Mistral has raised EUR 3 billion from Samsung, the Scaleup Europe Fund, and PSG Equity, reaching a EUR 21 billion valuation.
According to the announcement, South Korea-based Samsung Electronics led the round, with participation from the Scaleup Europe Fund, an EU-backed vehicle managed by EQT, and existing investor PSG Equity. The new valuation nearly doubles the EUR 11.7 billion figure Mistral achieved a year earlier, when Netherlands-based chip equipment maker ASML led its previous funding round.
Infrastructure expansion and revenue targets
Mistral's chief executive officer, Arthur Mensch, said the fresh capital would be directed toward building the company's own data centre capacity and expanding its computing infrastructure for rental to third parties. He indicated that owned compute capacity is expected to roughly double over the next five years as the company works toward training larger models at greater speed.
Mensch had previously projected that Mistral's annual recurring revenue would exceed the equivalent of over EUR 850 million (USD 1 billion) this year. Following the new funding and partnership agreements, he said the company expects to surpass that figure, though he did not provide an updated projection. Moreover, he added that the fundraising is intended to support growth extending into 2027.
The Samsung partnership follows a similar model to Mistral's existing relationship with ASML, where the AI developer's technology has been integrated into the manufacturer's production processes. Mensch indicated that a comparable approach would be pursued with Samsung, though further details were not disclosed.
Positioning around sovereign AI and open-weight models
Mistral has sought to position itself as an alternative to both US and Chinese AI providers, aligning with growing interest across Europe and other regions in ‘sovereign AI’ development. The Scaleup Europe Fund’s backers include the European Commission alongside corporate investors such as Novo Holdings and Santander.
The company is expected to continue to focus on open-weight AI models rather than the closed, proprietary systems used by OpenAI and Anthropic, a strategy that puts it in direct competition with increasingly capable Chinese AI labs. Mensch said Mistral’s upcoming model releases would be competitive with rivals in the space.
He noted that Chinese AI developers generally do not operate directly with enterprise customers outside China, and that while some Chinese models can be deployed on Mistral’s own infrastructure, this does not create significant dependency, since data remains within Mistral’s systems. Mensch further argued that long-term reliance on Chinese models carries uncertainty, citing unclear upgrade cycles and the possibility of future export restrictions.
Describing volatility in the AI sector as significant at present, Mensch said Mistral aims to offer customers assurance of continued access to improved models over time, which he said requires the company to keep training its own systems rather than depending on external providers.
