EU plans AI gigafactories: 200 billion euros to catch up
The EU Commission presented a new AI initiative in Paris. A total of 200 billion euros is to be mobilized, including for AI gigafactories.
At the AI summit in Paris on Tuesday, EU Commission President Ursula von der Leyen presented the new AI initiative in Europe
(Image: EU)
EU Commission President Ursula von der Leyen presented a new AI initiative called “InvestAI” at the AI Action Summit in Paris. The aim is to invest a total of 200 billion euros in European AI development. This is intended to create the necessary hardware infrastructure.
The core of the initiative is a new European fund with a volume of 20 billion euros, which is to finance the establishment of four AI Gigafactories. These facilities will each be equipped with around 100,000 of the latest generation of AI chips – around four times more than in current AI factories. The gigafactories will specialize in training particularly complex AI models, for example for applications in medicine and science.
Largest public-private partnership
“We want AI to be a force for good and for growth,” emphasized von der Leyen. The European approach is based on openness and cooperation. The project is based on the CERN model and is to be established as the largest public-private partnership for trustworthy AI worldwide.
Financing will be provided by a multi-layered fund. The initial funding comes from existing EU programs such as Digital Europe and Horizon Europe. Member states can also participate. Of the planned 200 billion euros, 150 billion will come from providers, investors, and industry.
Where Europe sees its opportunity
The EU already announced seven AI factories in December, including one in Germany. Five more are to follow shortly. Von der Leyen rejected the accusation that Europe is being left behind. Instead, competition is only just beginning and Europe is well positioned with its AI start-up scene. Europe's opportunity lies above all in the industry-specific application of AI.
Videos by heise
Investments in the hardware infrastructure should be flanked by education and training programs, as well as the development of common European data spaces. Regulation through the AI Act should not be an obstacle, but should instead guarantee security.
(mki)