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).
Latest update Oct 19, 2026
01
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), HSBC, 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, alongside a €1.2B second site under construction with EcoDataCenter in Borlänge, Sweden - Mistral's first infrastructure investment outside France, targeted to open in 2027. Annual recurring revenue passed $400M in early 2026, up roughly 20x year over year; CEO Arthur Mensch told CNBC on September 8, 2026 he expects to beat the $1 billion ARR target if current trends hold. On September 8, 2026 Mistral closed a €3 billion Series D led by Samsung Electronics, with the EU-backed Scaleup Europe Fund (managed by EQT) and existing investor PSG Equity as co-leads, at a post-money valuation of more than €21 billion (~$24 billion). The next day Samsung announced a formal on-premises partnership to deploy Mistral Large across its semiconductor design, manufacturing and engineering operations. The company says it now operates in 20 countries and supports more than 125 global enterprises. Enterprise reach also includes UK grocer Tesco (three-year deal and joint AI lab, December 2025) and a Southeast Asia push anchored in Singapore, where headcount is set to triple to 100 by end-2026. On 6 October 2026 Mistral opened a public preview of Mistral Large 4, also called Le Chonk. The company blog describes a 1-trillion-parameter model with 52 billion active parameters and says open weights will follow by the end of October. As of 16 October 2026 those weights were not published. Artificial Analysis scores the preview at 38 on its Intelligence Index and 50 on its Cyber Index, and describes the model as 1 trillion parameters with 49 billion active.
mistral.aiUpdated Oct 16, 2026
Analysis · 5 min read · Sep 3, 2026The long view on Mistral AI: strategy, technology and what to watch02
2026-2027
Use the September 8, 2026 €3 billion Series D to expand owned data-center capacity and train larger models. CEO Arthur Mensch told CNBC the long-term plan is to rely on capacity Mistral builds itself, with owned compute growing around 100% over the next five years, rather than remaining dependent on rented infrastructure. As of September 16, 2026 no new named site or GPU-count disclosure has followed the close.
2026-2027
On September 9, 2026 Samsung announced a formal strategic partnership, at a France-South Korea state summit in Paris, to integrate Mistral Large and related on-premises models across semiconductor design, manufacturing-process optimization and engineering workflows. That converts CEO Arthur Mensch's September 8 comment that Mistral would work with Samsung the way it already does with ASML into a named industrial deployment. Watch whether it produces named production-line deployments and whether it extends beyond Samsung's own fabs.
2026-2027
Follow through on the June 2026 agreement with Bpifrance and MGX to scale Campus AI beyond the 1.4 GW Fouju flagship to a roughly 3 GW network of AI factories across France - pick and stand up the second site (location still not fixed as of September 18, 2026) alongside an additional ~€7.5B of investment.
Later 2026
On September 16, 2026 Mozilla and Mistral put Mistral Small 4 inside Firefox Smart Window beta for users in the United States, Canada and France, Mistral's first mass consumer distribution channel outside enterprise and sovereign deals. Mistral said the United Kingdom and Germany are expected to follow later this year, with zero data retention and no default server-side chat storage. Watch whether those two markets actually open in 2026, whether Smart Window stays a recommended Mistral option rather than a lock-in, and whether browser distribution moves Mistral's consumer footprint off the enterprise-only path.
2026-2027
On September 10, 2026 Mistral and Cloudera announced a partnership to run and customize Mistral models, including Forge, on Cloudera's hybrid data platform across public cloud, on-prem, sovereign and air-gapped environments covering about 30 exabytes of customer-managed data. No financial terms or named first customers were disclosed. Watch whether the deal produces signed deployments in regulated industries that will not send data to a US-hosted API.
2026-2029
Execute the September 15, 2026 three-year program with TotalEnergies, more than €100 million, to train frontier and agentic AI models on TotalEnergies subsurface data for reservoir exploration and engineering. The companies said they will establish a joint scientific laboratory; no first-model delivery date was disclosed.
2026 and beyond
Follow through on the strategic collaboration announced August 24, 2026 with Saudi AI company HUMAIN, valued in the hundreds of millions of euros, spanning AI infrastructure, the development and localization of advanced AI models (starting with cybersecurity and voice), Arabic-language frontier models, and a joint go-to-market strategy for regulated industries across Saudi Arabia and the wider region. Mistral is also exploring using HUMAIN's data-center infrastructure to support growing local compute needs - its first major Gulf partnership. As of September 16, 2026 no named joint model, customer, or compute capacity has been disclosed. HUMAIN separately launched humain-m3, a MiniMax-based Arabic frontier model, in research preview on September 3, 2026, so Mistral's promised Arabic models remain outstanding.
