China's Renewable Energy Goals for AI: Challenges and Prospects
In the context of China's push to develop renewable energy for the artificial intelligence (AI) sector, grid operators are expressing concerns about potential risks. Efforts to increase renewable electricity supply for data centers could create significant challenges due to difficulties in forecasting peak demand at these facilities.
According to analysts and industry officials cited by Reuters, China's strategic goal of using renewable energy to meet most of the electricity demand for data centers by 2030 may not be practically feasible.
Challenges from Inflexible Data Center Demands
"As far as we understand, data centers cannot adjust their power consumption much," Pei Shanpeng, director of state-owned power giant State Power Investment Corporation, said at a recent industry conference in Beijing.
"GPUs are very expensive, so once purchased, operators want to use them as quickly and powerfully as possible," the official added.
This lack of flexibility in adjusting electricity demand creates significant challenges for the power grid, which needs to continuously balance supply and demand. High-intensity AI data centers often have unpredictable peak demand, making grid management difficult.
Innovative Solutions: Underwater Data Centers
To address energy issues, China has launched the world's first offshore wind-powered underwater data center. Using seawater cooling and renewable energy, this solution aims to reduce demand for energy, water, and land.
The prototype Shanghai Lingang underwater data center has a capacity of 24 MW, developed by HiCloud Technology and state-owned China Communications Construction. This project represents a new direction in optimizing energy efficiency for high-tech infrastructure.
Current Energy Situation for Data Centers in China
According to last year's report by the International Energy Agency (IEA), the electricity supply for data centers in China is primarily coal-based, with a rate of nearly 70% in 2025. Other energy sources include:
- Renewable energy: nearly 20%
- Nuclear energy: nearly 10%
- Natural gas: the remainder
The following table shows the energy allocation for data centers in China:
| Energy Type | Percentage (%) | Description |
|---|---|---|
| Coal | ~70% | Main electricity supply source |
| Renewable Energy | ~20% | Solar, wind, hydropower |
| Nuclear Energy | ~10% | Nuclear power |
| Natural Gas | ~0% | Supplementary energy source |
Future Renewable Energy Strategies
According to IEA forecasts, solar and wind power will add nearly 90 TWh of additional electricity for data centers by 2030. This growth is supported by:
- Increasing renewable energy share in the grid mix
- Regulations requiring data centers to be located in areas abundant with renewable energy
- Policies prioritizing data center construction in western China - a region rich in renewable energy resources
However, analysts and industry officials still believe the data center sector is not an ideal candidate for renewable energy due to its unpredictable peak demand characteristics.
In-depth Analysis of Technical Challenges
The core issue lies in the incompatibility between renewable energy characteristics and data center demands. Solar and wind energy are highly variable, depending on weather conditions, while AI data centers require stable and continuous power supply.
Additionally, data centers are often located in major urban areas to reduce network latency, while areas with abundant renewable energy potential are typically in remote western China. This geographical distribution creates challenges in electricity transmission and grid infrastructure.
Conclusion: Balancing Ambition and Reality
Although China has ambitious goals for transitioning to renewable energy for data centers, the technical and operational challenges cannot be overlooked. The development of innovative solutions like underwater data centers is a step in the right direction, but further research and investment in energy storage technologies, smart grid management, and other energy efficiency solutions are needed.
Balancing the ambition of sustainable AI development with grid operational reality will be key to China achieving its renewable energy goals for data centers by 2030.