The European Union is deploying a network of supercomputers to optimize artificial intelligence (AI) models across the continent, aiming to narrow the gap in this field with global rivals, as part of a multi-billion dollar push to regain technological competitiveness.
Thirteen "AI factories" - facilities enabling startups and smaller firms to train complex AI models - will begin operations next year, including one at Vienna Technical University.
"You can add many many computers over a network to get a supercomputer. Here we can use all the CPU and all the RAM of all the computers together for a single task," said Valentin Hirschbrich, IT infrastructure staff member at Vienna Technical University.
After U.S. President Donald Trump announced 500 billion dollars in private sector investment in AI infrastructure, the EU followed suit. The bloc has committed about two billion U.S. dollars to the initiative so far, with plans to build four larger "gigafactories" to further scale capacity and put Europe back on the international AI track.
"The boost that you can expect is - from our point of view - huge, because there's a lot of demand when it comes to training or fine-tuning of AI models," said Markus Stöhr, project manager at AI Factory Austria.
But some experts question its efficacy amid Europe's lack of homegrown cloud service giants.
"We do have American cloud providers that run local installations in Europe, but obviously it would be nice if we would also have European companies. I think for the future, that we are competitive in Europe, the importance is attracting and keeping talented people in Europe," said Edgar Weippl, Informatics professor at University of Vienna.
The EU-funded AI factory in Vienna is set to focus on trustworthy artificial intelligence, for instance AI models that run on public data and do not violate the privacy of individuals.
"You could go to cloud providers but then you always have this issue with data privacy. But we have an offer so we ensure that the data is here located in Vienna," Stöhr said.
Critics warn of risks in publicly funding hardware that may quickly become obsolete. Some argue software development offers better cost efficiency, but EU officials see the supercomputers as foundational for attracting AI investment.
EU launches supercomputer network to optimize AI models
