The European Commission formally opened its call for AI Gigafactory proposals on July 30. €10 billion ($11.4B) in public funding targets €30B with private investment. Seven planned sites must each host at least 100,000 AI chips. Bidding closes November 12, 2026. The earlier informal call drew 77 proposals across 16 member states.
The European Commission formally opened its call for AI Gigafactory proposals on July 30, putting €10 billion ($11.4 billion) of public money on the table with the expectation of attracting enough private investment to push the total past €30 billion. Seven sites are planned, and the bidding window closes November 12, 2026.
Each gigafactory must host at least 100,000 AI chips — roughly four times the scale of the largest current data-center deployments in the EU. The Commission isn't aiming for parity with US hyperscalers. It's aiming to establish a credible European alternative for AI compute at a moment when dependence on non-European cloud providers is becoming a strategic liability.
The Scale Problem #
Europe's AI infrastructure gap is real and widening. US hyperscalers — Amazon, Microsoft, Google — collectively deploy more AI compute capacity in a single quarter than exists in the entire EU. Chinese state-backed infrastructure projects are building at comparable scale. Europe's largest AI training clusters are an order of magnitude smaller than what's available in either the US or China.
The gigafactory program is designed to close that gap in one leap. Seven sites at 100,000 chips each means 700,000 AI accelerators deployed under European jurisdiction. If the private investment materializes at the expected ratio, the total capital deployment would make this one of the largest civilian infrastructure programs in EU history.
The Demand Signal #
The earlier informal call for proposals drew 77 submissions across 16 member states covering 60 potential sites. That's not just enthusiasm — it's evidence that European governments and private consortia have been waiting for a funding mechanism large enough to make industrial-scale AI infrastructure viable within EU borders.
The geographic spread matters. The Commission wants facilities distributed across member states rather than concentrated in a single country or region. This is partly politics — every member state wants the economic development that comes with hosting major infrastructure — but it's also resilience. Distributed infrastructure is harder to disrupt and reduces single-point-of-failure risk.
What 100,000 Chips Actually Buys #
A 100,000-chip cluster represents a significant AI training capability. For context, training runs for frontier models like GPT-5 class systems typically use clusters in the 20,000-50,000 GPU range, though next-generation training runs are pushing toward 100,000+. The gigafactories wouldn't just enable training — they'd make Europe a credible host for frontier AI research that currently can only happen in US or Chinese facilities.
The infrastructure also supports inference at scale. European companies building AI-powered products currently route most compute through US cloud providers. Local infrastructure at gigafactory scale would change the economics and latency characteristics for European AI deployment.
The November Deadline #
Bidding closes November 12, 2026, giving consortia roughly 15 weeks to assemble proposals. That's tight for infrastructure projects of this scale, but the earlier informal call means many potential bidders have been working on their submissions since before the formal opening.
The real test comes after November. Can the Commission move from proposal evaluation to actual construction on a timeline that keeps the program relevant? AI infrastructure cycles move at hyperscaler speed — 18-month buildouts from groundbreaking to operation. European public infrastructure projects haven't historically operated at that tempo.
The Broader Context #
The gigafactory program is Europe's answer to a simple question: who controls the physical infrastructure that AI runs on? Right now the answer is predominantly US companies, with Chinese state-backed infrastructure emerging as the only credible alternative at scale. The gigafactories won't match US or Chinese total capacity, but they would establish a third pole — infrastructure under European jurisdiction, governed by European regulations, accessible to European companies and researchers on European terms.
Whether that's worth €10 billion in public money depends on how much you think AI infrastructure sovereignty matters. The Commission has made its bet clear.
Q: What exactly is an AI Gigafactory?
Q: How much public funding is the EU committing?
Q: When does the bidding close?
Q: How much demand is there?
Q: Can Europe realistically compete with US and Chinese AI infrastructure?
Get AI news in your inbox
Daily digest of what matters in AI.