Europe’s AI Gigafactories: Operational Sovereignty Without Technological Control
The EU will finance and govern strategic computing capacity while remaining dependent on non-European processors, interconnects and software.
Artificial intelligence is often described as an immaterial industry based on algorithms, data and specialist knowledge. Its most advanced capabilities nevertheless depend on a highly concentrated physical infrastructure: accelerators, high-bandwidth memory, optical interconnects, cooling systems, electricity supply and data centres capable of coordinating tens of thousands of processors. The European procedure for selecting up to seven AI Gigafactories therefore represents a material change in policy. The European Union is no longer limiting its intervention to AI regulation, research funding or support for individual companies. It is investing in the infrastructure that will determine who can develop advanced models, which data can be processed and how computing resources are allocated. For the defence sector, this transition is significant because access to large-scale computing increasingly conditions intelligence analysis, simulation, systems engineering, cyber operations, autonomous platforms and materials research. The central distinction, however, is between operational sovereignty and technological sovereignty. Europe may control where computing takes place and who can use it without yet controlling the technologies on which the infrastructure depends.