French MPs have raised concerns about the strategic partnership between Paris-based startup Mistral AI and Microsoft, citing worries about competition and data sovereignty in the cloud sector. Mistral AI and Microsoft recently announced that Mistral AI's AI model, Large, will be available on Microsoft's cloud platform, Azure. However, the partnership has faced criticism over outsourcing data storage to non-EU companies and increasing the concentration of power among major tech players.

French lawmakers have voiced worries about the extraterritoriality of US laws, such as the Foreign Intelligence Law and the Cloud Act, which could potentially compromise the security of foreign data. To safeguard sensitive information, France has implemented the SecNumCloud certification and the 'cloud-to-the-centre' doctrine, requiring the storage of public sensitive data on certified French clouds. While some MPs believe that using Microsoft's cloud is acceptable as long as sensitive data is not shared on Azure, others argue that the 'cloud-to-the-centre' doctrine must be respected and non-sovereign clouds avoided.

Mistral AI has expressed its willingness to collaborate with French cloud provider OVHcloud, which has received the highest SecNumCloud certifications. Additionally, concerns have been raised about competition, with lawmakers emphasising the need to support European innovation and prevent the recurrence of historical monopolies. Some MPs have criticised Microsoft for alleged harmful practices in the cloud sector, as identified by the French competition authority. The authority is currently investigating competitive practices in the AI and cloud sectors, as well as Big Tech investments in AI startups.

The European Commission will also review the Mistral AI-Microsoft partnership. Despite the concerns, some MPs advise against rushing to conclusions, viewing the partnership as an opportunity for Mistral AI to grow. The Economy Minister, Bruno Le Maire, has praised Mistral AI's potential and expressed confidence in France's ability to become a leading nation in AI development. Furthermore, France, Germany, and Italy are advocating for mandatory codes of conduct for foundation models in the EU's AI law.