Mistral AI has closed a €3 billion round led by Samsung Electronics that values the Paris startup at more than €21 billion, nearly doubling the €11.7 billion price tag from its 2025 financing. The EU’s Scaleup Europe Fund and PSG Equity co-led, and ASML, which bought an 11 per cent stake for €1.3 billion last year, participated again. The company said the proceeds will fund model development and computing infrastructure.
Samsung leads with strategic intent
Samsung’s leadership of the round mirrors the ASML playbook: a strategic industrial backer writing a large check to secure preferred access. Mensch described the agreement as opening the door for further collaboration, the same phrasing he used for the ASML partnership. No breakdown of Samsung’s commitment versus the co-leads was disclosed, nor was the split between primary and secondary shares.
The sovereign pitch meets industrial demand
Mistral has positioned itself as a European sovereign alternative to OpenAI and Anthropic, a narrative that gained urgency after the Trump administration ordered Anthropic to block foreign-national access to its most advanced models this summer. The startup has converted that narrative into commercial traction with Airbus, BMW and ASML. Mensch said the funding also supports expansion in the US and Asia, though the sovereign framing remains the core differentiator in Europe.
Data centre bills are piling up
The capital intensity is visible in the balance sheet. Earlier this year Mistral raised US$830 million in debt for a data centre outside Paris built on Nvidia chips, and committed a further €1.2 billion for a facility in Sweden, its first infrastructure investment outside France. Mensch characterised the strategy as owning “significant assets,” a notable shift for a company that sells open-source models. The €3 billion equity round does not cover those commitments in full, implying further financing ahead.
What the terms don't say
The statement gives no premium to an undisturbed price, no break fee, no board composition changes and no liquidation preference details. For a round that nearly doubles valuation in under two years, the absence of term disclosure is itself a data point. Mensch’s claim that open source is more cost-efficient and easier to integrate into complex corporate systems remains untested at this capital scale.