Mistral Raises €3 Billion Led by Samsung to Scale Sovereign, Open-Weight AI

French AI company Mistral said on September 8 that it raised €3 billion in a Series D round at a post-money valuation of more than €21 billion, led by Samsung Electronics. For enterprise AI leaders, the deal matters less as a funding headline than as a signal: a large group of chipmakers, banks, sovereign funds and asset managers is betting that companies will want an open-weight, deploy-anywhere alternative to the major U.S. model providers, and that such an alternative needs its own compute to compete.

What did Mistral announce?

According to Mistral's announcement, the round is the largest equity raise ever completed by a European technology company, three years after Mistral launched. Samsung led, with the Scaleup Europe Fund, managed by EQT, and existing investor PSG Equity as co-leads. New investors include Advent, funds and accounts managed by BlackRock, and the Grand Duchy of Luxembourg. Existing backers that took part include a16z, ASML, Bpifrance, BNP Paribas CIB, General Catalyst, Lightspeed, Nvidia and Salesforce Ventures.

W.Media reported that the round equals about $3.49 billion and values the company at roughly $24.4 billion. Mistral said the money will fund frontier research, more compute for training larger models, infrastructure and products, and commercial and international expansion. It said it operates in 20 countries and works with more than 125 enterprises, including Airbus, ASML and HSBC. The company did not disclose revenue.

Chief executive Arthur Mensch framed the raise in a post on X, quoted by AI Weekly, as a way to "make open and sovereign AI the technology frontier."

How does the round compare with Mistral's earlier funding?

The jump is steep. Mistral's previous rounds were a €105 million seed in June 2023, a €385 million Series A that December, a €600 million Series B in June 2024 and a €1.7 billion Series C in September 2025, according to Orrick, which advised the company. ASML led the Series C.

Much of the new money is likely to go into hardware. Vestbee reported that Mistral secured $830 million in debt financing in March to buy Nvidia chips for its first data center, a facility with 13,800 GB300 GPUs and 44 megawatts of powered capacity. That is modest next to the U.S. hyperscalers, but it gives Mistral a growing ability to sell models, inference and compute as one package.

Why does an open-weight, sovereign pitch resonate with enterprises?

Mistral's argument to buyers is control. Its announcement describes a sovereign AI offering in which data stays within an organization's boundaries, models can be customized, compute is private and predictable, and production systems are auditable. Its open-weight models can be run on premises, at the edge or in any cloud.

That pitch lines up with what practitioners have been telling AI Connect audiences. In a AI Connect webinar in July, Aaron Delp, Global Head of AI Technical Marketing at Mistral AI, put the company's case plainly: "This isn't renting AI. This is owning your AI." Speakers without a stake in Mistral have made a similar point. Rajesh Sinha, founder of Fulcrum Digital, told The AI Enterprise Conference 2026 that companies running many agents need their own oversight layer, warning against relying on any one vendor's tooling: "Have your own control panel, which is vendor agnostic." His seven-layer sovereign AI architecture assumes many workloads will run inside the customer's firewall.

The appeal is strongest in regulated industries and in countries with data residency rules, where sending sensitive data to an API hosted elsewhere is a hard sell. Earlier AI Connect sessions on sovereign AI for government found the same demand for models, data and inference under the customer's control.

What are the open questions for buyers?

Not everyone reads the raise as a pure growth story. AI Weekly noted commentary on social media suggesting the round reflects a shift away from competing head-on at the frontier, though Mistral's own language says the opposite. In June, before terms were set, European Business Magazine argued that Mistral's valuation rested more on strategic position than on current revenue, which it described as a fraction of what U.S. rivals report.

For enterprise buyers, three points are worth watching:

• Capacity. A single 44-megawatt site is a start. Customers evaluating Mistral's hosted inference should ask where capacity will come from and on what timeline.
• Model quality versus control. Open weights and deployment choice only pay off if the models are good enough for the task. Benchmark against your own workloads, not leaderboards.
• Investor ties. Samsung, ASML and Nvidia are strategic shareholders. That may help with chip supply, but procurement teams should understand how those relationships shape the roadmap.

What should enterprise leaders do now?

The practical takeaway is not to switch vendors because of one funding round. It is that a credible, well-funded open-weight option now exists alongside the closed U.S. platforms, which strengthens buyers' negotiating position and makes a multi-model architecture easier to justify. Companies that keep identity, policy, observability and cost tracking in their own layer will be best placed to use that choice, whichever models they run next year.

Frequently asked questions

How much did Mistral raise in its Series D?

Mistral announced a €3 billion Series D on September 8, 2026, at a post-money valuation of more than €21 billion. W.Media put the round at about $3.49 billion.

Who invested in Mistral's €3 billion round?

Samsung Electronics led, with the EQT-managed Scaleup Europe Fund and PSG Equity as co-leads. New investors include Advent, BlackRock-managed funds and Luxembourg; existing backers include a16z, ASML, Nvidia and Salesforce Ventures.

Why does Mistral's funding matter to enterprises?

Mistral sells open-weight models that can run on premises, at the edge or in any cloud. More funding for compute makes it a stronger alternative for companies that want control over data, models and deployment.

Filed under: AI Strategy & ROI, Data & AI Infrastructure, AI Governance & Trust