Q3 2026
Bring the dedicated 10 MW inference facility at Les Ulis (Essonne), inside Digital Realty's Paris South campus and powered by Nvidia GB300 NVL72 accelerators, online to complement the 44 MW Bruyères-le-Châtel training site and give Mistral direct control over serving capacity as usage scales. Re-verified as of September 18, 2026: still targeting Q3 2026 (the window closes September 30), with no independent confirmation yet that the site has opened. Mistral's own AI Now Summit page still lists a Q3 2026 opening.
Q3 2026
H2 2026
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. CEO Arthur Mensch told CNBC on September 8, 2026 that models due 'very soon' will be 'very competitive,' but as of September 16, 2026 no parameter count, benchmarks, license terms, or public release date has been disclosed.
H2 2026
End of 2026
Sustain the ~20x revenue ramp - from ~$20M (early 2025) past $400M (early 2026) - to management's target of more than $1B in annual recurring revenue by the end of 2026. CEO Arthur Mensch told CNBC on September 8, 2026 he expects to beat that target if current trends hold, and declined to give a revised figure. No newer ARR number has been disclosed since early 2026 as of September 16, 2026.
2027
Bring the €1.2B EcoDataCenter-partnered facility in Borlänge online - Mistral's first infrastructure investment outside France - with a 250 MW first phase (Nvidia Vera Rubin GPUs) on a site that can scale to 600 MW, expected to create roughly 200 long-term jobs plus 400 construction jobs over a 7-10 year buildout. Re-verified as of September 18, 2026: Mistral Compute's product page still lists the Sweden site as 'in motion' from February 2026, with no new opening date beyond 2027.
End of 2027
Convert the European enterprise coalition announced August 11, 2026 - anchor members ASML, Amadeus, CMA CGM, Caisse des Dépôts and Capgemini committing to multi-year 'European Compute Unit' purchases - into signed multi-year compute commitments. Mistral says the coalition's commitments will underwrite 200 MW of infrastructure by the end of 2027, on the way to a full gigawatt by 2030, alongside the separate regional-inference and Priority Tier offerings launched the same day. Re-verified as of September 18, 2026: Mistral Compute still advertises 1 GW of sovereign EU capacity by 2030, with no new named coalition MW disclosure since August 11.
Date not announced
TargetFollow-on contract not yet signed
03
On Oct. 15, 2026 Mistral released Timestral, time-series models in 5 million, 28 million, 123 million and 209 million parameter sizes. The company says the 209 million model ranks first among pretrained zero-shot models on GIFT-Eval, TIME and fev-bench, and that forecasts run on a CPU. A technical report is still to come, and no independent leaderboard check was opened.
Mistral launched a public preview of Mistral Large 4 (Le Chonk) on 6 October 2026, describing a 1-trillion-parameter natively multimodal mixture-of-experts model with 49 billion active parameters, trained from scratch on 3,800 NVIDIA Grace Blackwell GPUs in its European datacenters. The company said the preview API is live on Mistral Studio and that open weights will follow by the end of October after red-teaming with cybersecurity partners and state authorities; Axios, citing Mistral, named 27 October and quoted VP of science Pierre Stock saying the lab is not yet at the closed-model frontier. Independent evaluation of the preview API was not completed this run; SOTA-among-open-weights claims remain company-reported.
On September 28, 2026 Mistral said it opened a Munich hub for Physics AI and Industrial AI research plus applied engineers working with enterprise partners, and that it is recruiting there. The company said Munich teams will work with BMW on crash simulations and with Siemens Energy on industrial AI, and that it has a TUM wind-tunnel partnership for automotive aerodynamic digital twins; Heise confirmed those named partners from the announcement and noted Mistral did not disclose headcount or investment. BMW's crash-simulation partnership was already public in May 2026, Siemens Energy and TUM have not independently confirmed the new Munich work, and Mistral restated its existing 1 GW European compute target for 2030.
