Nvidia CEO says he is ‘disappointed’ after report China ordered halt to use of company’s AI chips

Nvidia Chief Executive Jensen Huang said he was "disappointed" after reports that China’s Cyberspace Administration had instructed major technology companies to stop using AI chips manufactured by the U.S. firm for the Chinese market.

The comments came Wednesday as Huang, who is accompanying U.S. President Donald Trump on a state visit to the United Kingdom, told reporters he would be "patient" in response to the move and would back U.S. efforts to resolve the dispute. He added he would tell the same to the president if asked. The report was published by the Financial Times on Wednesday morning.

The FT said the Cyberspace Administration had told companies including DeepSeek, Tencent and Alibaba to halt purchases or use of Nvidia chips that were specifically manufactured for China. Nvidia shares were down more than 1% in premarket trading following the report.

Nvidia, whose chips power data centres and many AI applications worldwide, has become central to the global artificial-intelligence boom. The company surpassed a $4 trillion valuation earlier in 2025 and has been at the centre of intensifying U.S.-China technology frictions. The U.S. previously restricted sales of Nvidia's most advanced chips to China; the restrictions were modified in July when the U.S. administration reached a deal allowing certain sales while imposing conditions.

Under that arrangement, Nvidia agreed to remit a share of revenues tied to Chinese sales to the U.S. government, a payment framework described by the company as unprecedented. Nvidia has said it must pay 15% of its Chinese revenues under the agreement reached in the summer.

Huang said he supported U.S. efforts to manage geopolitical issues and would back government approaches to resolving the dispute between the countries. He did not offer further detail on how Nvidia would adjust its business in China should the Cyberspace Administration guidance be confirmed and enforced.

China has been accelerating its own development of domestic AI chips as part of a broader strategy to reduce reliance on foreign technology and to compete with U.S. firms. Major Chinese technology companies had been among customers that ordered Nvidia hardware prior to the reported guidance, according to the Financial Times.

Beyond direct sales, Nvidia has announced investments and partnerships tied to the U.K. and other markets. The company signed commitments in the U.K. that include supplying chips to the Stargate data centre project in north-east England, a collaboration involving OpenAI, Arm and NScale. Huang and other technology executives, including Microsoft CEO Satya Nadella, are attending events in London connected to the state visit.

U.S.-China tensions over advanced semiconductors have involved a mix of export controls, licensing requirements and corporate agreements. Regulators in both countries have increasingly scrutinised the flow of advanced AI hardware and the software and data that run on it. Industry analysts say the measures reflect broader geopolitical competition over AI capabilities and infrastructure, and they are monitoring how corporate arrangements and regulatory guidance will affect deployment and investment.

Nvidia has said previously that it aims to comply with applicable laws and export controls while serving global customer demand. The company did not immediately provide new public comment beyond Huang’s remarks on Wednesday.

Any official confirmation from Chinese authorities about the Financial Times report would clarify whether the guidance is an immediate, enforceable prohibition or part of broader regulatory review. Nvidia’s customers and partners in China and elsewhere will watch for further details that could affect hardware deployment, data-centre operations and long-term procurement plans.

The situation underscores the challenges technology companies face in navigating competing legal, commercial and strategic priorities as governments seek to shape the global landscape for AI hardware and software.