Mistral Seeks $3.5 Billion to Build European AI Infrastructure

Mistral AI is in talks to raise about 3 billion euros (about $3.5 billion) at a valuation of about 20 billion euros (about $23.1 billion), a figure that could go higher depending on investor demand, Bloomberg reported Friday (June 12), citing unnamed sources.

The French AI startup’s discussions with investors are at an early stage, and the terms could change, according to the report.

Mistral did not immediately reply to PYMNTS’ request for comment.

The company was valued at 11.7 billion euros (about $13.5 billion) in a September Series C funding round in which it raised 1.7 billion euros (about $2 billion). Mistral said at the time in a press release that it would use the new funding to fuel its scientific research.

The round was led by semiconductor equipment manufacturer ASML, and Mistral CEO Arthur Mensch said in the release that the two companies operate in the same value chain.

“We have the ambition to help ASML and its numerous partners solve current and future engineering challenges through AI, and ultimately to advance the full semiconductor and AI value chain,” Mensch said.

Advertisement: Scroll to Continue

According to Friday’s Bloomberg report, Mistral was founded in 2023 by researchers from Google DeepMind and Meta, has positioned itself as a European alternative in the AI market, and has focused on serving as a provider of AI infrastructure for European governments and companies.

Mistral has discussed offering European banks and other institutions a cybersecurity-focused AI model as an alternative to Anthropic’s Mythos, and Mensch said, “We must have control over this technology,” per the report.

The company introduced a system called Forge in March, saying it enables enterprises to build AI models that are trained on their proprietary knowledge rather than publicly available data.

Describing the benefits of Forge in a press release, Mistral said that by building models trained on their own knowledge, enterprises can retain control over the models, data and intellectual property; build agents that can navigate internal systems, use tools correctly and make decisions within the organization’s constraints; build both dense models and mixture-of-experts ones; and continuously refine models as needed.

“As organizations integrate AI agents into core operations, the ability to encode institutional knowledge into model behavior will become increasingly important,” Mistral said in the release.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.