The European Space Agency said on September 23, 2026 that it and Mistral signed a letter of intent in Paris on September 16 to deepen AI cooperation. ESA Director General Josef Aschbacher and Mistral CEO Arthur Mensch signed the framework, which covers exploring Mistral models and infrastructure for Earth observation, engineering and space AI on European systems. The LoI builds on existing ESA-Mistral work including the Orbit assistant and EVE Earth Virtual Expert. It is not a production contract.
Mozilla and Mistral announced on September 16 that Mistral Small 4 is joining the beta of Firefox Smart Window, Mozilla's AI browsing assistant, as a selectable model for users in the United States and Canada, while France gets wider beta access and official French-language support. The United Kingdom and Germany are expected later this year, and Mozilla stressed that Firefox users can still choose among other models rather than being handed a default. Mozilla said it picked the model partly on multilingual performance, and Mistral agreed to zero data retention on the prompts and browsing context Firefox forwards to it.
TotalEnergies and Mistral announced on September 15 a three-year joint program of more than 100 million euros to build frontier AI models, including agentic systems, for oil and gas reservoir exploration, characterization, and development. The companies will set up a joint scientific laboratory pairing TotalEnergies subsurface teams with Mistral, drawing on nearly 10 petaflops of data and about a century of geoscience know-how. Chairman Patrick Pouyanne said the work is meant to generate multiple development scenarios and to extend the life of existing projects.
Mistral AI said September 10, 2026 it will run inference on Cloudera's hybrid platform so customers can deploy models in private or public cloud, on premises, or in air-gapped environments and customize them on proprietary data. Cloudera's Kamal Brar said the partnership targets customers sitting on about 30 exabytes of Cloudera-managed data. No deal value was disclosed.
Mistral said on September 8, 2026 that it raised 3 billion euros in a Series D led by Samsung Electronics, with the EU-backed Scaleup Europe Fund managed by EQT and existing investor PSG Equity as co-leads, at a post-money valuation of more than 21 billion euros (about $24 billion). The company called it the largest equity fundraising by a privately owned European technology firm, nearly doubling the 11.7 billion euro valuation from its ASML-led Series C a year earlier. CEO Arthur Mensch told CNBC the capital will go toward owned data centers and larger models; CFO Johan Bergqvist said Mistral is on track for $1 billion of annual recurring revenue by year-end and that Microsoft did not join this round.
Mistral AI announced a strategic collaboration on August 24 with Saudi AI company HUMAIN covering AI infrastructure, model development and deployment across Saudi Arabia and the wider region. The deal includes plans for Arabic-language frontier models and joint go-to-market efforts in regulated industries, with Mistral also exploring HUMAIN's data centers for regional compute capacity.
Mistral AI announced on August 11 a three-part expansion of its infrastructure business: regional inference endpoints letting customers choose whether workloads run in Europe or the US, a new uptime-guaranteed Priority Tier for mission-critical deployments, and a coalition of European enterprises making multi-year compute commitments that Mistral says will underwrite 200 megawatts of infrastructure by the end of 2027 and a full gigawatt by 2030. The company also said it will begin hosting third-party open models on its platform, starting with Z.ai's GLM-5.2, as it positions itself as the sovereign alternative to US and Chinese AI providers for European customers.
Mistral AI released Shieldstral on August 4, a 3-billion-parameter open-weight multimodal safety classifier the company says outperforms models up to seven times its size by treating content moderation as a policy-adaptive question-answering task. Released under an Apache 2.0 license, Shieldstral accepts plain-language moderation policies at inference time rather than requiring retraining, unifies text and image safety scoring, and is designed to run on a single 16GB GPU.
Microsoft and Mistral AI announced a major expansion of their strategic partnership, underpinned by a new multibillion-dollar agreement to build out AI infrastructure in Europe, including thousands of Nvidia Vera Rubin GPUs in European data centers. Mistral's models - including open-weight Mistral Medium 3.5 and its OCR 4 document model - will be available across Microsoft Foundry, Copilot Studio and Azure, with cloud, customer-controlled and fully disconnected deployment options aimed at regulated sectors such as financial services, healthcare and manufacturing.
04
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'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.
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, a €1.2B second site under construction with EcoDataCenter in Borlänge, Sweden (250-600 MW, Nvidia Vera Rubin GPUs, opening 2027) - Mistral's first infrastructure investment outside France - and the 'Campus AI' joint venture with MGX, Bpifrance and Nvidia - a 1.4 GW flagship site at Fouju now under construction, scaled in June 2026 to a roughly 3 GW nationwide network with a second site pending.
