Mistral AI Annualized Revenue Run-Rate (ARR)
End 2024-Jan 2026
Annualized Run-Rate (USD M)
Updated at 2026-01
Mistral AI Funding Rounds and Valuation Trajectory
| Round (date, lead) | Amount Raised | Post-Money Valuation | Notable Investors |
|---|---|---|---|
| Seed (2023-06, Lightspeed) | EUR 105M ($113M) | $260M | Lightspeed, Eric Schmidt, Xavier Niel, JCDecaux, Sofina, Exor Ventures |
| Series A (2023-12, Andreessen Horowitz) | EUR 385M ($415M) | $2B | a16z, Lightspeed, NVIDIA, Salesforce, BNP Paribas, Bpifrance |
| Series B (2024-06, General Catalyst) | EUR 600M ($640M) | $6B | General Catalyst, Lightspeed, a16z, Bpifrance, Cisco, IBM, Salesforce |
| Series C (2025-09, ASML) | EUR 1.7B (~$2B) | EUR 11.7B (~$14B) | ASML (lead, EUR 1.3B, ~11%), NVIDIA, DST Global, General Catalyst, a16z, Index, Lightspeed, Bpifrance |
| Conventional debt (2026-03, bank syndicate) | $830M (debt) | n/a (debt, not priced equity) | Syndicate of seven banks; funds 13,800 Nvidia GB300 GPUs for a Paris-area data center |
Updated at 2026-03
Le Chat (Vibe) Consumer Traction and Customer Base
| Metric | Figure | As of | Note |
|---|---|---|---|
| Le Chat app downloads to 1M | 1,000,000 | Feb 2025 | Reached in ~13-14 days; topped French iOS App Store |
| Macron public endorsement | Yes | Feb 2025 | President urged 'download Le Chat rather than ChatGPT' before AI Action Summit |
| Total customers | ~450,000 | Mid-2025 | Across consumer, developer, and enterprise |
| High-value enterprise customers | ~1,031 | Mid-2025 | European regulated industry and French-state heavy |
| Revenue from Europe | ~60% | 2025-2026 | Structural geographic skew vs US peers |
| Consumer product rename | Le Chat to Vibe | May 2026 | Repositioned as AI chat plus coding agent |
Updated at 2026-05
Mistral Model Releases: Open-Weights vs Proprietary
| Date | Model | License | Significance |
|---|---|---|---|
| Sep 2023 | Mistral 7B | Open (Apache 2.0) | First release; beats LLaMA 2 13B; established the open-weights brand |
| Dec 2023 | Mixtral 8x7B | Open (Apache 2.0) | First open-weights SoTA mixture-of-experts model; beats GPT-3.5 on key benchmarks |
| Feb 2024 | Mistral Large + Microsoft/Azure deal | Proprietary (closed) | First closed frontier model; Azure-first launch sparked EU open-weights backlash |
| Jul 2024 | Mistral Large 2 + Codestral | Proprietary (research-license weights) | Frontier-tier flagship plus dedicated coding model; multi-cloud distribution |
| Nov 2024 | Pixtral + Ministral family | Mixed (open + proprietary) | Native multimodal plus small-model 'AI on the edge' family |
| May 2025 | Magistral (reasoning) | Mixed (open small + proprietary) | First European entry in the o1-class reasoning space |
| Dec 2025 | Mistral Large 3 (675B/41B MoE) + Ministral 3 | Open (Apache 2.0) | Sparse MoE flagship trained on 3,000 H200s; swing back to open weights |
Updated at 2025-12
Mistral European Sovereign Compute Buildout
| Program / Facility | Announced | Capacity / Commitment | Status |
|---|---|---|---|
| Mistral Compute (sovereign GPU cloud) | Oct 2025 | EU bare-metal GPU cloud on ASML + Nvidia reference architecture | GA |
| Bruyeres-le-Chatel data center (Essonne) | Mar 2026 | 13,800 Nvidia GB300 GPUs; 44 MW | Planned go-live Q2 2026 |
| $830M debt financing (seven banks) | Mar 2026 | Debt to fund the GB300 GPU purchase above | Closed |
| Paris-region AI campus JV (Bpifrance, MGX, Nvidia) | 2025-2026 | 1.4 GW campus | Construction expected H2 2026; operations from 2028 |
| European deployed-capacity target | Stated 2026 | ~200 MW across Europe | Targeted by end-2027 |
| Proprietary chip exploration | May 2026 | Custom silicon to reduce Nvidia dependence (RUMORED scope) | Exploratory per CEO Arthur Mensch |
Updated at 2026-03
Mistral AI in AI Software
OpenAI
$PRIVATE
Largest AI lab by revenue. Approximately $13B annualized run-rate revenue (mid-2025) from a mix of ChatGPT consumer subscriptions, ChatGPT Enterprise / Team, the OpenAI API, and Microsoft Azure OpenAI revenue share. ChatGPT reaches roughly 800M weekly active users (late 2025). Builder of the GPT family (GPT-4o, GPT-4.5, GPT-5, o-series reasoning models) and Sora video. Microsoft is the dominant compute and distribution partner; SoftBank led the $40B March 2025 primary round at $300B.
