China has issued new rules requiring data centres that receive government funding to use only domestically-made artificial intelligence chips, two sources familiar with the matter told Reuters. The move marks one of Beijing’s strongest efforts yet to remove foreign technology from its critical infrastructure and accelerate its goal of AI chip self-sufficiency.
Removal of Foreign Chips From New Projects
According to the sources, Chinese regulators have instructed data centre projects that are less than 30% complete to remove any installed foreign chips or cancel planned purchases. Projects at more advanced stages will be assessed individually. The new rule could significantly impact foreign chipmakers such as Nvidia, AMD and Intel, who have long supplied chips for China’s fast-growing AI data centre sector.
It remains unclear whether this guidance applies nationwide or to certain regions. Regulators, including the Cyberspace Administration of China and the National Development and Reform Commission, have not responded to requests for comment.
Impact on Nvidia and U.S.-China Chip Tensions
The new directive could deal a major blow to Nvidia, whose AI chips, including the H20, B200 and H200 models, are widely used in AI data processing. Nvidia had been seeking to recover its market share in China after U.S. export controls sharply limited its sales. However, these restrictions already reduced Nvidia’s market share in China to zero from 95% in 2022.
The move comes shortly after U.S. President Donald Trump said that Washington would allow China to buy some Nvidia chips, but not the most advanced ones. Despite that, China’s latest measure effectively shuts out foreign chip suppliers from government-funded projects, creating new opportunities for local firms such as Huawei, Cambricon, MetaX, Moore Threads and Enflame.
Boon and Risks for Domestic Firms
Since 2021, Chinese AI data centre projects have received more than $100 billion in state funding, according to a Reuters review. Many of these projects will now need to rely solely on domestic chip producers. While this could boost local manufacturers’ sales, it also poses challenges. Developers accustomed to Nvidia’s reliable hardware and software may find the transition difficult.
Some projects have already been paused as a result of the directive, including one planned in a northwestern province that had intended to deploy Nvidia chips. Meanwhile, U.S. sanctions continue to restrict China’s access to advanced chipmaking tools, limiting the country’s capacity to produce high-end processors.
Beijing has made similar moves before, such as discouraging purchases of Nvidia’s top chips and banning Micron products from critical infrastructure. The latest rule deepens the divide between U.S. and Chinese technology ecosystems, reinforcing Beijing’s long-term plan to achieve full technological independence.
(with inputs from Reuters)
