Big Tech's AI ambitions pose a major power test for Europe

SoftBank's recent commitment to invest 75 billion euros in AI infrastructure in France highlights the nation's emerging role as a pivotal center for artificial intelligence. However, analysts warn that the substantial energy requirements of these technologies could strain the European power grid. On Saturday, SoftBank revealed plans to construct 3.1 gigawatts worth of AI data centers in the northern region of Hauts-de-France, with new facilities planned for Dunkirk, Bosquel, and Bouchain, expected to be operational by 2031. France, where over 60% of electricity is generated from nuclear power, is strategically positioned to handle such energy-intensive initiatives, especially as Europe faces soaring electricity prices for industrial use. This energy dependency is critical, as the costs associated with running data centers make investments sensitive to electricity rates. Experts have indicated that companies are increasingly drawn to regions with more affordable energy prices. In fact, energy costs for industrial operations in Europe averaged about twice those in the U.S. last year, and were approximately 50% higher than in China and India, as reported by the International Energy Agency. Analysts suggest that enhancing the role of nuclear power could be vital for European countries aiming to meet the energy demands of data centers. Currently, nuclear energy constitutes only 11.8% of Europe’s overall energy mix, with oil and gas still making up over a third, according to Eurostat data. Big Tech companies in the U.S. have shown interest in small modular reactors (SMRs) as a solution for their energy needs. Amazon, for instance, announced a partnership with Dominion Energy in 2024 to explore SMR development, while Google entered into a similar agreement with Kairos Power and the TVA for a new nuclear facility in 2025. These smaller, factory-built reactors provide a more efficient alternative to traditional power plants, typically generating 300 megawatts or less and designed for quicker, cost-effective deployment. Despite their advantages, rolling out SMRs on a large scale presents challenges. Data center operators are currently assessing their energy needs over the next decade, as this will dictate their operational viability. Tania Arora, a partner at Baker McKenzie, remarked that while many operators are considering SMRs, the lack of any operational units outside of China and Russia poses a significant risk. In addition to energy availability, the talent pool in London is another key factor driving Big Tech's investment in Europe. Several AI companies from the U.S. have announced plans to expand in London, with Nvidia-backed Runway declaring its intention to establish a new European headquarters there. Anthropic, OpenAI, and Google are also planning to open new offices in the city, further solidifying London's status as a tech hub.