Subscribe
For years, Europe has tried to become the world’s leading regulatory power in artificial intelligence. On July 30, Brussels finally took a step towards becoming a technological power as well.
The European Commission has launched a call to establish up to seven AI Gigafactories across the continent. The initiative could mobilise around €10 billion in European and national funding, with the aim of attracting at least another €20 billion in private investment. The new facilities will combine advanced processors, cloud technologies, software, high-speed connectivity and data centres, allowing European companies and researchers to train the most sophisticated artificial intelligence models.
It is good news. But it is also an admission of how much time Europe has lost.
While the United States built global platforms and China developed an industrial ecosystem supported by enormous public and private investment, the European Union concentrated much of its energy on regulation. With the AI Act, Brussels proved that it knew how artificial intelligence should be governed before demonstrating that European companies were capable of developing and commercialising it at scale.
The gigafactories therefore mark a necessary change of direction. Europe is beginning to understand that technological sovereignty cannot be created through a directive. It must be supported by capital, infrastructure, energy, research and productive capacity.
In this sector, however, the answer cannot simply be a return to the nation state. No European country, acting alone, possesses the financial, industrial and computing scale required to compete with the United States and China. Not even Germany or France can independently build a complete artificial intelligence supply chain, finance the infrastructure required by the most advanced models and create a market comparable in size to the American or Chinese ones.
The European dimension is therefore not an obstacle to overcome, but a strategic necessity.
The real question is not whether Brussels should coordinate investment. It is what kind of Europe should do so. The choice is between a Europe capable of acting as a power, concentrating resources in decisive sectors, and a Europe that behaves merely as an administration, centralising rules without centralising economic strength or industrial capacity.
The gigafactories could represent a step towards the first model. Computing capacity, semiconductor supply, advanced research, technological security and capital markets all require continental scale. These are areas in which fragmentation across 27 national strategies would condemn Europe to irrelevance.
Centralising resources, however, must not mean planning technological development entirely from above. The Commission should create the conditions in which companies, universities and investors can compete, rather than bureaucratically deciding which technologies or industrial models should prevail.
The new facilities must be located where energy, networks, scientific expertise and companies capable of using them already exist. If projects are instead selected according to geographical balance, political distribution of funds or compensation between governments, the gigafactories risk becoming expensive monuments to European inefficiency.
There is also a contradiction that Brussels cannot ignore. AMD, Nvidia and Qualcomm have expressed interest in supplying the chips required by the new centres. All three are American companies. Europe may therefore host the infrastructure while continuing to depend on foreign suppliers for its most important components.
This does not make the project pointless. It does, however, demonstrate that technological sovereignty cannot be reduced to the geographical location of a server. It requires a broader strategy covering semiconductors, cloud services, software, capital and workforce training.
The second major challenge is energy. Data centres dedicated to artificial intelligence require enormous quantities of stable and competitively priced electricity. Europe cannot declare AI to be a strategic priority while maintaining energy costs that discourage industrial and digital investment.
A continent that wants to host the next generation of supercomputers must invest in electricity grids, expand generation capacity and abandon ideological prejudices against particular energy sources, including nuclear power. Without energy realism, Europe’s technological ambitions will remain confined to Commission press releases.
Computing capacity alone will not be enough either. Only a limited share of European companies currently uses artificial intelligence. Europe needs businesses capable of turning these infrastructures into products, industrial applications and competitive services. It needs deeper capital markets and fewer barriers to the growth of start-ups, which are too often acquired by foreign groups before they can become European champions.
The Commission’s initiative therefore deserves support. Not because Brussels should replace national governments or the market, but because in certain sectors Europe must finally learn to operate at the scale of a power.
The principle should be simple: more Europe where scale, negotiating strength and strategic investment are necessary; less Europe where centralisation produces only new obligations and procedures.
The seven gigafactories could mark the beginning of a more mature approach: a Europe united in building its technological strength, but sufficiently free to allow companies, researchers and nations to innovate along different paths.
To compete with Washington and Beijing, Europe does not have to choose between continental cooperation and economic freedom. It must learn to combine both.
Technological sovereignty cannot be regulated into existence. But nor can it be built by 27 states proceeding separately.
It must be built together, without being suffocated by bureaucracy.