EU to build AI gigafactories in €20bn push to catch up with US and China

 

EU to build AI gigafactories in €20bn push to catch up with US and China

 

Up to five sites with power-hungry supercomputers and datacentres planned to drive AI ‘moonshots’

 

Jennifer Rankin in Brussels

Wed 9 Apr 2025 09.40 EDT

https://www.theguardian.com/technology/2025/apr/09/eu-to-build-ai-gigafactories-20bn-push-catch-up-us-china

 

The EU has revealed details of a €20bn (£17bn) plan to create new sites equipped with vast supercomputers in Europe to develop the next generation of artificial intelligence models, while opening the door to amending its landmark law that regulates the technology.

 

Publishing a strategy to turn Europe into an “AI continent”, the European Commission vice-president Henna Virkkunen said the technology was at the heart of making Europe more competitive, secure and technologically sovereign, adding: “The global race for AI is far from over.”

 

The EU is attempting to catch up with the US and China, which have taken a lead in pioneering the technology that increasingly powers shopping websites and self-driving cars, generates text, and is predicted to play a transformative role in healthcare, security, defence and advanced manufacturing, among other sectors.

 

The US has a commanding lead in AI, far ahead of China. A report from Stanford University this week said 40 “notable AI models” – meaning influential – were produced by institutions in the US in 2024, compared with 15 in China and three in Europe (all French).

 

In a separate 2024 report, Stanford found that no EU country made the top five for “vibrancy” in AI, a metric that considered private investment, patents and research. It ranked the UK in third place behind the US and China, with France in sixth place and Germany eighth.

 

The EU has already embarked on a plan to build 13 AI factories – sites with supercomputers and datacentres, where researchers develop and test AI models.

 

The new AI “gigafactories” would be much larger, targeting what the commission called “moonshots”: significant innovations in healthcare, biotech, industry, robotics and scientific discovery.

 

While the best-performing AI factories have supercomputers equipped with up to 25,000 advanced AI processors, a gigafactory would exceed 100,000 AI processors, the strategy document said.

 

These power-hungry facilities, which can require huge amounts of water for cooling, should run “as much as possible” on a green energy supply, an EU official said, with plans for “recycling” water if it was used.

 

Last year 47% of electricity generated in the EU came from renewable sources, but campaigners fear that power-hungry datacentres could undermine Europe’s climate ambitions.

 

“Europe has made some strides towards fully renewable electricity, and still has a long way to go, but this could all be undone by ballooning energy consumption,” said John Hyland, a Greenpeace spokesperson. “Every power-guzzling datacentre risks being a lifeline to polluting power plants, well illustrated in Ireland where they consume over a fifth of the electricity.”

 

EU officials envisage three to five AI gigafactories in the EU at a cost of around €3-5bn each, compared with €600m for the biggest AI factory. The commission is seeking to raise €20bn, using public money to incentivise private investors, via a facility provided by the European Investment Bank. The bloc is also seeking to develop European-made AI semiconductors, advanced chips that are vital for AI factories.

 

But the commission faced criticism for indicating an openness to rethink its AI Act. The regulation, which aims to control the risks posed by AI, was approved by the European parliament only last year and will not be fully applicable until August 2027. On Wednesday, the commission said it was launching a consultation on “possible simplification” of the AI Act, as part of a wider drive to cut redtape amid concerns about Europe’s lagging economic growth.

 

The European Consumer Organisation (BEUC) said the commission needed to focus on the implementation and enforcement of the AI Act. “Much of the AI Act does not even apply yet but the commission is already signalling that it is open to cutting some of its requirements. When we know that consumers worry about how AI could be used to harm them, the commission should instead be trying to reassure consumers,” said Frederico Oliveira da Silva at BEUC.

 

Virkkunen said the commission was “very committed” to the risk-based approach and the main principles of the AI Act, but was looking at whether there were some administrative burdens or reporting obligations that could be cut to make life easier for businesses.