
Mistral AI
Arthur Mensch
Data verified as of Jul 17, 2026
Summary
Mistral AI is Europe's leading frontier AI lab, founded in April 2023 in Paris by CEO Arthur Mensch (ex-Google DeepMind) with chief scientist Guillaume Lample and CTO Timothée Lacroix (both ex-Meta). It champions open-weight models: its December 2025 Mistral 3 generation includes Mistral Large 3, a 675B-parameter sparse mixture-of-experts model released under Apache 2.0, alongside commercial and specialist models (Medium 3.5, Small 4, Magistral reasoning, Codestral/Devstral coding, Voxtral audio, OCR 4 documents and the new Robostral robotics line). Its consumer assistant Le Chat was rebranded 'Vibe' in May 2026 and repositioned as an agentic work-and-code platform. Positioned as Europe's sovereign alternative to US labs, Mistral has deep industrial and government ties — lead Series C investor ASML (~11% stake), Airbus, BMW, Stellantis, CMA CGM (a €100M deal) and a framework agreement with France's Ministry of Armed Forces — and is building its own Nvidia-powered infrastructure through Mistral Compute, whose first 44 MW data center south of Paris came online in mid-2026. Annual recurring revenue passed $400M in early 2026, up roughly 20x year over year, with a $1B+ target by year-end, and in June 2026 the company was reported to be in talks to raise about €3B at a ~€20B valuation.
Main Products

Mistral's frontier model generation, headlined by Mistral Large 3 — a 675B-parameter sparse mixture-of-experts model (41B active) released open-weight under Apache 2.0 in December 2025 — alongside Mistral Medium 3.5 (the 'first flagship merged model,' default in Vibe since May 2026) and the efficient Mistral Small 4. A next-generation 'fat but sparse' MoE frontier model entered partner early access in July 2026.
Mistral Large 3 and the Mistral 3 family are generally available via La Plateforme, major clouds and open-weight download; Medium 3.5 (April–May 2026) powers Vibe by default, and the next open-weight frontier model is in early access with research, government and industry partners as of July 2026.

Mistral's consumer and enterprise AI assistant, rebranded from Le Chat to Vibe on May 28, 2026 and repositioned as 'one agent for long-running, multi-step work.' Work Mode connects to Google Workspace, Outlook, SharePoint, Slack and GitHub to execute multi-step tasks autonomously; Code Mode runs persistent sandboxed coding sessions through to merged pull requests, with a VS Code extension and CLI.
Available on web and mobile in Free, Pro (€14.99/mo), Team (€24.99/user/mo) and Enterprise tiers; existing Le Chat conversations, settings and plans carried over at the rebrand. Remote agents run on Mistral Medium 3.5.

