On June 16, 2026, a study by Estonian researchers published by the Financial Times found that Mistral's top AI model...

On June 16, 2026, a study by Estonian researchers published by the Financial Times found that Mistral's top AI model ranks 47 out of 60 tested models at filtering Russian disinformation, indicating a significant safety vulnerability in open-source models (FT). Concurrently, Bloomberg reported that France plans €655M in AI investments through 2030 and will deploy a Mistral-powered chatbot for state services, signaling government adoption (Bloomberg). However, the study's findings could undermine trust in Mistral's models for sensitive applications, particularly in government and security contexts. The negative safety assessment, while based on a single source, is corroborated by the FT's reputation; the government contract simultaneously presents a positive adoption signal. Given the conflicting signals, the trust risk from the disinformation weakness is likely to weigh on Mistral's reputation and may slow broader enterprise adoption, suggesting a downward direction for trust and potential future revenue.