“We believe the right approach is to make models open and let the community build on top of them.”
“Europe should not be dependent on American or Chinese companies for its AI infrastructure.”
“The cost of training frontier models will come down dramatically. This is not a game only for the biggest companies.”
“Regulation should focus on applications, not on the technology itself. You don't regulate engines, you regulate cars.”
“We want to be the Linux of AI -- open, efficient, and everywhere.”
“Open-source models are not less safe. They are more safe because thousands of researchers can study them.”
Arthur Mensch has emerged as the most prominent voice for European AI independence and the open-weight model philosophy. As CEO of Mistral AI, which he co-founded after leaving Google DeepMind, Mensch has built the company into a serious competitor to American AI labs while maintaining a commitment to releasing model weights publicly.
His public statements consistently emphasize three themes: the superiority of open models for safety and innovation, Europe’s need for technological sovereignty in AI, and the importance of regulating AI applications rather than the underlying technology. This final point put him at odds with the EU AI Act’s approach during its drafting, where he argued that regulating foundation models themselves was counterproductive.
Mensch’s analogy comparing AI regulation to automotive regulation — focusing on the car rather than the engine — became influential in European policy circles. He argued that the same model could be used safely or dangerously depending on its application, making application-level regulation more logical than model-level restrictions.
On the business side, Mensch challenges the narrative that frontier AI requires billions in capital and can only be built by a handful of American companies. Mistral’s efficient models, which rival larger competitors at a fraction of the training cost, provide evidence for his claim that the AI development game is more open than incumbents suggest.
His positioning as a European AI champion has given him outsized influence in Brussels, where his views on regulation and open-source are taken seriously by policymakers seeking a European perspective distinct from American tech industry lobbying.