Mistral AI
Language models · Model developers
How they use Nebius AI Cloud
Nebius named Mistral AI among its managed cloud clients in its August 2024 financial results. Its investor materials described Mistral AI using GPU infrastructure for model training. The company appeared again on a Nebius client slide in 2025.
In a Nebius brochure, Mistral AI CTO Timothee Lacroix said Nebius compute and network infrastructure ranked among the strongest options the team had tried for stability and allowed it to train models efficiently. His testimonial states: “and allowed us to train our models efficiently.”
Sources
4 sourcesSEC-filed Oct 2024 deck
18 Oct 2024 · US SEC · Web page
Describes Mistral AI's use of Nebius infrastructure.
Open original sourceNebius brochure: Mistral AI customer testimonial
19 Oct 2026 · Nebius · Web page
In Nebius’s brochure, Mistral AI CTO Timothee Lacroix reports that Nebius compute and network infrastructure performed strongly on stability and enabled efficient model training.
Open original sourceNebius Group_press release_01-08-2024.pdf
15 Oct 2026 · assets.nebius.com · PDF
Nebius Group’s August 1, 2024 release named Mistral AI among Nebius AI’s managed clients.
Open original sourceCompany context
mistral.aiMistral AI develops language and multimodal AI models, software and computing infrastructure for developers, businesses and governments. Its products include Vibe, an assistant for research, analysis and coding; Studio, for building and managing AI applications and agents; and an inference API offering Mistral and third-party models. Forge helps organizations customize and train models, while Compute provides dedicated GPUs in Europe. Customers use these tools to automate work, process documents and build tailored AI systems with control over their data and deployment.
Mistral AI — ProductsDocumented financing · private company
Recent updates
21 updatesMistral pledges AI tools and research support to European RAISE initiative
Mistral joined the European Commission’s RAISE initiative, pledging open-weight models, Vibe seats and API credits for European researchers. It also committed to PhD internships, research roles and university collaborations.
European CommissionCERN and Mistral announce physics AI partnership
CERN and Mistral signed a statement of intent for a long-term partnership developing foundation models for scientific discovery. Their initial ambition is a Large Physics Model trained on simulated particle collisions.
CERNMistral launches Timestral forecasting models
Mistral released four time-series foundation models ranging from 5 million to 209 million parameters. They support probabilistic multivariate forecasting, run on CPUs, and can be adapted through Mistral Forge.
Mistral