The paper argues that global AI competition has extended from technological rivalry in algorithms and chips to a broader contest over power infrastructure. As energy consumption in AI training and inference continues to rise, global data center electricity demand has grown rapidly, making power supply a key variable determining the investment structure and geographical distribution of the AI industry.

Although Europe has strong policy ambitions to expand computing capacity, it faces severe electricity constraints. These include a temporal mismatch between power infrastructure development and data center construction cycles, a spatial mismatch due to the over-concentration of computing resources in core hubs, and a structural mismatch between the intermittency of renewable energy and the stable electricity demand required by AI. These issues are compounded by aging power grids and the dual challenge of electricity prices that are significantly higher and more volatile than those in China and the United States, leading to a continued decline in Europe’s share of the global data center market.

In response, the European Union and its member states have implemented multi-dimensional measures, including accelerating grid expansion and reforming grid access rules, increasing investment in nuclear energy and promoting self-built microgrids for data centers, improving electricity market mechanisms to enhance system flexibility, and introducing energy efficiency standards to strictly control the energy consumption of computing power.

Finally, the paper predicts that the geographical distribution of computing power in Europe will become more balanced, that the EU’s goal of tripling computing capacity is unlikely to be achieved, and that the long-term development of the industry will depend on balancing computing expansion with the energy transition. Constraints related to water resources and public environmental concerns will also continue to play a limiting role.

Translated by Gaoyuan LI

Full text in Chinese available at:

https://fddi.fudan.edu.cn/d5/87/c18965a775559/page.htm