The €75B headline vs. the contract that doesn't exist
The number doing the heavy lifting in every headline — up to €75 billion, 5 GW — is a stated ambition, not a signed and bonded commitment. Only €45 billion is anchored to a Phase 1 milestone for 3.1 GW by 2031, with the remaining 2 GW (and roughly €30B) gated on future negotiations [1]. Multiple outlets point out there is no documented €75B contract; the ceiling could shift as Phase 2 talks evolve and is contingent on a six-year milestone path [4][5]. That matters because SoftBank's recent track record is precisely about pledge-vs-delivery. The Decoder notes plainly that 'SoftBank's ambitious global spending has yet to actually materialize,' citing reports that the US Stargate project is stalling and lenders are pulling back from OpenAI-backed loans [5]. The Next Web is more blunt: 'Son's track record on vision is extraordinary. His track record on execution is more complicated' [4]. The financing context reinforces the skepticism — SoftBank reportedly scaled back a planned $10B OpenAI-stake margin loan to as low as $6B amid creditor hesitation [4], while still taking on roughly $40B in bridge loans for OpenAI [3]. Reddit's French-language communities surfaced the WeWork analogy almost immediately — supportive voices celebrated the digital sovereignty narrative, but skeptics consistently invoked SoftBank's Vision Fund history and questioned how much of the headline actually lands as built kilowatts. Read together: €45B is the number to underwrite; €75B is the number to discount.



