Friday, August 7, 2026
DarkSubscribe
AI Infrastructure · News & Analysis
HomePolicyReport
Policy · Report

China restricts state-funded datacenters from purchasing foreign AI chips, accelerating domestic chip development.

Market fragmentation and supply chain realignment; non-Chinese infrastructure builders face reduced access to state-sector cloud expansion.
Trade pressSlicast · November 5, 2025 · Global · Source: bworldonline.com
importance 85

The Chinese government has issued guidance requiring new data center projects receiving state funds to use only domestically-made artificial intelligence chips, according to two sources familiar with the matter. In recent weeks, Chinese regulatory authorities have ordered data centers less than 30% complete to remove all installed foreign chips or cancel plans to purchase them, while projects in more advanced stages will be decided on a case-by-case basis. The move represents one of China's most aggressive steps yet to eliminate foreign technology from its critical infrastructure and achieve AI chip self-sufficiency, amid a pause in trade hostilities between Washington and Beijing.

The decision could substantially reshape the AI chip market, particularly for Nvidia, whose current share of the Chinese AI chip market is zero compared to 95% in 2022. US President Donald Trump said in an interview following talks with Chinese President Xi Jinping that Washington will "let them deal with Nvidia but not in terms of the most advanced" chips. The new guidance covers Nvidia's H20 chips, the most advanced AI chip the US firm is allowed to sell to China, and also applies to more powerful processors such as the B200 and H200, which remain available through grey-market channels despite US export restrictions. Excluding Nvidia and other foreign chipmakers including AMD and Intel from major state projects would eliminate a significant portion of their China revenue, even as discussions continue about resuming advanced chip sales to China.

China's AI data center projects have drawn over $100 billion in state funding since 2021, with most facilities receiving some form of state support for construction. Some projects have already been suspended before breaking ground as a result of the directive, including a facility in a northwestern province that had planned to deploy Nvidia chips and was being developed by a private technology company that received state funding. The sources did not specify whether the guidance applies nationwide or only to certain provinces, and declined to be named due to the sensitivity of the matter.

The directive opens market opportunities for domestic Chinese chipmakers including Huawei Technologies, Shanghai-listed Cambricon, and startups such as MetaX, Moore Threads, and Enflame. Products from these Chinese companies already rival some of Nvidia's offerings, but they have struggled to gain market share, as developers accustomed to Nvidia's reliable software ecosystem have been reluctant to adopt domestic alternatives. However, the move also risks widening the US-China gap in AI computing power, as US tech giants like Microsoft, Meta, and OpenAI have spent or allocated hundreds of billions of dollars to build data centers powered by Nvidia's most advanced chips. Meanwhile, leading Chinese chip manufacturers like SMIC are facing supply constraints due to US sanctions on semiconductor manufacturing equipment that have hit advanced chip production capacity.

Beijing's latest action follows a series of measures to reduce its dependence on US technology. In the current year, China discouraged local tech giants from purchasing advanced Nvidia chips over security concerns while displaying a new data center powered solely by domestic AI chips. In 2023, Beijing banned the use of Micron's products in its critical infrastructure, leading the largest US memory chipmaker to exit the server chip market in China. The US has justified its restrictions by alleging that the Chinese military would use advanced chips to increase its capabilities.

Read the original
China restricts state-funded datacenters from… · Slicast