Nvidia chief executive Jensen Huang says the company’s share of China’s artificial intelligence accelerator market has effectively collapsed to zero, delivering one of the starkest warnings yet about the consequences of Washington’s escalating semiconductor export restrictions.
“In China, we have now dropped to zero,” Huang said during an interview with the Special Competitive Studies Project.
The remarks mark one of the clearest acknowledgements yet of how severely U.S. export controls have disrupted Nvidia’s once-dominant position in China, previously one of its most strategically important markets.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
Just two years ago, Nvidia controlled the overwhelming majority of China’s AI accelerator market, supplying the advanced GPUs used by Chinese cloud giants, research institutions, and AI startups to train large language models and power hyperscale computing infrastructure.
Now, Huang says Washington’s restrictions have effectively erased Nvidia’s direct commercial foothold in the country.
“Conceding an entire market the size of China probably does not make a lot of strategic sense, so I think that has already largely backfired,” Huang said. “Maybe it made sense at the time, but I think the policy really needs to be dynamic and needs to stay with the times.”
His comments amount to a pointed critique of the increasingly aggressive semiconductor export controls imposed by successive U.S. administrations to curb China’s AI ambitions. Washington has argued the restrictions are necessary to prevent advanced American chips from supporting Chinese military modernization and strategic AI capabilities. But Huang suggested the policy may instead be accelerating China’s technological independence while weakening the global reach of American AI platforms.
The Nvidia chief’s frustration also comes after Washington recently softened parts of its hardline position and signaled willingness to permit exports of some downgraded Nvidia AI chips to China. The Trump administration moved earlier this year to allow shipments of Nvidia’s H20 AI chips to proceed, with White House economic adviser Kevin Hassett saying the decision was intended to preserve America’s technological edge and prevent China from fully replacing U.S. suppliers. The deal includes 15% revenue sharing with the U.S. government.
Washington later began issuing export licenses for some H20 shipments after Nvidia said it had received assurances that approvals would move forward. The H20 chip was specifically designed by Nvidia for the Chinese market after earlier export restrictions blocked sales of its more advanced processors.
But even those efforts to partially reopen the market appear to have run into growing resistance from Beijing itself. Chinese regulators later summoned Nvidia over alleged security concerns tied to the H20 chips, including fears around potential tracking, remote access, and so-called “backdoor” vulnerabilities.
The scrutiny highlighted a new layer of distrust emerging in the technology confrontation between the world’s two largest economies.
Beijing’s concerns are believed to have significantly complicated Nvidia’s efforts to rebuild its China business, even after Washington relaxed some licensing restrictions. The Cyberspace Administration of China reportedly questioned whether U.S.-designed chips could expose Chinese data or critical systems to surveillance or remote intervention capabilities. Nvidia denied the allegations and insisted its chips contain no “backdoors” that would allow remote control or unauthorized access.
For Beijing, the U.S. sanctions appear to have reinforced long-standing concerns that reliance on American technology creates strategic vulnerabilities. Chinese officials have repeatedly warned they would not bow to Washington’s pressure and instead would accelerate domestic innovation and semiconductor self-sufficiency.
The rapid rise of Chinese AI hardware firms now increasingly reflects that strategy. Companies such as Huawei, Cambricon, Moore Threads, and MetaX are aggressively expanding production as China attempts to reduce dependence on Nvidia and other U.S. suppliers.
Research firm Bernstein previously estimated Nvidia’s China AI GPU market share could collapse from 66% in 2024 to roughly 8% over time as domestic Chinese vendors move to satisfy as much as 80% of local demand.
Huang’s latest comments suggest the deterioration may have been even faster. The Nvidia chief argued that China retains formidable structural advantages in artificial intelligence regardless of hardware restrictions.
“American companies win around the world,” Huang said. “The argument there is that across the five-layer cake, there’s one particular layer that is too important because in the others, China can get ahead. They have cheaper energy. They have incredible talent.”
He pointed specifically to China’s deep engineering and research base.
“So, they have the number of science and math experts, and as a result of that, the number of AI researchers in China is quite extraordinary, it’s one of their national treasures,” Huang said.
The comments reflect growing concern inside parts of the U.S. technology sector that export controls may ultimately strengthen Chinese competitors rather than weaken them. Huang has increasingly argued that the real strategic advantage for the United States lies not merely in hardware leadership but in maintaining global dominance of software ecosystems such as Nvidia’s CUDA platform, which remains deeply embedded in AI development worldwide.
But the longer Chinese companies are shut out from American hardware, analysts say, the stronger the incentive becomes for Beijing to build parallel semiconductor and software ecosystems insulated from U.S. influence.
That fragmentation could eventually weaken the global dominance of American AI standards and reduce the international reach of U.S. technology companies. The geopolitical tensions are already reshaping supply chains and investment flows across the semiconductor sector.
The United States has tightened restrictions on advanced AI chip exports, while China has intensified efforts to localize semiconductor production and reduce dependence on foreign suppliers. At the same time, Washington’s own policy shifts have shown growing unease about completely surrendering the Chinese market to domestic competitors.
U.S. officials acknowledged that preventing China from buying American chips entirely could accelerate indigenous Chinese innovation and erode U.S. influence over global AI infrastructure.
That policy dilemma now sits at the center of the AI cold war.