The White House has informed other federal agencies that NVIDIA will not be allowed to sell its most advanced AI chip, the B30A, to Chinese customers. The move marks a major escalation in the US-China tech rivalry. The decision effectively extends the Washington export restrictions to include even scaled-down versions of the chip giant’s powerful AI hardware, according to The Information.
The B30A chip, though designed to meet previous limits on exports, can still be used in large clusters to train large language models. This is a use that US officials reportedly deem too advanced to export to China.
This action is part of the United States' broader strategy to halt the transfer of advanced technology to China, which could support the development of the latter’s AI-driven military capabilities.
NVIDIA had already distributed samples of the B30A chips to many major clients in China before the effective date of the order. Nevertheless, a company spokesperson told Reuters that NVIDIA currently holds ‘zero share’ in China’s highly competitive datacenter computing market, and the company has not included any revenue from the country in its financial forecasts.
Insiders told The Information that the California-based chipmaker is now working to redesign the B30A in the hope that the US administration will reconsider its stance. NVIDIA has seen record growth fueled by global demand for AI chips, but has faced repeated regulatory setbacks as Washington tightens its tech export rules to Beijing.
China has responded with new policies aimed at accelerating its chip self-sufficiency. According to a Reuters report, Beijing has ordered that all new data center projects receiving state funding must exclusively use domestically developed processors.
Those projects that are less than 30% complete must remove any installed foreign chips or scrap plans to buy them; those farther along will be considered on a case-by-case basis.
The policy effectively sidelines NVIDIA and other foreign chipmakers from one of the world’s fastest-growing AI infrastructure markets. This clampdown comes even as some advanced NVIDIA chips, barred under US rules, have continued to surface in China through grey market channels.
Also Read: Michael Burry Bets Big Against AI Boom With NVIDIA, Palantir Shorts
The recent limitations have increased the separation of technology between Washington and Beijing, leading to greater self-sufficiency for both nations. NVIDIA could lose one of the most significant markets for AI, the Chinese market; however, it might balance this loss with a rapidly growing demand in other regions.
China, on the other hand, is undergoing a rapid shift towards establishing an entirely home-grown semiconductor ecosystem that can power its AI dreams without relying on US technology.