Mistral AI is turning its push for European AI sovereignty into a bigger infrastructure bet, giving customers more control over where AI runs while seeking long-term commitments to build up to 1 gigawatt of European compute capacity by 2030.
The French AI company said Tuesday that its Mistral Regional Endpoints are now generally available, allowing customers to choose whether inference runs in Europe or the U.S. The feature is designed to help organizations manage data residency, regulatory and latency requirements.
Mistral has also introduced its Priority Tier in public preview for mission-critical workloads. The service gives eligible API requests priority queueing, custom rate limits, and an uptime service-level agreement. The tier costs 1.75 times standard list pricing and carries a 99.5% uptime SLA.
Regional processing, however, does not mean every part of an AI application stays within the selected geography. Mistral says limited, safeguarded transfers to subprocessors outside the region can occur. That distinction could matter for regulated customers, particularly as EMEA companies weigh AI adoption against data sovereignty risks.
Mistral opens its platform to rival models
Mistral is also expanding beyond its own models. The company said its infrastructure will support third-party open models, beginning with Z.ai's GLM-5.2.
That gives customers access to different models without necessarily moving workloads between separate infrastructure providers. Mistral said open weights allow customers to inspect, adapt and retain the intelligence they build with models.
“Different workloads need different models, and that will keep changing as the frontier evolves,” Matan Grinberg, CEO and co-founder of Factory, said in a statement. “Mistral allows us to run open models under strict regional controls and service commitments, making it easy for us to maintain data residency and compliance requirements while furthering our commitment to open source.”
A long-term bet on European compute
The larger move is Mistral's plan to bring European companies together to secure long-term AI capacity. Its European Compute Units turn multi-year customer commitments into access to Mistral-built infrastructure. The company is targeting up to 1 GW of capacity by 2030.
Mistral's infrastructure ambitions are substantial. VentureBeat reported that Mistral currently operates less than 200 megawatts of capacity and is targeting 200 megawatts by the end of 2027, followed by up to 1 GW by 2030.
By combining open models with regional processing and long-term compute commitments, Mistral is positioning itself as a European alternative to predominantly U.S.-based AI and cloud providers.
Read more: The European Union is also advancing new cloud and AI rules designed to reduce reliance on U.S. technology providers, providing broader policy context for Mistral’s sovereignty push.