China has reportedly intensified efforts to purge foreign technology from key digital infrastructure. According to individuals familiar with new government orders, Beijing has instructed that any state-funded data centre project must rely exclusively on Chinese-made artificial intelligence chips. Facilities still early in construction must remove foreign chips already installed or cancel related procurement plans. Projects near completion will be reviewed individually.
This directive marks a significant escalation in China's push for semiconductor independence. It comes at a moment of relative calm in US-China trade tensions, yet underscores China’s strategic resolve to secure its AI supply chain and reduce reliance on US technology. Washington’s export controls, especially restrictions on Nvidia’s advanced chips, have long frustrated Beijing. In turn, China now appears to be leveraging domestic investment and its regulatory authority to accelerate the adoption of homegrown hardware.
The US and China have been competing fiercely over high-end computing capacity. President Donald Trump recently stated that the US would only permit limited Nvidia chip sales to China, ensuring Beijing cannot access the cutting-edge. The latest move from Chinese authorities severely curtails Nvidia’s prospects in the world’s second-largest market, while opening the door wider for domestic contenders like Huawei and other Chinese chip startups.
It is uncertain whether this policy applies nationwide or only to certain provinces. Government bodies that typically supervise such matters have not publicly commented. Alongside Nvidia, companies including AMD and Intel are likely to face lost market opportunities in China’s expanding data-infrastructure sector.
State investment in AI computing hubs in China has exceeded $100 billion since 2021. Because most data-centre initiatives receive some public support, the order may affect a large portion of ongoing projects. In at least one case, a major facility that had planned to deploy Nvidia chips was paused before construction began. By restricting the participation of US suppliers, China hopes to insulate its technology ecosystem from geopolitical pressure, mirroring past actions such as banning Micron components from sensitive networks.
Despite lobbying efforts by Nvidia leadership arguing that controlled chip sales benefit US interests, the company’s foothold in China has collapsed from a dominant market share to zero. Now even Nvidia’s H20 chips, the most advanced permitted under US export rules, fall under the new ban. More powerful processors, officially restricted but available through grey-market channels, are also covered.
For China’s domestic semiconductor firms, the mandate is a major commercial opportunity. Huawei and emerging competitors may now gain access to contracts previously dominated by American suppliers. However, there are also risks. While Chinese chips are improving, developers continue to favour mature Nvidia ecosystems, and China faces manufacturing constraints due to US curbs on advanced fabrication equipment. Meanwhile, US companies are investing massive resources into AI clusters powered by cutting-edge processors, leaving Chinese data centres potentially trailing in the global performance race.
In sum, the policy signals China’s determination to harden its digital infrastructure against foreign dependence. It simultaneously forces rapid adoption of domestic AI processors while intensifying the technological divide between the two superpowers.
