Mistral News.

The US government ordered Anthropic to cut foreign nationals off from its most capable models, and Anthropic pulled them worldwide. Mistral CEO Arthur Mensch had warned for two years that US firms could switch off rivals' models. A viral 'Le Chaton Fat' meme about a fictional Mistral model made the point for him.

Mistral CEO Arthur Mensch positioned the company as a sovereign AI alternative after the US government ordered Anthropic to cut model access for foreign nationals. Mensch stated Mistral exists outside centralized state or corporate control. The Paris-based company offers open-source models that customers can run on their own infrastructure.

France plans €655 million in AI investments through 2030 and a Mistral-powered chatbot for state services. The country's security service will replace Palantir with Chapsvision.

Open-source AI models are among the worst at filtering Russian disinformation, according to Estonian researchers. Mistral's top model ranked 47th out of 60 tested models. Open-source generative models performed worse at removing false news than other models.

A study found that Mistral, Europe's AI champion, is vulnerable to Russian disinformation. The research identified specific weaknesses in the company's large language models. These vulnerabilities could allow the generation of misleading content targeting European audiences.

Mistral AI is in early talks to raise roughly €3bn ($3.5bn) at a valuation near €20bn ($23.2bn). The French open-source LLM startup, valued at €11.7bn last September, counts ASML as its largest shareholder. Mistral targets European governments and businesses as an AI infrastructure provider.

Mistral AI is in talks to raise €3 billion at a €20 billion valuation, nearly doubling its September 2025 valuation of €11.7 billion. The Paris-based AI lab plans to use the funds to build data centres and reach 1 gigawatt of computing capacity by 2030, competing with Amazon, Microsoft, and Google in physical infrastructure.