A new government directive aims to solve the massive energy challenge of China's AI boom by directly connecting data centers to dedicated nuclear and hydrogen power sources.
Back
China to explore direct nuclear power for 4 agencies' AI ambitions
Four of China’s top agencies have jointly issued an action plan to explore powering computing facilities directly with dedicated energy sources like nuclear and hydrogen, aiming to solve the growing power-demand crisis driven by the artificial intelligence sector.
The plan, announced by the National Energy Administration, the National Development and Reform Commission, the Ministry of Industry and Information Technology, and the National Data Bureau, establishes a formal framework for a “computing–electricity coordination” model.
The proposal encourages new data centers to be built with dedicated power supplies and to configure these sites with advanced energy storage to improve stability. The move follows a real-world test in the city of Zhongwei, where a 500 megawatt solar plant is already directly supplying a data center cluster, with plans to expand to 4.6 gigawatts of co-located wind, solar, and storage, according to Datang Corp.
This initiative is a core part of China's "East Data, West Computing" strategy, designed to shift power-hungry data centers to western regions rich in energy resources. If the direct-power model proves scalable, it could provide a template for slashing both the carbon footprint and operating costs for China's massive data center expansion, creating a significant advantage in the global AI race.
The AI Power Dilemma
The explosive growth of artificial intelligence has created a global energy challenge for technology giants. In the U.S., Microsoft is reportedly reconsidering its goal to match 100% of its hourly energy use with clean power by 2030, as the rapid buildout of AI data centers strains its ability to meet the target. As tech companies expand their data center fleets, many are turning to natural gas to meet the intense power demand, highlighting the difficulty of aligning AI growth with clean energy goals using conventional grid infrastructure.
From Blueprint to Reality
China's plan is not merely theoretical. In Zhongwei, Ningxia, China Datang Corp. has already commissioned the country's first large-scale project designed to directly supply a data center under this coordinated model. The 500 MW solar plant is expected to generate 970 gigawatt-hours annually, covering about 50% of the cloud facility's demand. Once the project's first phase is complete with an additional 1.5 GW of wind power, it is projected to generate 4.3 terawatt-hours annually—far exceeding the data center's expected 2.29 TWh consumption. This project serves as a real-world test of matching large-scale digital loads with dedicated, co-located renewable generation, moving beyond certificate-based green energy claims to a physical infrastructure model.
The government-led initiative is expected to stimulate major investment in China's nuclear, hydrogen, and energy storage sectors. It offers a dual benefit: securing a stable, long-term power source for its technology industry while simultaneously creating a new, dedicated demand base for its advanced energy projects.
This article is for informational purposes only and does not constitute investment advice.
[1] Four departments: Explore direct power supply for computing facilities from energy sources like nuclear power and hydrogen[2] Microsoft's AI data center push is colliding with its clean power goals[3] AI Data Center Growth Is Now a Power Infrastructure Problem[4] AI data center boom is leaving consumer electronics short of chips − even though they don't use the same kinds[5] China Datang brings online 500 MW solar farm for direct data center supply
