Technology & AIAnalysis

China Connects Green Power to AI Hubs as Computing Electricity Surges

Computing centers are projected to consume 800 billion kilowatt-hours by 2030, prompting direct green power integration.

Share
A white robot showcasing modern design on a sleek dark surface.
Photo by Pavel Danilyuk on Pexels

The Brief

China is accelerating the integration of renewable energy grids with national computing hubs to manage the power demands of expanding artificial intelligence systems. According to figures reported by People's Daily, daily token consumption across the country grew from 100 billion in early 2024 to 140 trillion in March 2026, driving computing power centers to consume 170 billion kilowatt-hours in 2025. The National Energy Administration projects computing electricity use will reach 800 billion kilowatt-hours by 2030, spurring policy support and direct green power supply infrastructure across western resource hubs.

Why it matters

The commercial scaling of large language models is fundamentally constrained by electricity generation and grid stability. Computing power centers require stable voltage and frequency while consuming immense power volumes. China's push to directly link renewable energy bases to western data center clusters tests whether high-demand digital infrastructure can expand rapidly without driving up carbon emissions or destabilizing local utility grids.

China context

China's 'Eastern Data, Western Computing' strategy establishes eight national computing hubs and ten data center clusters, aiming to align energy-intensive data processing with western renewable resources. By late spring 2026, non-fossil sources accounted for 62% of China's installed power capacity, and green certificates traded reached 930 million in 2025. National planners are now formalizing 'computing-electricity coordination' through joint ministry initiatives to establish direct green power connections.

Editor's View

EDITOR'S VIEW — Analysis and inference, not factual reporting. The operational challenge in China's AI infrastructure strategy is shifting from raw compute deployment to grid-level synchronicity. While western hubs like Ningxia, Gansu, and Inner Mongolia offer abundant solar and wind power, data center workloads require uninterrupted, high-precision power supply. The success of direct-supply initiatives will depend on whether storage and grid flexibility mechanisms can keep pace with volatile AI inference and training workloads.

What to watch

  • Approval and commissioning timelines for direct green power connections at national computing hubs in western provinces.
  • Grid flexibility and energy storage policies designed to manage the intermittent supply of renewable energy powering large-scale data centers.
  • Expansion of green power certificate trading mechanisms for national data center operators.

Key Takeaways

  • 1China's daily AI token queries reached 140 trillion in March 2026, up from 100 billion in early 2024.
  • 2Computing power electricity demand is projected to reach 800 billion kWh by 2030, representing about 6% of national electricity consumption.
  • 3Non-fossil energy accounted for 62% of China's total installed capacity of over 4 billion kW by late spring.
  • 4New data centers in western clusters like Horinger and Qingyang report green power consumption rates above 80%.
  • 5China Datang brought a 500,000-kW solar project online in Zhongwei, Ningxia, to supply computing facilities directly.
China is intensifying efforts to align its power infrastructure with the rapid expansion of artificial intelligence, as surging computing demand increases pressure on electricity supply systems. Data reported by People's Daily indicates that daily token queries across the country surged from 100 billion in early 2024 to 140 trillion in March 2026. In 2025, national computing power centers consumed 170 billion kilowatt-hours (kWh) of electricity, accounting for approximately 1.6% of the country's total power consumption. Projections by the National Energy Administration indicate that computing-related electricity use will expand by more than 100 billion kWh annually during the 15th Five-Year Plan period, reaching approximately 800 billion kWh by 2030, or roughly 6% of total national electricity demand. Data processing hardware demands high power stability, where minor fluctuations in voltage or frequency can lead to transmission errors and equipment disruptions. To address these demands, policymakers are pairing national power generation capacity with targeted spatial planning. As of May, China's total power generation capacity exceeded 4 billion kilowatts, with non-fossil sources accounting for 62% of the total. Under the 'Eastern Data, Western Computing' initiative, China has organized eight computing hubs and ten national data center clusters. In western hubs such as Horinger in Inner Mongolia and Qingyang in Gansu, newly constructed data centers have achieved green electricity consumption shares exceeding 80%, supported by a green power certificate market that recorded 930 million transactions in 2025. Policy support has also expanded. Earlier this year, 'computing-electricity coordination' was formally designated as a key task in the Government Work Report. In April, four central government agencies, including the National Development and Reform Commission, issued an action plan to coordinate the layout of major renewable energy bases with national computing hubs. In Ningxia's Zhongwei, the China Datang Zhongwei Cloud Base 500,000-kilowatt photovoltaic project commenced operations as the country's first large-scale direct green power supply facility dedicated to computing infrastructure.