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, prompt/skills versioning and governance (July 2026), and Agentic Search (August 2026). Forge, launched March 2026, lets enterprises build frontier-grade custom models on Mistral's stack.
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, Robostral Navigate (July 2026) - an 8B embodied-navigation model steering robots from a single RGB camera - and Shieldstral (August 2026), a 3B open-weight safety classifier that judges text and images against plain-language moderation policies at inference time instead of a fixed, retrained taxonomy.
Public-preview natively multimodal mixture-of-experts model. Mistral's 6 Oct 2026 blog calls it a 1-trillion-parameter model with 52 billion active parameters and says the preview API is on Mistral Studio, with open weights due by the end of October. The docs page lists API identifier mistral-large-4+1, 52B active parameters, 1.05T total parameters, a 1.6B vision encoder, and a 1M context window. The Hugging Face card, still an upcoming release on 16 Oct 2026, says 49 billion active parameters per token and 52 billion including embeddings and output layers, with a current ETA of 31 Oct 2026. Artificial Analysis describes the same preview as 1T total and 49B active, with a 512k context window, and scores it 38 on the Intelligence Index and 50 on the Cyber Index. Weights were not downloadable on 16 Oct 2026, so there is no open-weight release date or public license yet.
05
mid-2026 tracker estimates
(up from ~$20M in early 2025). CEO Arthur Mensch told CNBC on September 8, 2026 he expects to beat the $1 billion year-end ARR target if current trends hold, and declined to give a revised figure; ~60% of revenue comes from European customers across API usage, sovereign enterprise deployments and Vibe subscriptions. No newer ARR figure has been publicly disclosed since early 2026.
(~$24B) post-money after the September 8, 2026 Series D, up from €11.7B (~$13.8B) at the September 2025 Series C
(early 2026), ~20x growth year over year; CEO Arthur Mensch told CNBC on September 8, 2026 he expects to beat the $1 billion ARR target if current trends hold, and declined to give a revised figure
(~€2.8B through Series C plus €3B Series D) plus ~€723M ($830M) data-center debt
mid-2026 tracker estimates; Mistral has not published a later official headcount as of Sept 18, 2026
44 MW / 13,800 Nvidia GB300 GPUs at Bruyères-le-Châtel; a second, larger site (250-600 MW, Nvidia Vera Rubin GPUs) is under construction in Borlänge, Sweden via a €1.2B EcoDataCenter partnership announced Feb 2026, still targeted to open in 2027 as of Sept 18, 2026
1M downloads in first 14 days of the mobile app (Feb 2025); accounts, history and plans carried over automatically at the May 2026 Vibe rebrand; Vibe 2.0 (terminal-native coding agent powered by Devstral 2) shipped January 27, 2026. On Sept 16, 2026 Mistral Small 4 entered Firefox Smart Window beta in the US, Canada and France, Mistral's first large consumer distribution channel
TotalEnergies three-year program of more than €100M (announced Sept 15, 2026) alongside CMA CGM's €100M over five years; a July 2026 multibillion-dollar Microsoft deal to rent Mistral's European data-center capacity is larger but its value remains undisclosed as of September 2026
Framework only as of September 23, 2026: explore models, cloud and skills for Earth observation, engineering and space AI on European infrastructure; not a production contract
street address, hub headcount and investment undisclosed
Q3 2026 closed without a Mistral or Digital Realty confirmation that inference is live. Mistral's AI Now Summit page still lists a Q3 2026 opening as of 5 Oct 2026; newsroom has no opening post.
900-plus on mistral.ai/careers as of 30 Sep 2026 (30-plus nationalities; 50% female leaders)
Public preview since 6 Oct 2026, API mistral-large-4+1. Weights not published as of 16 Oct 2026. Blog: 1T / 52B active. Docs: 1.05T / 52B active / 1.6B vision encoder / 1M context. Hugging Face: 49B active per token (52B with embeddings), ETA 31 Oct 2026. Artificial Analysis: Intelligence Index 38, Cyber Index 50, 49B active, 512k context.
06
Operating status
Hypergrowth; Timestral released; Large 4 in preview
On Oct. 15, 2026 Mistral released Timestral, time-series models from 5 million to 209 million parameters. The company says the largest ranks first among pretrained zero-shot models on GIFT-Eval, TIME and fev-bench; that ranking is not independently confirmed. Mistral Large 4 remains a public preview, with open weights still due by the end of October.
