Europe opens its call for seven AI gigafactories worth 30 billion euro

Europe is putting real money on the table in the AI compute race. On July 30, 2026, the European Commission officially opened the call for up to seven AI gigafactories: facilities combining advanced chips, cloud and data centers to train next-generation AI.

The money and the timeline

The funding plan stacks public and private capital. Roughly 10 billion euro of public funding, drawn from the EU budget plus member states, is designed to unlock about 20 billion euro in private investment, for more than 30 billion euro in total.

The schedule is set: bids close November 12, winners are expected in early 2027, and the sites should be operational around mid-2028.

The interest is real

This is not a call into a void. 76 consortia across 10 countries, from Germany and France to Poland and Greece, have signaled preliminary interest. On the supply side, Nvidia, AMD and Qualcomm have signed letters of intent to provide chips.

That breadth matters politically as well as industrially. A gigafactory program spanning ten member states, rather than concentrating in one or two, is the kind of distribution that keeps EU-level funding politically sustainable.

The unspoken part: sovereignty

The strategic subtext is barely subtext at all. Europe currently trains its AI on American and Chinese compute. Its frontier-scale training runs happen on infrastructure owned, operated, or supplied by companies outside its jurisdiction. These seven sites are Europe’s bid for sovereignty: the ability to train next-generation models on European soil, under European rules, without depending on someone else’s export policies.

The open question

Can 30 billion euro close the compute gap? The honest answer is that nobody knows yet. Frontier compute buildouts elsewhere are measured in larger numbers and faster timelines, and mid-2028 is a long way off in AI years. The counterargument: Europe does not need to match the largest builders dollar for dollar to secure strategic autonomy for its own labs, researchers, and industries.

Too little too late, or the foundation of an independent European AI stack? Check back in 2028, when the doors are supposed to open.