EU Pledges €20B for AI Gigafactories: Scaling Compute, Control, and Capital

The MoU between the European Commission and the EIB Group marks a formal acceleration of Europe’s efforts to close the gap with global competitors in the infrastructure required to train next-generation AI models. Under the InvestAI programme, €20 billion of public capital is earmarked to enable four to five AI gigafactories, with auxiliary advisory and financing support from the EIB Group and EIF to make projects bankable. [1][3]

These gigafactories are ambitious in scale: roughly 100,000 cutting-edge AI chips per facility, around four times the size of existing “AI factories” currently under development within the EU. [1][2][4] Proposals from industry exceed expectations both in scale and scope—with 76 bids from 16 member states and €230 billion in planned investment, though many projects may never reach fruition. [5][6][2]

The synergy with other EU programmes—InvestEU, EuroHPC Joint Undertaking, TechEU—suggests a coordinated ecosystem approach. The EIB will use its advisory arm and capacity to provide loans or guarantees to attract private capital to complement public grants. [2][4] TechEU alone is set to mobilise €70 billion by 2027, with an aim of €250 billion together with partners. [4][7]

However, there are key open questions and risks. First, power supply: gigafactories with 100,000 advanced chips will likely need gigawatt-scale electricity supplies and robust infrastructure, which may not be equally available across all proposed sites. [4][6] Second, timeline: with formal calls expected early 2026, construction, deployment, and ramp-up will take years, potentially lagging rivals. Third, regulation: permitting, environmental assessments, and state aid rules may impede speed or raise costs. Fourth, private investment is critical to closing funding gaps; with public funds covering roughly 30%, failure to attract sufficient private capital threatens viability. [1][2][6]

Strategically, successful deployment of these gigafactories could shift the locus of AI compute infrastructure toward Europe, enabling domestic R&D, reducing dependence on imports of AI compute and chips, and giving European startups and research institutions more parity. At the same time, Europe’s global competitiveness in AI will partly depend on aligning regulatory, energy, and talent ecosystems to support these high-cost, high-scale undertakings.