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 over five years), TotalEnergies (more than €100M over three years, announced September 15, 2026), Stellantis in-car assistants, Airbus and BMW industrial engineering, UK grocer Tesco (three-year deal and joint AI lab, signed December 2025), HSBC, French utility EDF, 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. Mistral's September 8, 2026 Series D post said it supports more than 125 global enterprises across 20 countries.
GPU cloud for training and inference at scale on Mistral's own Nvidia-powered European data centers, sold to enterprises and public-sector customers that need sovereign AI infrastructure. Mistral Compute's product page (re-verified September 18, 2026) still lists 1 GW of sovereign EU capacity by 2030 and the Sweden EcoDataCenter site as in motion from February 2026.
07
How past announced dates turned out.
2/2announced dates hit on time
Close the reported ~€3B round at a ~€20B valuation
Target H2 2026 · Resolved Sep 8, 2026
Mistral announced a €3 billion Series D on September 8, 2026 led by Samsung Electronics, with the Scaleup Europe Fund and PSG Equity as co-leads, at a post-money valuation of more than €21 billion. The close exceeded the ~€20B valuation that had been in talks since June.
Start construction of the 1.4 GW Campus AI
Target H2 2026 (operations 2028) · Resolved Jun 2, 2026
Campus AI entered its implementation/construction phase on schedule: on June 2, 2026 Eiffage was awarded a >€120M contract to design, build and commission the Fouju site's 400/33kV high-voltage substation and shared infrastructure, with a groundbreaking ceremony and site works following in summer 2026 (within the announced H2 2026 construction-start window) and first-phase delivery targeted for spring 2028.
08
2026
Releases Timestral time-series models
On Oct. 15, 2026 Mistral released Timestral in four sizes, from 5 million to 209 million parameters, for probabilistic forecasts with a 32,768-step context. The company says the 209 million model ranks first among pretrained zero-shot models on GIFT-Eval, TIME and fev-bench, and that a technical report will follow.
Launches Mistral Large 4 preview, weights due end of October
On 6 October 2026 Mistral announced a public preview of Mistral Large 4 (Le Chonk), a 1-trillion-parameter natively multimodal mixture-of-experts model with 49 billion active parameters, trained on 3,800 NVIDIA Grace Blackwell GPUs in its European datacenters. The company said the preview API is available on Mistral Studio and that open weights will drop by the end of October after red-teaming with cybersecurity partners and state authorities. Axios, citing Mistral, named 27 October for weights and quoted Pierre Stock saying the lab is not yet at the closed-model frontier. Docs list Large 4 as an open-weight general-purpose multimodal model.
Opens Munich Physics AI and Industrial AI hub
On September 28, 2026 Mistral announced a Munich hub for Physics AI and Industrial AI research teams and applied engineers. It named BMW crash-simulation work, Siemens Energy industrial AI, and a TUM wind-tunnel digital-twin partnership, and said it is recruiting in the region. Heise reported the opening and that Mistral did not disclose staff numbers or investment. BMW's crash-simulation collaboration was already announced in May 2026. Siemens Energy and TUM were not independently confirmed. Headcount at the hub is undisclosed.
Signs ESA letter of intent on space AI
ESA said on September 23, 2026 that Director General Josef Aschbacher and Mistral CEO Arthur Mensch signed a letter of intent in Paris on September 16 to explore Mistral models and infrastructure for Earth observation, engineering and space AI on European systems, building on existing Orbit and EVE work. The LoI is a framework, not a production contract.
Mistral Small 4 comes to Firefox Smart Window
On September 16, 2026 Mozilla and Mistral announced a partnership putting Mistral Small 4 into the beta of Firefox Smart Window, Mozilla's AI browsing assistant, as a selectable model for users in the United States and Canada, while France gains wider beta access and official French-language support. The United Kingdom and Germany are expected later in 2026. Mozilla said it chose the model partly on multilingual performance and that users can still pick other models or a custom endpoint; Mistral agreed to zero data retention on the prompts, memories and browsing context Firefox forwards. It is Mistral's first large consumer distribution channel, a departure from its enterprise and sovereign-government focus.
Partners with TotalEnergies on frontier geoscience AI
On September 15, 2026 TotalEnergies and Mistral announced a three-year joint program of more than €100 million to develop frontier AI models, including agentic systems, for oil and gas reservoir exploration, characterization, and development. The companies will establish a joint scientific laboratory pairing TotalEnergies subsurface teams with Mistral, drawing on nearly 10 petaflops of data and about a century of geoscience know-how. No equity investment was disclosed.
