In a move that underscores China’s growing determination to become technologically self-sufficient, Beijing has reportedly ordered all data-centre projects that receive government funds to phase out the use of foreign-made artificial-intelligence chips. This new instruction, which insiders say was communicated quietly to provincial authorities and state-run companies, signals one of the strongest policy pushes yet toward ending China’s reliance on imported computing hardware.
According to people familiar with the matter, local administrators were told that data-centre projects supported by public investment must not install or purchase foreign AI chips. Projects that are still in early stages — less than 30 percent complete — have been asked to remove any imported hardware already fitted or cancel remaining foreign orders. Those that are further along will reportedly be reviewed individually, meaning that in some special cases there might be room for exceptions, though that remains unclear.
This policy forms part of a much broader effort by the Chinese government to reshape its digital ecosystem and tighten control over the backbone of its artificial-intelligence infrastructure. While the United States and China have shown occasional signs of easing trade tensions, the competition in advanced computing technology remains fierce. China’s leadership, seeing AI as a core driver of future power, wants to ensure that its computing capability cannot be compromised or restricted by foreign sanctions.
For years, China’s AI and cloud-computing industries have relied heavily on advanced chips designed by American firms such as Nvidia, AMD and Intel. These high-performance processors are the brains behind large-scale data processing, machine learning and generative-AI systems. But as Washington has tightened export restrictions — limiting the sale of top-tier chips that could be used for military or surveillance purposes — Beijing’s sense of urgency has grown. Officials now see dependence on foreign semiconductors as both an economic weakness and a national-security risk.
By telling state-funded projects to switch exclusively to Chinese-made AI chips, Beijing aims to nurture its domestic semiconductor sector while protecting strategic autonomy. Analysts describe this as an important step in building a parallel tech ecosystem — one that can function even if access to Western components is completely severed. The new policy also serves as a signal to local governments, private investors and industrial planners: China’s future digital infrastructure will be built on home-grown silicon, no matter the short-term cost.
Sources familiar with the new instruction said some ongoing projects have already been affected. In one example, a planned facility in the country’s north-western region — initially scheduled to use Nvidia hardware — has been paused indefinitely after receiving the order to comply with the domestic-only rule. Others may face expensive re-designs as they replace imported chips with Chinese alternatives.
The extent of this new policy is still uncertain. Insiders note that it may not apply to every province or private-sector project, as enforcement mechanisms differ regionally. What is known, however, is that Beijing’s intention is to make the policy nationwide over time, covering all publicly funded data-centre initiatives.
This campaign comes alongside a range of other measures aimed at encouraging the use of local technology. Reports indicate that some provinces are offering generous electricity-cost discounts — in some cases up to 50 percent off — to AI data centres that adopt Chinese-made processors. These incentives are part of a national industrial plan to boost domestic chip production, reduce import dependency and attract developers to invest in Chinese computing platforms.
China’s state funding for AI data-centre infrastructure has already surpassed the 100-billion-dollar threshold since 2021. Much of this capital is linked to regional projects built to support “digital-economy zones,” where AI, cloud and big-data companies cluster. Because nearly all such projects include some form of public investment, the new policy could impact a large share of new facilities planned for the next two years.
For foreign chip manufacturers, this development poses serious challenges. Nvidia, which once dominated China’s AI-chip market with a reported 95 percent share in 2022, is likely to see its presence shrink sharply unless it adapts to new restrictions. The company has already attempted to design “China-specific” chips that comply with U.S. export rules, but those versions have struggled to match the performance of its global products. With Beijing now cutting off state-backed projects from using any foreign chips, Nvidia’s path to maintain even partial access to the Chinese market is narrowing.
Meanwhile, domestic chipmakers are seizing the opportunity. Huawei Technologies, through its Ascend 910C chip, is positioning itself as a major local alternative to Nvidia’s popular AI GPUs. Other Chinese semiconductor firms such as Biren, Moore Threads and Tecorigin are racing to improve their designs and production capacity. Beijing has also set a target to triple its AI-chip output by 2026, showing just how serious the government is about reducing dependency on foreign hardware.
Industry experts say that, beyond simple substitution, China is trying to develop a fully integrated domestic AI-infrastructure ecosystem — spanning everything from chip design and fabrication to data-centre construction, cloud software, and algorithmic research. The government’s hope is that by investing across all layers simultaneously, it can achieve “technological sovereignty” within a decade.
This transformation, however, will not be easy. Analysts note that China still lags behind leading global chip producers in certain key areas, particularly in cutting-edge lithography and design software. Developing chips that match the performance of Nvidia’s H100 or AMD’s Instinct series remains a major challenge. Yet, with enormous state support and policy protection, Chinese manufacturers are expected to close some of that gap over time.
The geopolitical implications are equally significant. By tying access to its large AI market to the use of domestic chips, China is effectively redrawing the map of global supply chains. Countries and companies that wish to continue doing business in the Chinese digital sector will increasingly have to adapt to its new self-reliance framework. This could further fragment the global tech landscape into two partially disconnected systems — one dominated by U.S. and Western suppliers, and another centred on Chinese hardware and standards.
What makes this move particularly notable is its timing. Around the world, AI is becoming an essential engine for innovation and productivity — from autonomous driving to healthcare analytics and national defence. Control over AI-computing infrastructure is now a strategic advantage. By ensuring that its AI data centres depend on domestic components, China is not only protecting its technological independence but also reinforcing its ability to dictate the pace and direction of future AI development within its borders.
Ultimately, the decision to restrict foreign chips in state-funded data centres reflects China’s long-term vision: to stand at the forefront of global AI innovation without external constraints. It demonstrates how deeply technology has become intertwined with national strategy — and how the race for AI supremacy is shaping the next chapter of global economic competition.
Whether this policy will deliver the intended results remains to be seen. But one thing is clear: Beijing’s new directive marks a turning point in the evolution of China’s digital economy. It signals a future in which every byte of data and every watt of computing power is expected to serve not just business efficiency but also the country’s broader goal of technological independence and global influence.