On 20 October 2026, a “power event” at Google Cloud in London affected banks and fintech companies across Europe. Not long after, two things happened: (1) the French Ministry of the Interior took local government websites offline until critical vulnerabilities in the software running them were fixed; (2) France’s OVHcloud announced its first AI model for early 2027, and its CEO, Octave Klaba, argued that Europe’s AI infrastructure must be built on a European scale, since any national strategy would be “doomed to failure“, while Alphabet’s legal chief, Kent Walker, argued that European concerns about American dominance in AI could be addressed through partnerships.
Although France publicly champions digital sovereignty, incidents such as these demonstrate just how dependent Europe remains on infrastructure and software it does not control. This can also be observed in French law itself. SecNumCloud, France’s most stringent cloud security certification, which requires protection against non-European laws with extraterritorial reach, was granted in December 2025 to S3NS, a subsidiary of Thales created in partnership with Google Cloud. This shows that, in practice, such sovereignty may ultimately depend less on who develops the technology than on who legally controls it.
However, not everyone in the European technology sector accepts this view. Mistral CEO Arthur Mensch warned that leaving cutting-edge AI in American and Chinese hands would be a “fundamental mistake”, condemning Europe to “a regime of economic dependence unprecedented in the history of civilisations”. And France appears to be seeking greater independence in the sector. Nevertheless, if Klaba is right, success in this endeavour will depend on demand and investment on a European scale. The question remains whether the EU will continue pursuing sovereignty through partnerships with American giants or choose to develop its own technology.