Partners with Cloudera on sovereign enterprise AI
On September 10, 2026 Mistral said it will run inference on Cloudera's hybrid data platform so customers can deploy models in private or public cloud, on premises, or in air-gapped environments and customize them on proprietary data. Cloudera's Kamal Brar said the partnership is aimed at customers sitting on about 30 exabytes of Cloudera-managed data. No deal value was disclosed.
Samsung lead ties Mistral to chip-manufacturing AI
Samsung Electronics led Mistral's September 8, 2026 Series D. The next day, at a France-South Korea state summit in Paris, Samsung announced a formal strategic partnership to integrate Mistral Large and related on-premises models across semiconductor design, manufacturing-process optimization and engineering workflows, extending Mistral's industrial-sovereign pitch from European chip equipment (ASML) into Korean semiconductor manufacturing.
2026
Closes €3B Series D led by Samsung at more than €21B
On September 8, 2026 Mistral announced a €3 billion Series D led by Samsung Electronics, with the EU-backed Scaleup Europe Fund managed by EQT and existing investor PSG Equity as co-leads, at a post-money valuation of more than €21 billion (about $24 billion). The company called it the largest equity fundraising by a privately owned European technology firm. CEO Arthur Mensch told CNBC the capital will fund owned data centers and larger models; CFO Johan Bergqvist said Mistral is on track for $1 billion of annual recurring revenue by year-end and that Microsoft did not participate.
Strategic collaboration with Saudi Arabia's HUMAIN
On August 24, 2026 Mistral announces a strategic collaboration with Saudi AI company HUMAIN covering AI infrastructure, model development and deployment across Saudi Arabia and the wider Middle East. The deal includes plans for Arabic-language frontier models and joint go-to-market efforts in regulated industries, with Mistral also exploring HUMAIN's data centers for regional compute capacity.
Unveils sovereign AI infrastructure coalition
On August 11, 2026 Mistral announces a three-part infrastructure expansion: regional inference endpoints letting customers choose whether workloads run in Europe or the US, an uptime-guaranteed 'Priority Tier' for mission-critical deployments, and a coalition of European enterprises making multi-year compute commitments that Mistral says will underwrite 200 MW of infrastructure by end-2027 and a full gigawatt by 2030. Mistral also says it will begin hosting third-party open models on its platform, starting with Z.ai's GLM-5.2.
Fouju Campus AI enters construction phase
On June 2, 2026, Campus AI selects Eiffage - via a consortium of Eiffage Energie Systemes and Eiffage Genie Civil - to design, build and commission the Fouju site's shared infrastructure, including a 400/33kV, 1.4 GW high-voltage substation, in a contract worth over €120M, marking the project's move from planning into its implementation phase ahead of a first-phase delivery targeted for spring 2028.
Campus AI scales to a 3 GW nationwide network
On June 1, 2026, at the Choose France summit, Bpifrance, Mistral and Abu Dhabi-backed MGX agree to expand the Campus AI joint venture beyond the flagship 1.4 GW Fouju site to a roughly 3 GW network of 'AI factories' across France - an additional ~€7.5B of investment - with a second site to be selected; as a shareholder, Mistral secures priority access to all compute deployed across the program rather than renting capacity on the open market.
Microsoft partnership expands with multibillion-dollar Europe AI deal
On July 21, 2026 Microsoft and Mistral announced a major expansion of their strategic partnership, underpinned by a new multibillion-dollar agreement to build out AI infrastructure in Europe - including thousands of Nvidia Vera Rubin GPUs in European data centers. Mistral's models, including open-weight Mistral Medium 3.5 and OCR 4, will be available across Microsoft Foundry, Copilot Studio and Azure, with cloud, customer-controlled and fully disconnected deployment options aimed at regulated sectors such as financial services, healthcare and manufacturing.
Physical-AI push: Robostral Navigate ships
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.
First sovereign data center online; €3B raise reported in talks
Mistral Compute's Bruyères-le-Châtel facility begins operations around mid-2026 on schedule, and in June Bloomberg reports Mistral is in early talks to raise about €3B at a roughly €20B valuation - nearly double its Series C mark.
Le Chat becomes Vibe; Airbus, BMW and ASML industrial deals
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.
