Fluidstack Targeting EU Sovereign Compute
CoreWeave of Europe
This is really a race to become Europe’s default sovereign AI utility, not just another GPU reseller. Fluidstack is moving from brokering spare chips for startups to operating dedicated clusters for governments and model companies, which matters because the biggest EU buyers increasingly want compute that sits in Europe, runs on long term contracted capacity, and can be financed like infrastructure rather than bought one server at a time.
-
The strongest evidence is customer and project mix. Fluidstack signed an MoU with the French government in February 2025 to build a 1 GW AI supercomputer in France, aimed at one of Europe’s largest sovereign AI clusters, and that shifts it into the kind of nation scale capacity market where leadership is decided.
-
Leadership in the EU will not look like AWS style horizontal dominance. It will look like owning scarce local GPU supply for a small set of buyers, governments, frontier model labs, and regulated enterprises. Mistral already uses Fluidstack for Mistral Compute, showing how European model companies can bundle local models with local infrastructure.
-
The competitive opening is timing. CoreWeave is larger globally, with $3.52B in trailing revenue as of June 2025 and 250,000 plus GPUs across 33 facilities in the US and Europe, but Europe’s local providers still hold only about 15% of cloud infrastructure revenue, which leaves room for a focused specialist to win the sovereign slice as policy pressure rises.
From here, the market is likely to split in two. US hyperscalers will keep most general cloud spend, while a smaller but strategically important sovereign compute layer forms around European governments, labs, and regulated industries. If Fluidstack keeps converting financing into dedicated EU capacity and anchor contracts, it can become the local infrastructure backbone those buyers standardize on.