Samsung Leads Mistral AI’s €3 Billion Funding Round Valuing Company Above €21 Billion
Image: Zonebourse

Technology and Science · updated 3h ago · 3 min read

Samsung Leads Mistral AI’s €3 Billion Funding Round Valuing Company Above €21 Billion

Happened

Samsung-led €3 billion funding values Mistral above €21 billion. EU Scaleup Europe Fund and PSG Equity co-led the round.

Split on

Whether Mistral hosts/uses Chinese models.

Left out

2 of 3 outlets skipped it: france considers a golden share to keep strategic control.

21outlets compared

Brief IACNBCCTechCXO DigitalpulseDiario FinancieroDiarioBitcoinEquitypanditEuronews

Same story, two versions

tap a side to read it in full

Le Monde.frLe Monde.fr

“But today, the start-up sells computing power and hosts Chinese models,” criticized

Translated from the original.

Read the original

CNBCCNBC

Chinese AI labs are not actually operating or working with enterprise customers outside of China.
Read the original
VS

Le Monde frames Mistral as hosting Chinese models, while CNBC stresses limits on Chinese enterprise use.

Samsung-led funding surge

French AI company Mistral said it raised €3 billion in fresh funding led by Samsung Electronics, valuing the company at more than €21 billion as it positions itself as a European alternative to rivals like OpenAI and Anthropic.

Mistral on Tuesday said it raised 3 billion euros ($3.5 billion) in fresh funding led by memory chip giant Samsung

CNBCCNBC

Reuters reported that Mistral raised €3 billion at a valuation of around €21 billion ($24 billion) in what the three-year-old AI company said was the biggest equity funding round by a privately owned European technology firm.

Image from Brief IA
Brief IABrief IA

Mistral’s chief financial officer Johan Bergqvist told Reuters that the funds will be used to power Mistral’s models and invest in so-called frontier research as the company seeks to compete with larger rivals.

Arthur Mensch, CEO of Mistral, told CNBC that the funding would go toward building more infrastructure, including its own data centers, and renting out computing capacity.

Mensch said, "Long term, the plan is to fully rely on capacity that we are building ourselves," adding that the amount of compute Mistral owns would grow "around 100% in the next five years."

Compute, models, and revenue

Mensch told CNBC that Mistral will train "bigger and faster models" and that it expects its annual recurring revenue to exceed $1 billion this year, while also declining to give a revised revenue figure.

Reuters reported that Mistral is on track for $1 billion of annual recurring revenue by the end of the year, and that its more than 125 customers can download and customise models on their own servers.

Image from CNBC
CNBCCNBC

CNBC said Mistral is looking to take a different approach to OpenAI and Anthropic by working with individual companies to create custom AI tools and integrate them with their businesses.C

Reuters described Mistral’s open models as allowing customers to run them on their own servers, and said the company’s client base is becoming more geographically diverse with growth particularly in Asia and North America.

Mensch told CNBC that Mistral’s plan is to be "a trusted partner for our customers," and said the only way to guarantee access to better models is by continuing to train its models itself.C

Sovereignty and strategic stakes

The funding round also sharpened the political and strategic framing around technology independence, with Reuters quoting Bergqvist saying, "The fact that the EU or Europe have to have their own kind of AI provider in the game is important."

"The fact that the EU or Europe have to have their own kind of AI provider in the game is important," Bergqvist said

ReutersReuters

Reuters said the push for technological independence was highlighted in June by a U.S. decision to limit foreign access to two advanced models developed by Anthropic, and Bergqvist said Europe must maintain access and not compromise its supply chain.

CNBC reported that Mistral is building out its own data center footprint and is focusing on open-weight AI models versus proprietary closed systems from OpenAI and Anthropic.C

The company told CNBC that Chinese AI labs are not actually operating or working with enterprise customers outside of China, and that deploying Chinese models on Mistral’s infrastructure "is not creating a very strong dependency" because the data stays with Mistral.

In the same CNBC interview, Mensch said, "Now we also need to be a trusted partner for our customers," tying the stakes of the new funding to customers’ demand for certification that they will get access to better models over time.C