Officer Collette Kress confirmed during the company’s latest earnings call that the U.S. government has approved licenses allowing NVIDIA to ship its H200 AI GPUs to customers in China. However, Kress made it clear that approval does not guarantee sales. NVIDIA has not yet generated revenue from those potential shipments, and the company remains unsure whether China will allow the chips to be imported.
“While the U.S. government has approved licenses for H200 to be shipped to China-based customers, we have yet to generate any revenue. And we are uncertain whether any imports will be allowed into the country,” Kress said during the call. She added that NVIDIA is not including China data center compute revenue in its outlook, continuing the same approach as the previous quarter.
The situation highlights the complicated position NVIDIA now faces in one of the world’s most important artificial intelligence markets. U.S. export controls have restricted the sale of NVIDIA’s most advanced AI GPUs to China, forcing the company to create modified, China-specific chips. But even those products have run into roadblocks, with Chinese authorities reportedly discouraging or restricting some purchases as the country pushes to strengthen its domestic semiconductor industry.
In an interview with CNBC, Huang said Chinese chipmakers are gaining momentum because NVIDIA has been pushed out of the market.
“The local Chinese ecosystem is doing quite well, because we’ve evacuated that market,” Huang said. He also predicted that Huawei could “have an extraordinary year coming up.”
Huawei has become one of the biggest beneficiaries of China’s drive to reduce reliance on foreign AI hardware. As NVIDIA’s access to Chinese customers has narrowed, domestic companies have stepped in to fill the gap. Huawei’s AI chips have received growing attention from local cloud providers, technology companies, and government-backed buyers looking for alternatives to U.S. hardware.
Still, China’s chip industry faces significant challenges. U.S. restrictions on advanced semiconductor manufacturing equipment have made it harder for companies such as Huawei to produce chips using the most cutting-edge process technologies. As a result, many domestic AI accelerators are believed to rely on older manufacturing methods, which can limit performance and efficiency compared with the latest U.S.-designed GPUs.
Industry estimates suggest that the most powerful U.S. AI chips remain far ahead of Huawei’s current offerings. Some projections indicate that top NVIDIA AI GPUs may deliver several times the performance of Huawei’s alternatives. Even if Huawei were able to produce millions of AI chips in the coming years, its total computing capacity would still represent only a small fraction of NVIDIA’s global AI computing footprint.
Huang has repeatedly emphasized that NVIDIA supports U.S. leadership in artificial intelligence. Speaking earlier this month at a major business conference in Los Angeles, he said China should not receive NVIDIA’s newest AI chips, adding that the company is a strong supporter of the United States having “the first, the most and the best” technology.
The H200 GPUs recently cleared for potential shipment to China are based on NVIDIA’s older Hopper architecture. Since then, NVIDIA has moved forward with newer AI chip platforms, including Blackwell and future designs such as Vera Rubin. These next-generation products are expected to power the next wave of AI data centers, large language models, and advanced enterprise AI systems.
The uncertainty around China comes at a time when NVIDIA remains the dominant force in the global AI chip market. Demand for its GPUs has soared as cloud providers, startups, and large enterprises race to build and deploy artificial intelligence systems. Data center revenue has become the core driver of NVIDIA’s growth, with AI training and inference workloads fueling massive orders for its hardware.
However, the rise of agentic AI could reshape parts of the market. Agentic AI refers to systems that can plan, reason, take actions, and complete complex tasks with less human input. As these workloads evolve, the role of CPUs is becoming more important again alongside GPUs. That shift has renewed investor and industry attention on companies such as Intel, where CEO Lip-Bu Tan is working to revive competitiveness and reposition the company for the next phase of AI computing.
For NVIDIA, China remains both a major missed opportunity and a strategic challenge. The company says it would be happy to serve Chinese customers where allowed, but export controls and import uncertainty have made future revenue difficult to predict. Meanwhile, Huawei and other Chinese chipmakers are using the opening to expand their presence in the domestic AI market.
The key question now is whether NVIDIA can regain meaningful access to China with approved products like the H200, or whether China’s AI hardware market will continue shifting toward local suppliers. For now, Huang’s message is clear: NVIDIA’s absence has created room for Chinese competitors to rise, and Huawei may be one of the biggest winners.
