Mistral AI wants to build 1 gigawatt of European compute by 2030 — and lock in customers now.

Mistral AI aims to establish a European AI infrastructure worth 1 gigawatt by 2030, positioning itself as a provider of critical resources for enterprises and governments. The company announced plans to offer regional inference endpoints, a Priority Tier with uptime guarantees, and a coalition of European enterprises committing to multi-year compute contracts, which will fund 200 megawatts of infrastructure by the end of 2027 and 1 gigawatt by the end of 2030.

Mistral will also begin hosting third-party open models, with the first being GLM-5.2 from Chinese AI lab Z.ai. This shift in focus from model development to providing AI infrastructure comes with significant capital requirements, estimated at $38 billion for a typical one-gigawatt data center. To finance this expansion, Mistral is forming an anchor group of European enterprises that will commit to long-term contracts, or European Compute Units (ECUs), which will guarantee their access to the infrastructure.