Sovereign European AI cloud built with Nvidia: GPU capacity for training and inference operated on Mistral's own data centers, anchored by the 44 MW Bruyères-le-Châtel site (13,800 Nvidia GB300 GPUs) in Essonne, a 10 MW inference facility at Les Ulis due in Q3 2026, and a planned 1.4 GW 'Campus AI' joint venture with MGX, Bpifrance and Nvidia targeting operations by 2028.
First site at Bruyères-le-Châtel came online around mid-2026 as scheduled, funded partly by the March 2026 $830M debt raise; Mistral targets ~200 MW of European capacity by 2027 ahead of the 1.4 GW campus, whose construction is slated to begin in H2 2026.
Mistral's developer and enterprise platform: pay-per-token API access to the full model catalog plus AI Studio for building production AI — Workflows (April 2026), built-in and custom MCP connectors, Search Toolkit, and prompt/skills versioning and governance (July 2026). Forge, launched March 2026, lets enterprises build frontier-grade custom models on Mistral's stack.
Actively expanding through 2026 with enterprise-grade governance (prompt/skill system of record, connector controls) and the Forge custom-model program.
A broad lineup of purpose-built models: Codestral and Devstral for coding, Magistral for step-by-step reasoning, Voxtral for speech (with an open-weight TTS model added March 2026), Mistral OCR 4 for document intelligence (June 2026), Leanstral for formal proofs, and Robostral Navigate (July 2026) — an 8B embodied-navigation model steering robots from a single RGB camera — marking Mistral's push into physical AI.
Most specialist models ship open-weight alongside commercial API versions; Robostral Navigate launched July 8, 2026 as the first model in Mistral's robotics line.
What's Next
Publicly release the next-gen open-weight frontier model
Graduate the new 'fat but sparse' mixture-of-experts frontier model — in early access with research, government and industry partners since July 2026 — to a broad open-weight release, Mistral's next attempt to close the capability gap with OpenAI, Anthropic and Google.
Start construction of the 1.4 GW Campus AI
Begin building the 1.4 GW Paris-region AI campus joint venture with MGX, Bpifrance and Nvidia — Europe's largest announced AI data-center project — with construction slated to start in the second half of 2026 and operations targeted for 2028, toward ~200 MW of Mistral capacity by 2027 along the way.
Operations & Revenue
Mistral is scaling fast on all fronts: ARR passed $400M in early 2026 (targeting $1B+ by year-end), its first sovereign 44 MW data center came online south of Paris in mid-2026 with a 10 MW inference site due in Q3 2026 and a 1.4 GW campus planned for 2028, and it is in reported talks for a ~€3B raise at a ~€20B valuation. Its strategy pairs frontier-class open-weight models (Apache 2.0 Mistral Large 3) with deep European enterprise, industrial and defense relationships.
Revenue Streams
Usage-based, per-token access to the Mistral model catalog (Large 3, Medium 3.5, Small 4, Magistral, Codestral, Voxtral, OCR 4) plus AI Studio tooling — Workflows, connectors/MCP, prompt and skills management — for developers and enterprises.
Vibe subscriptions (Pro / Team / Enterprise)
Paid tiers of the Vibe assistant (formerly Le Chat) — Free, Pro at €14.99/month, Team at €24.99/user/month and custom Enterprise — spanning agentic Work Mode (Google Workspace, Outlook, Slack, SharePoint) and Code Mode (GitHub, sandboxed sessions).
Sovereign & industrial deployments
Large bespoke deployments and 'AI factory' engagements — CMA CGM (€100M), Stellantis in-car assistants, Airbus and BMW industrial engineering, the French armed forces framework agreement, Helsing defense work and government programs like AI for Citizens — often run on-premises or on French-controlled infrastructure. Includes Forge, its system for enterprises building frontier-grade custom models.
Key Metrics
Timeline
In January 2026 France's Ministry of Armed Forces signs a framework agreement giving the armed forces, defense directorates and affiliated institutions (CEA, ONERA, SHOM) access to Mistral's models, run exclusively on French-controlled infrastructure — cementing Mistral's role in European defense AI alongside its Helsing partnership.
In early 2026 Mistral's annual recurring revenue climbs past $400M — up from roughly $20M a year earlier — driven by La Plateforme API usage, sovereign enterprise deployments and Vibe subscriptions, with ~60% of revenue from Europe and a $1B+ ARR target for year-end 2026.
In March 2026 Mistral raises ~€723M ($830M) in debt from a consortium of seven banks (Bpifrance, BNP Paribas, Crédit Agricole CIB, HSBC, La Banque Postale, MUFG, Natixis CIB) to fund its first own data center at Bruyères-le-Châtel — 13,800 Nvidia GB300 GPUs and 44 MW — and expands its Nvidia partnership to accelerate open frontier models.
At its AI Now Summit on May 28, 2026 Mistral rebrands Le Chat as 'Vibe' — an agentic assistant with autonomous Work and Code modes — and announces industrial AI partnerships with Airbus (design, manufacturing, on-board systems), BMW (a 'Large Industry Model' for engineering) and ASML, plus the Emmi physics-AI acquisition and a 10 MW inference data center at Les Ulis.
On July 8, 2026 Mistral releases Robostral Navigate, its first embodied-navigation model — an 8B simulation-trained model that steers robots using a single RGB camera and plain-language instructions — extending its specialist lineup (Leanstral proofs, OCR 4, Voxtral TTS) into robotics.
At VivaTech in June 2025 Mistral announces Mistral Compute, a sovereign European AI cloud built with Nvidia — anchored by data-center capacity in the Essonne region south of Paris — so European governments and enterprises can train and serve AI on European infrastructure.
In September 2025 Mistral closes a €1.7B (~$2B) Series C led by Dutch lithography giant ASML — which takes an ~11% stake and a strategic partnership — at a €11.7B (~$13.8B) post-money valuation, with Nvidia, DST Global, Andreessen Horowitz, Bpifrance, General Catalyst, Index Ventures and Lightspeed participating.
On December 2, 2025 Mistral releases the Mistral 3 family, headlined by Mistral Large 3 — a 675B-parameter sparse mixture-of-experts model (41B active) published under the permissive Apache 2.0 license — its strongest claim yet to frontier-class open-weight AI.
Mistral launches its commercial flagship Mistral Large and the Le Chat assistant in February 2024 alongside a distribution partnership (and €15M convertible investment) with Microsoft to serve models on Azure. In June it raises €600M (~$640M) in mixed equity and debt led by General Catalyst at a ~$6B valuation.
Funding
Cumulative disclosed raise · dated rounds