This week, Mistral announced it raised €3 billion in a Series D funding round, pushing its post-money valuation past €21 billion. With the influx of cash — $3.5 billion in US dollars — it plans to expand its frontier research, scale compute capacity for model training, and grow its infrastructure.
The funding round was led by Samsung Electronics, with Scaleup Europe Fund, managed by EQT, and existing investor PSG Equity acting as co-leads.
Mistral’s allocation strategy signals a bet on an alternative path to concentration in AI power. While some open-weight advocates argue that releasing model weights democratises access and reduces dependence on proprietary APIs, critics note that running powerful models still requires enormous compute capacity, which remains concentrated among a handful of labs, chip suppliers, and infrastructure providers.
Dario Amodei, CEO and co-founder of Anthropic, challenged the open-weight vision last month in an exchange on X, writing that open weights “are nowhere near a sufficient solution because they simply shift the concentration somewhat to those with the most compute and chips.”
Mistral’s response is to build more of the stack. The company claims to be the only AI company in the world building the full stack required to answer how organizations can use AI for mission-critical needs without surrendering control of infrastructure and intelligence. Its stack covers open-weight models, the underlying compute and infrastructure, and downstream products that bring models into production.
Looking ahead, Mistral says it aims to free customers from dependence on a single vendor’s roadmap, pricing, and availability, allowing them to build on its stack “without exposing their most valuable data, workflows and” — the sentence cuts off in the original announcement.
The round represents a major vote of confidence in Europe’s ability to compete at the frontier of AI development, and underscores the growing recognition that open-weight models alone may not be enough without corresponding investment in compute infrastructure.