Intelligence boost: What will Europe’s AI factories deliver for R&I?

The rapid development of artificial intelligence (AI) has unleashed a flow of innovation that currently seems unstoppable – and as has been highlighted in last year’s Draghi report and elsewhere, the European Union is fighting to gain its spot in the sun in a hugely competitive market environment. As signalled in her February 11 announcement of the new InvestAI initiative – including its proposed four “AI gigafactories” – Commission President Ursula von der Leyen clearly understands that now is the time for Europe to scale up its efforts to become a global leader in AI innovation, not least given the ripple effects emanating from Washington DC under the second Trump administration.

Against this backdrop, one of the key first steps to strengthen the EU’s AI ecosystem is the launch of the AI Factories initiative, a €1.5 billion plan to harness the continent’s advanced computing facilities and their vast volumes of high quality, curated data to accelerate the development curve of Europe’s most innovative companies and entrepreneurial researchers. As of December 2024, seven institutions have been selected to establish the first generation of European AI Factories, with another five due to be announced in the weeks ahead. Building on a set of existing and planned supercomputers, the objective of the factories will be to open their doors to the R&I ecosystem around them, and enable end users to exploit the unique resources and capabilities of these cutting-edge infrastructures. 

Given its infancy, the initiative is surrounded by a number of strategic questions about the way forward, not least when it comes to money, focus and sustainability. To begin with, the estimated €1.5 billion fund to establish the seven factories will need to be leveraged into a major inflow of private capital to create adequate momentum and make Europe a leader in AI development. But with startups and companies historically looking at the US as a main destination for growth, a framework will be needed to give companies the conditions and incentives to grow and expand in Europe based on whatever outputs they gain from the AI factories. Clarity will also be required about how the factories will link in with other important EU initiatives, such as the AI testing and experimentation facilities (TEFs).

In parallel, will Europe look to prescribe tailored missions and objectives to its AI factories, in line with specific policy priorities?  Each host location has its own unique set-up within a given regional R&I ecosystem, and as such may not be suited to providing a full spectrum of services across disciplines and innovation areas. Given the ambitious timeline for implementation, a related concern is whether intended end users have the internal bandwidth, know-how and skills to work with such advanced facilities. And on top of the human factor comes the environmental one: increasing the size and capacity of supercomputing centres and their data infrastructure will imply a major increase in energy consumption and costs, which may rapidly absorb available budgets. 

On March 25, Science|Business will convene members of its international Network, EU institutions and other key stakeholders to discuss these questions and more, as a means to explore the opportunity landscape and potential barriers to success for Europe’s AI community.