Europe’s AI sovereignty is under threat. Could Mistral be the answer?

In May 2024, Arthur Mensch, the CEO of Paris-based AI company Mistral, delivered an urgent warning to the French National Assembly, claiming that Europe had only two years to build its own AI infrastructure or face the risk of becoming a "vassal state" to the United States. This warning was based on the prospect of foreign AI models becoming intertwined with the country's military forces, which could prove irreparable.

Just one month later, the U.S. Commerce Department temporarily blocked foreign access to Anthropic's powerful AI model, Mythos, highlighting the risks of such dependency.

Mistral, founded in April 2023 by Mensch, his college friend Timothée Lacroix, and Meta alum Guillaume Lample, aims to address Europe's lack of homegrown research labs capable of competing with American front-runners in building large language models. The company released its debut model, Mistral 7B, in 2023, which outperformed Meta's Llama 2 13B in benchmark tests despite being about half the size. In December 2023, Mistral raised €385 million ($440 million), pushing its valuation past €2 billion ($2.3 billion).

Despite its ambitions of becoming a sovereign AI champion, Mistral still relies on U.S. technology for chips and cloud-computing infrastructure, and its models are not yet competitive with the bleeding-edge AI systems from Anthropic, OpenAI, or Google DeepMind. Additionally, around 40% of Mistral's revenue comes from non-European clients, raising questions about its ability to compete globally.

However, recent events and a shift in large enterprises' approach to purchasing AI models may present an opportunity for Mistral to prove its critics wrong and establish itself as Europe's most viable alternative in the AI landscape.