Europe weighs AI sovereignty beyond US dependence

The European Commission opened bidding in late July for up to seven AI gigafactories, offering 10 billion euros ($11.5 billion) in public funding, seeking to mobilize 20 billion euros more and requiring at least 100,000 advanced AI processors per site. The move reflects a broader European push to turn AI sovereignty from a policy phrase into infrastructure.

Europe has strong universities, a large single market, ASML and Mistral, but its AI ecosystem remains heavily tied to American cloud, computing and model providers. European providers’ share of Europe’s cloud market fell from 29 percent in 2017 to 15 percent by 2022, reinforcing concerns that access to US platforms could become leverage in the future.

Greater sovereignty is framed as strategic optionality rather than isolation. Open-weight models are presented as one practical tool, including Moonshot’s Kimi K3, released in July at 2.8 trillion parameters, because downloaded model weights can run on European hardware under European law. The argument also calls for evidence-based risk assessment of Chinese technology and limited cooperation in non-sensitive sectors, rather than nationality-based exclusion that narrows Europe’s choices.