← All companies

Frontier / Company profile

Mistral AI

What Mistral AI does

Mistral AI is a France-headquartered AI lab and applied AI company focused on building “open” and frontier commercial language and multimodal models, then packaging them into developer tooling and enterprise products for production use. Its product strategy spans (1) open-weight and API-accessible foundation models, (2) an AI application platform for building and operating assistants/agents, and (3) infrastructure intended to help enterprises deploy and control inference and customization closer to where their data and workloads run. The company’s latest direction emphasizes both productized agent workflows (Studio/Vibe) and “sovereign AI” infrastructure (regional endpoints, priority tiers, and longer-term compute commitments), alongside partnerships that expand access to model training and distribution capabilities.

News

Company record

Aug 11, 2026 · Official · Mistral AIIn-region inference, open models, and new European infrastructure for sovereign AI.

Mistral published an update describing in-region inference, expanded open model access, and plans for longer-term compute commitments in Europe, including a capacity target stated for 2030.

May 28, 2026 · Official · MistralAI Now Summit 2026

Mistral announced an AI stack for industrial engineering and described partnerships with Airbus, BMW, and ASML in its summary for the event.

May 23, 2026 · Official · Mistral AIEmmi joins Mistral to accelerate the AI-native industry

Mistral announced its acquisition of Emmi AI, describing the move as strengthening its position for industrial engineering and “physics AI” research.

May 22, 2026 · Official · Mistral AIConnect the dots: Build with built-in and custom MCPs in Studio

Mistral announced built-in and custom MCP connectors in Studio for integrating enterprise tools/data, with governance and monitoring inside Studio.

Apr 27, 2026 · Official · Mistral AIWorkflows for work that runs the business (public preview)

Mistral released Workflows in public preview, describing it as a way to run work processes that persist state and can resume after events.

Mar 23, 2026 · Official · Mistral AISpeaking of Voxtral (Voxtral TTS)

Mistral released Voxtral TTS, describing it as its first text-to-speech model with multilingual voice generation and low latency, available to test in Mistral Studio.

Mar 17, 2026 · Official · Mistral AIIntroducing Forge

Mistral introduced Forge, positioning it as a system for enterprises to build frontier-grade custom models grounded in their proprietary knowledge.

Mar 16, 2026 · Official · Mistral AIMistral AI partners with NVIDIA to accelerate open frontier models

Mistral announced its participation as a founding member in the NVIDIA Nemotron Coalition and said it would co-develop frontier open-source models with NVIDIA.

Show 3 earlier updates
Source map · 4 recurring channels · 13 references

Funding

Latest disclosed valuation$11.7B
Tracked capital$3.7B6 sourced rounds
DateRoundRaisedValuationLead / investorsEvidence
Mar 30, 2026Debt financing (data center / project debt)$830M
Feb 27, 2024Strategic investment (Series A add-on / conversion into next equity)$16.3M$1

What it builds

Vibe

Vibe is Mistral’s AI agent product for long-horizon work; it is positioned as a unified assistant experience for individuals and teams, including coding/work modes.

source ↗
Vibe Code

Vibe for code is Mistral’s coding-focused agent experience for terminal/IDE workflows.

source ↗
Studio

Studio is Mistral’s developer and enterprise platform to build, test, and deploy AI agents and applications, including programmatic access to Mistral models.

source ↗
Forge

Forge is Mistral’s system for enterprises to build frontier-grade custom models grounded in proprietary knowledge, with end-to-end data, evaluation, and deployment-oriented capabilities.

source ↗
Compute

Compute is Mistral’s infrastructure offering for training and inference, aimed at enterprise deployment options and capacity control.

source ↗
Studio Connect (built-in and custom MCP connectors)

Studio Connect extends Studio with connectors built from Model Context Protocol (MCP) servers, enabling enterprise data/tool integration governed and monitored inside Studio.

source ↗

Milestones & partnerships

Aug 11, 2026
Regional inference and compute commitments

Mistral stated it was expanding in-region inference and open model access, and described a compute-capacity target by 2030.

May 22, 2026
Studio Connect connectors via MCP

Mistral announced built-in and custom MCP connectors in Studio for integrating enterprise tools and data governed inside Studio.

Mar 23, 2026
Voxtral TTS launched

Mistral released Voxtral TTS for multilingual voice generation and described it as available to test in Studio.

Mar 17, 2026
Forge launched

Mistral introduced Forge as its enterprise model customization system centered on proprietary knowledge grounding and evaluation.

Sep 09, 2025
Series C announced and closed

Mistral’s Series C announcement set a 11.7B€ post-money valuation and an ASML-led round with listed participants.

ASML
Feb 06, 2025
Le Chat re-positioned as a unified assistant

Mistral’s about page notes “Le Chat (new)” on February 6, 2025.

May 29, 2024
Mistral AI Non-Production License (MNPL) introduced

Mistral introduced MNPL, describing it as a licensing approach for non-commercial use to support sustainable business growth.

Feb 26, 2024
Mistral Large release and Azure distribution

Mistral announced Mistral Large and said it would be available via Microsoft Azure (Azure AI Studio and Azure ML) alongside its own platform access.

Microsoft (Azure)
Sep 27, 2023
Mistral 7B release

Mistral’s about page lists “7B.” as a key date entry on September 27, 2023.

Jun 13, 2023
Seed round closed

Mistral’s about page lists a seed round in June 2023; external coverage attributes it to a €105M seed led by Lightspeed.

Apr 01, 2023
Mistral “born” (company formation)

Mistral’s about page states that “In April 2023, Mistral was born” following co-founder lab roots.