Skip to content
TECH CEO Daily
AIFunding

Mistral AI raises €3 billion in Samsung-led round, valuing it above €21 billion

The Paris model maker says the money goes to research and compute it owns, as it pitches European customers on control over where their AI runs.

By · Editor

· 3 min read · Fact-checked

The 60-second brief

  • 1Mistral AI closed a €3 billion Series D led by Samsung Electronics at a post-money valuation above €21 billion.
  • 2CEO Arthur Mensch told CNBC Mistral plans to rely fully on self-built compute over the long term.
  • 3Buyers gain a better-funded European supplier, but its reported targets of 200 megawatts by end-2027 and 1 gigawatt by 2030 trail the 2.16-gigawatt Queensland campus where Anthropic agreed to lease a building.

The news

Mistral AI raises €3 billion in new equity, the French model developer announced on September 8, 2026, in a Series D led by Samsung Electronics (005930.KS) that values it above €21 billion post-money. The money is earmarked for research and computing capacity.

Samsung was joined as co-leads by the Scaleup Europe Fund, which is managed by EQT, and by existing backer PSG Equity. Mistral named Advent, funds managed by BlackRock (BLK) and the Grand Duchy of Luxembourg as new investors. Returning shareholders include a16z, ASML (ASML), Bpifrance, General Catalyst, Lightspeed, Nvidia (NVDA) and Salesforce Ventures.

In its announcement, Mistral said the capital will go to frontier research, compute for training models, infrastructure and international expansion. Chief executive Arthur Mensch told CNBC, as reported by Quartz, that the company will build and own data centers while renting extra capacity in the meantime, with owned compute growing roughly 100% over five years.

That build-out is already under way. According to Quartz, Mistral's first facility at Bruyères-le-Châtel, near Paris, houses 13,800 Nvidia GB300 chips, a second site in Sweden carries a €1.2 billion investment, and the company is targeting 200 megawatts of European capacity by the end of 2027. TechCrunch reported a longer-range goal of 1 gigawatt in Europe by 2030.

The commercial case rests on control. Mistral says it serves more than 125 enterprises across 20 countries, naming Airbus, ASML and HSBC, and pitches customers on running AI for critical work without surrendering the underlying infrastructure. Mensch said Mistral is on track to pass $1 billion in annual recurring revenue before year-end, according to Quartz, which put the prior Series C valuation at €11.7 billion a year earlier.

TechCrunch also reported that Mistral launched tools in August 2026 that let customers choose the region where their AI queries are processed, and that it has begun hosting other companies' open-weight models on its infrastructure, including Chinese models. Open-weight models are those whose trained parameters are published for anyone to download and run.

The numbers

Series D size
€3 billion
Post-money valuation
More than €21 billion
Prior (Series C) valuation, per Quartz
€11.7 billion
Enterprise customers
More than 125, across 20 countries
Planned European capacity by end-2027, per Quartz
200 megawatts
Nvidia GB300 chips at the Paris-area site, per Quartz
13,800

Why CEOs should care

For CIOs and procurement teams, a well-capitalized European supplier changes the shortlist for regulated workloads. Before signing, ask Mistral which region each workload will run in, whether it runs on owned or rented capacity during the build-out, and what service levels apply while the company adds data centers. Mistral's target of 200 megawatts by the end of 2027, as reported by Quartz, is modest next to the 2.16-gigawatt Queensland campus where Anthropic has agreed to lease a building, so capacity guarantees matter.

CFOs should read the round as a vendor-viability signal and a pricing question. The valuation has risen from €11.7 billion to more than €21 billion in about a year, according to Quartz, and fresh equity lowers the risk that a key supplier stalls mid-contract. But a company that owns its data centers carries heavy fixed costs, so multi-year agreements should include price protections and clear exit terms.

For CISOs and boards, the sovereignty pitch cuts both ways. Choosing where data is processed helps with European data-protection rules and supplier-concentration risk tied to US providers. Hosting third-party open-weight models, including Chinese ones, gives technical teams more choice but calls for a written policy on model provenance, testing and approved uses before anyone deploys them.

The bigger picture

The round shows how sovereign AI, meaning AI infrastructure a country or region controls itself, has become a commercial category rather than a policy talking point. It also shows hardware makers buying into model developers: Samsung leads, while Nvidia and ASML are returning backers. A recent US deal shows the gap in scale. Eight days after Mistral's announcement, Anthropic agreed to lease Building One at a proposed Queensland data-center campus with 2.16 gigawatts of peak capacity, according to ABC News, which noted that the lease is subject to Foreign Investment Review Board approval and that the full campus is expected to take four to six years to build. That leaves Mistral competing on control and locality rather than raw size.

What’s next

Three markers will show whether the strategy holds: whether Mistral passes the $1 billion recurring-revenue mark Mensch forecast for year-end, progress on the Swedish site and the 200-megawatt target for 2027, and whether Samsung's lead investment turns into a commercial partnership. Mistral has not announced such a partnership.

What “Fact-checked” means

Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.

What we checked
Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
How
A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
Who
The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, . A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
If something is wrong
“Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error

How we fact-check →

Companies in this story

Mistral AISamsung ElectronicsArthur MenschSovereign AI

Earlier coverage of Mistral AI

All Mistral AI coverage →

Written by

Editor · Technology & Business Writer

Hussein is a writer and business technology enthusiast focused on the intersection of technology, entrepreneurship, finance, artificial intelligence, and digital innovation.

CoversAICybersecurityBig TechSaaSStartupsFintech

About this story. Researched from primary sources whenever they are available and fact-checked before publication.

Published by Tech CEO Daily, an independent publication. Masthead · Editorial standards

Follow Tech CEO Daily on Facebook for the day’s top stories in your feed.

Free newsletters

The technology briefing for people running businesses.

Daily, weekly, bi-weekly or monthly. You choose.

How often

The Daily Brief · Monday to Saturday, 7 a.m. ET

Free forever. One click to unsubscribe. We never sell your email.