China's Renewable Energy Challenge for AI Data Centers
China is advancing its strategic plans to utilize renewable energy to power artificial intelligence (AI) data centers, yet electricity grid operators are expressing concerns about potential risks stemming from the difficulty in predicting peak electricity demand from these facilities.
According to industry analysts and officials, China's strategic goal of meeting most of the electricity demands for data centers with renewable energy by 2030 may face significant implementation challenges.
Difficulties in Managing Data Center Electricity Demand
"As far as we understand, data centers really cannot adjust much in terms of electricity load," Reuters quoted Pei Shanpeng, Director of the State Power Investment Corporation, when speaking at a recent industry conference in Beijing.
"GPUs are very expensive, so once purchased, operators want to use them as quickly and intensively as possible," the official added.
Efforts to Power Data Centers with Renewable Energy
China is planning to leverage the renewable energy boom to power its data centers. The country has recently launched the world's first underwater data center, operated by offshore wind power, utilizing seawater for cooling and renewable electricity to reduce demand for energy, water, and land.
The Shanghai Lingshang underwater data center, with a capacity of 24 MW, was developed by HiCloud Technology and state-owned China Communications Construction Group.
Data on Electricity Supply for Data Centers in China
According to a 2022 report by the International Energy Agency (IEA), electricity supply for data centers in China will predominantly come from coal, accounting for nearly 70% by 2025, followed by renewable energy at nearly 20%, nuclear power at nearly 10%, and natural gas making up the remainder.
By 2030, solar and wind could add nearly 90 TWh of additional electricity for data centers, "supported by the increasing share of renewable energy in the electricity mix, provincial-level renewable energy quota requirements, and policies prioritizing the construction of data centers in western China - where renewable energy resources are abundant," the IEA stated.
Electricity Supply Sources for Data Centers in China (2025)
| Energy Source | Percentage |
|---|---|
| Coal | 70% |
| Renewable Energy | 20% |
| Nuclear Power | 10% |
| Natural Gas | Remainder |
Challenges in Renewable Energy Integration
However, analysts and industry officials argue that the data center sector is not well-suited for renewable energy due to a lack of transparency regarding peak electricity demand from these major electricity consumers.
The core issue lies in the fact that data centers, particularly those using GPUs for AI, have high and stable electricity demands that are difficult to flexibly adjust according to the output of intermittent renewable energy sources like solar and wind.
The inability to accurately predict peak electricity demand from data centers presents significant challenges for grid operators in balancing supply and demand, especially as the share of renewable energy in the electricity system continues to grow.
Conclusion
While China has ambitious goals for using renewable energy to power its data centers, technical challenges regarding load adjustment capabilities and peak demand forecasting could pose substantial barriers to achieving the 2030 targets.
To address this issue, solutions such as developing large-scale energy storage technologies, improving demand forecasting capabilities for data centers, or designing more flexible data centers may need to be considered.
In the meantime, projects like the Shanghai Lingshang underwater data center demonstrate innovative approaches to reducing energy consumption, but additional solutions are still needed to address the challenges of integration with the power grid.