# Mistral Compute / AI Cloud

Official pricing: https://mistral.ai/products/aicloud/  
Category: gpu-cloud · Isolation: bare-metal

## Pricing regimes (raw)

- **Dedicated AI capacity** (resource) [sales, commit]

## Features

Yes: persistent disk, volumes, ≥24 h sessions, custom image (Docker or snapshot), your own Docker/OCI image, full VM (own kernel), Docker inside, nested virtualization, SSH, static egress IP, SOC 2, HIPAA, SSO, GPU, arm64, Python, audit logs, EU data residency, gVisor or VM (no shared kernel), firewall inside (nftables), shared volumes, FUSE mounts, kernel features (eBPF, mounts, cgroups)

No: snapshots, memory snapshots, fork/clone, pause/resume, idle auto-stop, start from your own snapshot, browser, desktop GUI, computer-use API, browser + desktop control, anti-bot stealth, CAPTCHA solving, residential IPs, preinstalled agents, public IPv4, HTTPS preview URLs, self-hosting, BYOC, open source, Windows, macOS, Node.js, HTTP method/path egress rules, wake on request, live resize, live fork (no pause), memory fork, MCP server, usage per sandbox (API), automatic snapshots, snapshots on demand, agent harness API, hosted agent API (their own agent), inbound access rules, secret proxy, secret proxy for any API, extra volumes

Unknown: everything else. Evidence (source + quote) per feature: https://battleships.dev/data/providers/mistral-compute.json → feature_evidence

## Company

- Ownership: independent
- Stage: series-d+ [source](https://mistral.ai/news/mistral-makes-sovereign-open-weight-ai-to-frontier/)
- Accelerators: none found
- Total raised: $6.8B (estimated: Sum of reported equity rounds: ~$117M seed (Jun 2023), ~$428M (Dec 2023), $16M Microsoft (Feb 2024), ~$645M (Jun 2024), ~$2.3B Series C led by ASML (Sep 2025, €2B at €12B/$14B valuation), and €3B/~$3.3B Series D led by Samsung (Sep 2026, >€21B valuation); EUR→USD figures per Wikipedia/press, not company-stated in USD. Excludes the separate $830M debt facility.) [source](https://en.wikipedia.org/wiki/Mistral_AI)
- Last round: series-d, $3.3B, 2026-09, led by Samsung Electronics, Scaleup Europe Fund (EQT), PSG Equity (confirmed) [source](https://mistral.ai/news/mistral-makes-sovereign-open-weight-ai-to-frontier/)
- Investors: Samsung Electronics, ASML, NVIDIA, General Catalyst, Andreessen Horowitz (a16z), Lightspeed Venture Partners, BlackRock, Advent, DST Global, Index Ventures, Bpifrance, Salesforce Ventures, BNP Paribas CIB, Eurazeo, Grand Duchy of Luxembourg, Xavier Niel, Eric Schmidt, Conviction, Redpoint Ventures, SV Angel
- Revenue: $400M (confirmed: CEO Arthur Mensch statement, reported by press) [source](https://www.trendingtopics.eu/mistral-increases-arr-to-400-million-builds-its-own-ai-data-centers/) ARR, 2026-02

French sovereign-AI champion; product on this card ('Mistral Compute' / 'Mistral AI Cloud', dedicated GB200/GB300/B300 capacity) is a division of Mistral AI, not a separately-financed entity, so all funding/ownership figures are for the whole company. Company also disclosed a target of ~$1B ARR by end of 2026 (Le Monde) and a separate $830M debt facility (Mar 2026) for European data centers, distinct from the equity rounds above.

## Caveats

- Compute list price, currency, minimum commitment, GPU SKU and ordinary CPU availability unknown.
- Unknowns remain null, not free; generic infrastructure modes are opt-in alt.
- Worked example assumes resources are released between active windows unless a retained-instance scenario is explicitly discussed.
- VAT basis: Unknown: reviewed pricing source does not explicitly establish VAT/tax inclusion.
- Residency: European dedicated capacity marketed; exact contracted facility, data control and subprocessor scope require contract.
- Status 2026-09-29: live. Product now marketed as 'Mistral AI Cloud' (mistral.ai/products/mistral-compute redirects to /products/aicloud/). Timeline on the page: GB200 serving production Feb 2026, first external customers onboarded Mar 2026. The AI Cloud page also sells Regional Endpoints / Priority Tier model inference (token-priced, not a sandbox/VM) and SLURM-based cluster access; no per-GPU-hour, per-vCPU or storage/egress price is published anywhere.

## How this provider charges


As of 2026-09-28. Price basis: public list/reference unless labeled historical, illustrative, promo or unpriced.

Mistral Compute offers reserveable European AI capacity; no public self-serve 4-vCPU/8-GiB sandbox price recovered. Chat/Vibe subscriptions, model-token prices and Koyeb CPU tariffs are different products. Do not merge them.

## Currency, VAT and residency
- Native catalog: EUR. FX: 1 EUR = 1.1378 USD, ECB 2026-09-28 ([source](https://www.ecb.europa.eu/stats/policy_and_exchange_rates/euro_reference_exchange_rates/html/index.cs.html)). USD-listed catalogs remain independent quotes.
- Tax: Unknown: reviewed pricing source does not explicitly establish VAT/tax inclusion.
- Residency/restrictions: European dedicated capacity marketed; exact contracted facility, data control and subprocessor scope require contract.

## Regimes
| Regime | When it applies | How billed | Native / USD numbers | Source |
|---|---|---|---|---|
| dedicated-ai | Dedicated AI capacity | allocated wall time; None second increment | EUR None/vCPU-h + None/GiB-h → USD None + None; null means unknown | https://mistral.ai/products/aicloud/ |

## Gotchas
- Compute list price, currency, minimum commitment, GPU SKU and ordinary CPU availability unknown.
- Unknowns remain null, not free; generic infrastructure modes are opt-in alt.
- Worked example assumes resources are released between active windows unless a retained-instance scenario is explicitly discussed.
- VAT basis: Unknown: reviewed pricing source does not explicitly establish VAT/tax inclusion.
- Residency: European dedicated capacity marketed; exact contracted facility, data control and subprocessor scope require contract.

## Required worked example
4 vCPU / 8 GiB requested; 50 simultaneous workers × 8 hours/day × 22 days = **8,800 instance-hours**, 1,100 daily sessions. CPU utilization 30% does not reduce allocated-resource charges. Snapshot retention is **50 GiB total for 730 hours**, not 50 GiB per worker; egress is **100 GiB total**. Root disk size was not specified, so root/volume costs are not silently assigned zero.
- dedicated-ai: compute **unknown**; cannot produce a comparable quote.
- 50 GiB retained snapshots: **unknown/unverified feature or price**, not USD 0.
- 100 GiB egress: **unknown**, not USD 0.
- Full all-in bill: **null** where any required storage/network/fee or eligibility fact is unresolved. Quota raises may be necessary for 200 aggregate vCPU / 400 GiB RAM.

## Sources
- https://mistral.ai/products/aicloud/
- https://mistral.ai/pricing/