EU bets $11.4 billion on seven AI gigafactories to close its gap with the US and China

Brussels is putting €10 billion ($11.4 billion) of public money into seven new AI gigafactories. The goal, announced Thursday by the European Commission, is to draw another €20 billion ($22.8 billion) in private capital for a combined target near €30 billion (about $34.2 billion).

Each planned facility will house at least 100,000 cutting-edge AI chips, roughly four times the processing power of the data centers currently running inside the EU, according to the Commission's announcement carried by the AP wire (AP via ClickOrlando). Firms can now bid for the contracts to build the sites.

"AI gigafactories are a strategic necessity for Europe," said Henna Virkkunen. Brussels is framing the program as protection against dependence on foreign AI providers that officials warn could be "weaponized" against Europeans.

The EU already runs 19 AI data centers, ranging from Finland to Spain. The seven new gigafactories will more than double that network, and each one will outclass what is in place today. The most concrete way to see the gap is to look at what the bloc now calls one of its largest existing AI sites: the Paris campus run by France's Mistral.

The Commission's own June 2026 report to the European Parliament names the layers the €10 billion does not buy. It says EU businesses and public authorities will continue to depend on US AI providers, that all five of the bloc's largest cloud service providers are American, and that this dependence "will continue to expose data to third-country access." The same report concludes that Europe manufactures few of the components that go into a modern data center and that European electricity can cost two to three times the levels in the United States or China.

The structural drags are the reason a 2025 assessment by the U.S. Federal Reserve, cited in the AP wire, finds Europe lagging the United States and China in crucial AI development sectors. The U.S. still captures the lion's share of private AI investment, and China holds a large electrical power capacity for data centers. The wire also frames this as part of a wider fight: U.S. President Donald Trump has railed against EU tech regulations, and China has restricted the supply of minerals critical to the sector.

The €10 billion buys one specific layer: raw compute. Each gigafactory will be a frontier-scale site for training and running large AI models, sized to anchor tenants that today rent capacity from U.S. hyperscalers. The risk is that a bigger, faster data center inside the EU sits on top of an American cloud layer, imported chips, and the same expensive power grid. The Commission's June report concedes that those layers will not be replaced by the program.

The bid window is now open, and the test of the bet will be how much of the €20 billion private draw arrives. A high draw would create second-order effects the June report did not price in: a steady demand signal for new EU renewable buildout to feed the gigafactories, anchor tenants for European AI labs such as Mistral, and a domestic order book large enough to attract component manufacturing the bloc currently lacks. A thin draw would leave the seven sites running on a single layer, compute without sovereignty, against the same U.S. and Chinese rivals the program names as the target.