SoftBank’s $85B France Bet Puts Power at Center of AI Race

extract

SoftBank’s $85B France Bet Puts Power at Center of AI Race The company plans up to 5 GW of AI infrastructure in France, where EDF is turning former power plant sites into data center campuses and electricity has become a competitive advantage. SoftBank plans to invest up to €75 billion ($85 billion) in French AI infrastructure and develop up to 5 GW of data center capacity across the country, a massive buildout that underscores how access to power is becoming one of the industry’s most important competitive advantages. The Japanese technology giant said the initiative will begin with three sites in Dunkirk, Bosquel, and Bouchain, targeting 3.1 GW of capacity by 2031, before expanding nationwide to 5 GW. The first phase calls for €45 billion ($51 billion) in investment across the three sites by 2031. The announcement arrived alongside a separate decision by French utility EDF to select SoftBank as the preferred developer for a 400 MW data center campus at the site of its former Bouchain thermal power plant. France is only the latest stop in SoftBank’s AI infrastructure campaign. The company is a leading investor in Stargate, the US AI infrastructure initiative targeting up to $500 billion in investment and 10 GW of data center capacity. OpenAI, Oracle, and SoftBank said in September that Stargate had already reached nearly 7 GW of planned capacity and more than $400 billion in investment commitments. SoftBank has also invested in SB Energy projects tied to data center and power infrastructure development. Power first “France’s position as a major energy producer and exporter was absolutely decisive in our decision,” SoftBank founder and CEO Masayoshi Son said in a statement announcing the investment. "By combining abundant, competitive, decarbonized power with strategic sites and industrial expertise, France has all the assets required to become one of the world’s AI capitals.” “There are many factors that go into site selection – proximity to customers, availability and cost of real estate, availability and cost of power, networking infrastructure, ease of doing business, local financial incentives, political stability, minimizing the impact of natural hazards, local regulatory and data residency requirements,” said John Dinsdale, chief analyst and research director at Synergy Research Group. “But in the last couple of years, availability of power has become an ever more important decision criterion,” he told Data Center Knowledge. Jigar Shah, former director of the US Department of Energy’s Loan Programs Office, said the announcement reinforces the growing role of electricity availability in AI infrastructure development. “Power availability has become the primary site-selection variable for large-scale AI infrastructure investment,” Shah told Data Center Knowledge. Access to power and water has become critical to data center development, according to Roy Illsley, chief analyst at Omdia. “In some countries there is the power, but not in the correct place,” Illsley told Data Center Knowledge. Cost and reliability also play a role in site selection decisions, he said. Projects of this scale rise or fall on access to electricity, water and the infrastructure needed to deliver both at campus scale. Utility strategy For years, data center developers identified land and then sought power. Now, many are chasing power first. The Bouchain project offers a clear example of how that search is affecting site selection. However, rather than simply providing electricity to a third-party development, EDF packaged a former power plant site with existing grid infrastructure and launched a competitive process to identify a developer. SoftBank emerged as the preferred bidder for a campus expected to support up to 400 MW of capacity. EDF has spent the past year marketing utility-owned sites for data center development, using existing grid infrastructure and industrial properties to attract large AI projects. At Bouchain, that strategy has landed one of the world’s most aggressive AI infrastructure investors.