Singapore headcount to triple as Southeast Asia hub
In May 2026 Mistral says it will roughly triple its Singapore headcount from about 40 to 100 by the end of 2026, hiring mostly local engineers and sales staff, after investing 'multiple dozens of millions of dollars' there since opening the office in January 2025. Singapore is Mistral's largest Asia-Pacific market, ahead of Australia and South Korea, and is positioned as its hub for sovereign-AI deployment and expansion across Southeast Asia's manufacturing and financial-services sectors.
$830M debt raise; deeper Nvidia partnership
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.
ARR passes $400M, up ~20x year over year
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.
First overseas data center: €1.2B Sweden investment with EcoDataCenter
On February 11, 2026 Mistral announces its first infrastructure investment outside France: a €1.2B (~$1.4B) long-term deal with EcoDataCenter to build an AI-focused data center on a 20-hectare former paper-mill site in Borlänge, Sweden, hosting Nvidia's next-generation Vera Rubin GPUs. The site offers 250 MW in its first phase with room to scale to 600 MW, is expected to open in 2027, and is projected to create about 200 long-term jobs plus 400 construction jobs over a 7-10 year buildout.
French armed forces framework agreement
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.
2025
Three-year Tesco partnership and joint AI lab
In December 2025 UK grocer Tesco signs a three-year generative-AI agreement with Mistral, giving Tesco access to Mistral's commercial model family and establishing a joint 'AI lab' where teams from both companies co-develop use cases spanning Clubcard personalization, faster data analysis, internal tooling, and operational efficiency.
Mistral 3 generation: frontier-class open weights
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.
€1.7B Series C led by ASML at €11.7B valuation
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.
Mistral Compute announced with Nvidia
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.
Le Chat mobile app hits 1M downloads in 14 days
Mistral ships Le Chat mobile apps in February 2025; boosted by public promotion from President Macron, the assistant reaches 1 million downloads in its first 14 days and tops the French iOS App Store.
2024
Mistral Large, Microsoft deal and €600M Series B
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.
2023
In September 2023 Mistral releases Mistral 7B, a free, openly licensed model that outperforms larger rivals and establishes its open-weight strategy. In December it raises a €385M (~$415M) Series A led by Andreessen Horowitz at a ~$2B valuation.
Founded in Paris; record €105M seed
Arthur Mensch (Google DeepMind) and Meta researchers Guillaume Lample and Timothée Lacroix found Mistral AI in April 2023. Weeks later it closes a €105M (~$113M) seed round led by Lightspeed - then Europe's largest-ever seed - at a ~$260M valuation.
09
$7.5Bdisclosed across 7 rounds with a stated amount
Cumulative disclosed raise · dated rounds
Seed
Jun 2023
€105M (~$113M) at ~$260M valuation
Lightspeed (lead), Xavier Niel, JCDecaux Holding, Rodolphe Saadé, Motier Ventures, Exor, Sofina, First Minute Capital, LocalGlobe
Series A
Dec 2023
€385M (~$415M) at ~$2B valuation
Andreessen Horowitz (lead), Lightspeed, Salesforce, BNP Paribas, General Catalyst, Elad Gil
Microsoft convertible investment
Feb 2024
Microsoft (alongside Azure distribution partnership)
Series B
Jun 2024
€600M (~$640M, equity + debt) at ~$6B valuation
General Catalyst (lead), Lightspeed, Andreessen Horowitz, Nvidia, Samsung Venture, Salesforce, Bpifrance
Series C
Sep 2025
€1.7B (~$2B) at €11.7B (~$13.8B) valuation
ASML (lead, ~11% stake), Nvidia, DST Global, Andreessen Horowitz, Bpifrance, General Catalyst, Index Ventures, Lightspeed
Debt financing (data centers)
Mar 2026
Bpifrance, BNP Paribas, Crédit Agricole CIB, HSBC, La Banque Postale, MUFG, Natixis CIB (bank consortium)
Series D
Sep 2026
€3B (~$3.5B) at more than €21B (~$24B) valuation
Samsung Electronics (lead), Scaleup Europe Fund / EQT (co-lead), PSG Equity (co-lead); new investors Advent, funds and accounts managed by BlackRock, and the Grand Duchy of Luxembourg; existing investors including a16z, ASML, Nvidia, Salesforce Ventures, Bpifrance, General Catalyst, DST Global, Index Ventures and Lightspeed
