Mistral AI pursues €3 billion funding round at €20 billion valuation to challenge US rivals

Mistral AI Targets Major Funding Round to Fuel European AI Ambitions

Mistral AI is in early discussions to raise approximately €3 billion ($3.5 billion) at a €20 billion valuation, according to Bloomberg reports citing anonymous sources 1

1

2

2

. This AI funding round would nearly double the €11.7 billion valuation the French AI startup received during its Series C round last September, when semiconductor equipment giant ASML invested €1.3 billion for an 11% stake 2

2

4

4

. The discussions remain at an early stage, and terms could shift based on investor demand, with the valuation potentially climbing even higher 2

2

.

Founded in 2023 by researchers from Google DeepMind and Meta, Mistral AI has positioned itself as a sovereign European alternative to American AI dominance 4

4

. The company's CEO Arthur Mensch has made clear his vision: Europe must own and operate its own AI infrastructure rather than depend on Silicon Valley 2

2

. This strategy has taken on added urgency as European countries distance themselves from American tech, creating opportunities for homegrown alternatives.

Building European AI Infrastructure Through Owned Data Centers

The capital raise would replenish a war chest that Mistral AI is burning rapidly to build physical infrastructure. The company is constructing data centers it owns and operates, including a major facility near Paris and a €1.2 billion build-out in Sweden 2

2

. This ownership model distinguishes Mistral from competitors who rent compute capacity. The Paris-based company has partnered with France's army, Luxembourg's government, and several major European companies to deliver tailored solutions 1

1

.

Mistral's pitch has become increasingly industrial, signing deals with Airbus, BMW, and other European manufacturing giants for AI model development tailored to engineering and manufacturing processes 2

2

4

4

. The company is also developing an alternative to Anthropic's Mythos model for European banks, software designed to identify cybersecurity vulnerabilities. Mensch described this capability as a national security risk, stating last month, "We must have control over this technology" .

Physics AI Push Expands Mistral's Technical Frontier

The funding discussions come shortly after Mistral AI announced plans to develop physics AI products for industrial engineers 3

3

. This technology will enable engineers to generate multiple product design variations and test them in simulations. In May, the company acquired startup Emmi, which developed tools for building physics AI models optimized to solve partial differential equations—functions useful for describing complex physical phenomena 3

3

.

Mistral researchers have published multiple papers on physics AI, including datasets related to computational fluid dynamics for analyzing aerodynamic properties of cars and aircraft 3

3

. The company has released over half a dozen open-weight models, with the newest addition, Mistral Medium 3.5, debuting in April with 128 billion parameters 3

3

. These open-weight models allow anyone to customize them as needed, reflecting a more transparent approach than American rivals 1

1

.

Mistral generates revenue through paid cloud services powered by its models, including Mistral Vibe, an AI assistant that can summarize documents, generate code, and troubleshoot factory equipment 3

3

. The company also offers Studio, an AI agent development platform, alongside AI-optimized cloud infrastructure. In March, Mistral introduced Forge, a system enabling enterprises to build AI models trained on proprietary knowledge rather than publicly available data, allowing organizations to retain control over models, data, and intellectual property 5

5

.

The Scale Gap Mistral Must Bridge to Compete with US and Chinese Rivals

Despite its ambitions, Mistral AI faces a significant scale disadvantage. The company has raised only about $4 billion to date, a fraction of what OpenAI ($186 billion) and Anthropic ($161.25 billion) have secured 1

1

. These American labs also command vastly higher valuations—OpenAI last raised at $852 billion in March, while Anthropic was worth $965 billion last month 4

4

. Both plan to go public this year, following SpaceX's recent IPO at approximately $1.8 trillion 2

2

.

Mistral's models and chatbot have gained far less traction with businesses and consumers than those from OpenAI, Anthropic, and Chinese competitors 4

4

. The company also faces fresh competition from Prometheus Inc., which raised $12 billion this week and was launched by Amazon.com founder Jeff Bezos to automate hardware engineering tasks similar to those Mistral targets with physics AI 3

3

.

To compensate for these disadvantages, Mensch is pushing on multiple fronts. He told CNBC this week that Mistral is exploring designing its own AI chips for the first time to reduce reliance on Nvidia 2

2

. Whether €3 billion buys enough compute capacity and market adoption to stay competitive remains the critical question investors must weigh. Mistral's earlier backers include France's state bank Bpifrance and prominent US venture firms including Lightspeed Venture Partners, General Catalyst, and Andreessen Horowitz 4

4

.