VN — VnExpress
How China is tackling AI's thirst for electricity
Policy1 min read
China is using many plans to feed the growing power needs of AI data centers with its large supply of renewable energy. The main idea is to move computing work to places where clean power is cheap and plentiful.
What is happening
Chinese officials are using several ideas to make better use of clean energy, such as wind and solar power. AI data centers need huge amounts of electricity, so China is linking computing power with the power grid. In 2026, China added this plan to its government work report and its 15th five-year plan (2026 to 2030) for the first time.
The 'East Data, West Computing' plan
China started this plan in 2022 to balance computing work and electricity across the country. Urgent tasks, such as live chatbots and real-time AI answers, stay near users in eastern China. Slower tasks, such as training AI models and processing data offline, can move to data centers in the west. The west has cheap renewable power and cold weather that helps cool computers.
Key facts from the pilots
- The Ulanqab wind and solar farms have 300 megawatts (MW) of capacity, with a 45 MW energy storage system.
- A test in the Lingang area of Shanghai moved AI computing tasks between network nodes in under 3 minutes. The cluster had 100 processing cards.
- The Yajiang-1 AI supercomputer center sits at nearly 3,600 meters above sea level in Tibet.
Why it matters
AI needs a lot of electricity, and power grids can get overloaded at peak times. If computing work can shift between regions, China can use more clean energy and keep services running. It can also lower power bills and reduce the strain on the eastern grid.
What to watch
- Whether the pilot projects can expand to more data centers across China.
- How much the new long-distance data links cut costs and network delays.
- Whether storage batteries can keep wind and solar power steady over longer periods.