French government plans to hire sovereign AI companies like Mistral, excludes OpenAI

France is drawing a line in the silicon. The French government is moving to exclude OpenAI and other US-based AI providers from key public-sector contracts, instead funneling work toward homegrown companies like Mistral AI as part of an aggressive “sovereign AI” strategy designed to keep critical data and infrastructure under French and European control.

Defense deals and digital bureaucrats

The cornerstone of this push landed in January 2026, when the French Ministry of the Armed Forces awarded Mistral a framework agreement granting defense agencies and related institutions access to its AI models and services. The catch: everything runs exclusively on sovereign French infrastructure. No data leaves the country, no foreign company touches the pipeline.

Mistral CEO Arthur Mensch drove the point home in May 2026, warning France’s National Assembly against dependence on foreign AI for military applications.

France also plans to deploy a Mistral-powered AI chatbot to roughly one million civil servants, essentially embedding a domestic large language model into the daily workflow of the entire French bureaucracy.

The money behind the mandate

France’s broader AI strategy includes over €109 billion in investments, a figure announced at the February 2025 Paris AI Action Summit.

Mistral was founded in Paris in 2023. By early 2026, the company’s valuation exceeded $14 billion. Led by Mensch, a former DeepMind researcher, Mistral has positioned itself not just as a credible open-weight model provider but as a geopolitical instrument of French tech policy.

The company is also collaborating with German enterprise software giant SAP on AI tools for public administrations. That Franco-German initiative was announced in November 2025.