Mistral AI acquires Emmi AI to build an industrial physics-AI stack

Why it matters

Founders on both sides are betting that physics-informed AI can move from demos to production. If Mistral can merge Emmi's domain models with its platform, European manufacturers get a faster path to real-time sims and digital twins.

Mistral AI is acquiring Emmi AI, bringing cofounder Johannes Brandstetter and a 30-plus research team into Mistral AI (@MistralAI) to build what both companies call a leading AI stack for industrial engineering in a company post.

Brandstetter, Emmi AI's cofounder and Chief Science Officer out of Linz, framed the move as a bet on AI4Science going from lab demos to production. "This is a pivotal moment for the future of industrial engineering... By integrating our expertise into Mistral's world-class AI ecosystem, we are positioned to revolutionize core R&D," he said in the announcement. On Mistral's side, cofounder and CEO Arthur Mensch (@arthurmensch) called the deal a way to become the partner of choice for manufacturers in high-stakes sectors and to deliver a fully integrated platform.

What Emmi built

Emmi AI emerged in Austria as a physics AI shop focused on industrial engineering: models that speed up and sometimes replace traditional simulations in workflows across energy, automotive, semiconductors, and aerospace. The team has worked on problems like grid stabilization, injection molding, and automotive safety testing, aiming for real-time simulations and richer digital twins. Mistral cofounder and CSO Guillaume Lample said the combined stack is meant to "deliver real-time simulations and sophisticated digital twins" that break through long-standing technical bottlenecks.

How it will integrate at Mistral

Per the announcement, Emmi's researchers and engineers will join Mistral's Science and Applied AI groups this month. Mistral will make Linz an official office alongside Paris, London, Amsterdam, Munich, San Francisco, and Singapore, and plans to keep hiring in Austria, Germany, and Lithuania where Emmi has been operating.

Mensch positioned the combined offering as a full-stack platform for industrial customers, pairing Mistral's foundation-model platform with Emmi's physics-informed systems for simulation-heavy R&D. The pitch: compress cycles in design, testing, and validation for components and systems where milliseconds or microns matter.

The founder path to this deal

Emmi AI spent the past two years carving out a niche in engineering AI from Linz with a tight research team and industry pilots. In April 2025, the company said it raised a EUR 15 million seed round from 3VC, Speedinvest, Serena, and PUSH, signaling investor appetite for physics-first AI in Europe. Brandstetter has been vocal about bringing AI4Science into real-world manufacturing constraints, and the acquisition puts that thesis inside a larger platform with distribution into enterprise accounts.

For Mensch and Mistral, this is a step into deep industrial verticals while keeping the company European at its center of gravity. Folding in Emmi's domain models and talent gives Mistral a clearer story for aerospace, automotive, and chip customers who need deterministic performance and faster iteration cycles.

What customers should watch

• Near-term, expect Mistral to productize Emmi's physics AI into tools aimed at simulation and digital twin workflows.
• A Linz hub and ongoing hiring suggest more Europe-based industrial R&D capacity.
• If the integration lands, the combined stack could reduce time-to-validation in areas like crash testing, thermal management, and power systems.

Deal terms were not disclosed in the announcement.