Jensen Huang: Nvidia Has 'Zero Percent' Market Share in China as Export Controls Backfire
Key Takeaways
- ▸Nvidia's market share in China collapsed from 66% (2024) to 0%, a direct result of U.S. export controls and rapid scaling of Chinese competitors
- ▸Chinese companies (Huawei, Cambricon, Moore Threads, MetaX) are advancing both AI silicon and software to capture up to 80% of domestic GPU demand
- ▸Huang argues that absolute market denial accelerates China's self-sufficiency in AI chips and software rather than limiting its capabilities, potentially making the export control strategy counterproductive
Summary
Nvidia CEO Jensen Huang revealed that the company's market share in China's AI accelerator market has collapsed to 0%, down from an estimated 66% in 2024. The dramatic decline comes as a direct result of U.S. export controls on advanced semiconductors and the rapid rise of Chinese domestic competitors like Huawei, Cambricon, Moore Threads, and MetaX, which are targeting up to 80% of China's AI GPU market. Huang argued in an interview with the Special Competitive Studies Project that the export control policy has proven strategically counterproductive, accelerating China's drive toward self-sufficiency while eliminating opportunities for American companies to maintain influence in the Chinese market.
Despite acknowledging Nvidia's collapse in direct China sales, Huang emphasized that China remains a formidable AI competitor with significant advantages: cheaper energy costs, extraordinary talent, and an exceptionally large pool of AI researchers. He cautioned that U.S. export restrictions may ultimately slow global AI deployment and suggested that American technological leadership would be better served by maintaining market presence rather than total market denial. Huang warned that continued export controls could inadvertently push China to advance its own GPU and software stacks more aggressively, while the U.S. loses leverage over the evolution of AI technology globally.
- China's structural competitive advantages—lower energy costs, abundant technical talent, and a large pool of AI researchers—may matter more than hardware access restrictions
Editorial Opinion
Huang makes a compelling case that hardline export restrictions may be strategically self-defeating, though his claim of 'zero percent' direct Nvidia sales likely obscures significant gray-market flows. The real question isn't whether the U.S. can prevent Chinese access to chips—clearly it cannot—but whether export controls actually constrain China's AI progress or simply accelerate its independence. If Huang is right that China's fundamental advantages lie in energy, talent density, and researcher quantity, then the U.S. competitive edge may depend less on chip embargoes and more on dominating the entire AI software stack and ecosystem.
