The race for AI infrastructure dominance just got a jolt of adrenaline that stretches all the way from Tokyo to the hills of northern France. SoftBank Group has announced plans to invest up to €75 billion (roughly $87 billion) to build out data center capacity across the country, marking what the Japanese conglomerate calls its largest AI infrastructure play in Europe to date.
The scale is staggering. SoftBank aims to develop and operate up to 5 gigawatts of additional data center capacity. The first phase alone — targeting sites in Dunkirk (Loon-Plage), Bosquel, and Bouchain — will deliver 3.1 gigawatts to the Hauts-de-France region by 2031. To put that in perspective, a single gigawatt can power roughly 750,000 homes. This is not a toe-dip; it’s a cannonball.
Why France, Why Now
France has been working overtime to position itself as Europe’s AI hub, and this deal validates that strategy in a big way. French economic minister Roland Lescure called the announcement a “testament to President Emmanuel Macron’s ambition to position France as a leading destination all along the AI value chain.” And he’s not wrong — the country has been pushing hard on everything from AI talent development to energy infrastructure that can handle the colossal demands of modern machine learning workloads.
But there’s a deeper story here. SoftBank is both an investor in and a customer of OpenAI, and the company clearly sees AI compute as the most strategic asset of the next decade. By building data centers in France rather than, say, Ireland or Germany, SoftBank is betting on the country’s nuclear-heavy energy grid — which offers both reliability and relatively low carbon intensity — as a competitive advantage for power-hungry AI training clusters.
The Regulatory Tightrope
None of this is happening in a vacuum. While SoftBank ploughs billions into French soil, the mood around data center construction is shifting in other parts of the world. In the United States, public opposition to data center builds is heating up over environmental concerns, questions about grid strain, and the impact on utility prices for everyday consumers.
SoftBank is navigating these tensions with a two-pronged approach: in Ohio, it’s building a data center powered by a new 9.2-gigawatt natural gas plant. In France, it’s tapping into a grid that already runs largely on nuclear and renewables. Different strategies for different regulatory realities — but the appetite for compute capacity is the same everywhere.
What This Means for Startups
For European founders, this is both an opportunity and a warning signal. On the opportunity side, billions flowing into French data infrastructure means lower latency, better access to compute, and a stronger ecosystem for AI-native startups building in Europe. If you’re training models or running inference-heavy applications, having world-class data centers in your backyard is a genuine competitive advantage.
On the warning side, the sheer scale of capital required to play in the AI infrastructure game is stratospheric. SoftBank is spending more on French data centers than most countries’ entire annual tech budgets. This concentration of resources raises questions about who gets access to compute and at what price — a dynamic that matters deeply for early-stage startups without enterprise negotiating power.
The Takeaway for Founders
The era of assuming compute will be cheap, abundant, and local is ending. As massive players like SoftBank reshape the infrastructure landscape, startups need to think strategically about where they locate, which cloud regions they prioritize, and how they negotiate compute costs. The winners won’t just be the best product builders — they’ll be the ones who understand the physical geography of AI before their competitors do.
France just became a much more important dot on that map.
This article is based on reporting from TechCrunch.
