SoftBank’s €45B Phase in €75B France AI Data Center Push, Reboots Son

SoftBank Group founder Masayoshi Son told La Tribune du Dimanche that the group will invest €45 billion over the next five years to develop AI data‑center capacity in France, with an overall program that could expand to €75 billion. The initial phase is expected to deliver roughly 3.1 gigawatts (GW) of computing capacity in the Hauts‑de‑France region.

According to the interview and company statements, the rollout targets multiple sites including Le Bosquel and Dunkirk, with some facilities anticipated to be operational by 2028 and 2031. The €45 billion initial outlay is roughly $52.5 billion; SoftBank envisages further expansion across France to reach a total of about 5 GW and a potential €75 billion commitment over time. Strategic partners named in coverage include state energy provider EDF and Schneider Electric.

The scale of the plan underscores France’s appeal as a location for energy‑intensive AI infrastructure: its largely nuclear power mix offers competitive industrial power prices and grid capacity advantages for large data centers. The announcement, timed around France’s Choose France investment forum, is likely to boost local employment and downstream industrial projects while also prompting scrutiny of grid capacity, permitting timelines and regional supply chains.

SoftBank’s move is part of a broader global push into AI infrastructure. The group has increased capital allocation to computing and AI‑related assets and has already committed substantial sums to AI ventures, reinforcing a strategy that pairs software and cloud investments with physical compute capacity. For Europe, the project represents a notable step toward reducing dependence on non‑European cloud and compute providers.

Market observers say the immediate implications will be felt across energy, construction and data‑center equipment markets, with potential upward pressure on project financing needs and on local procurement. Over the medium term, a successful buildout could reshape the regional data‑center landscape and accelerate industrial policies aimed at on‑shoring critical AI infrastructure; however, execution risks — from permitting to grid upgrades — remain the main uncertainties to monitor.