The European Commission is raising $20 billion to construct four “AI gigafactories” as part of Europe’s strategy to catch up with the U.S. and China in artificial intelligence. However, some industry experts question whether developing these factories is a viable strategy.
The proposal for large public access data centres was unveiled by European Commission President Ursula von der Leyen last month, but it faces challenges such as securing hardware and finding suitable sites.
Bertin Martens from economic think tank Bruegel pointed out, ‘Even if we were to build this major computing facility in Europe, question arises: what will we do with it once it’s operational?’ He emphasized that there’s a dependency on local firms like France’s Nvidia-backed Mistral, which are expected to create AI models compliant with EU regulations that are stricter than those in the U.S. or China.
The absence of large European cloud service companies like Google and Amazon, or firms with expansive customer bases such as OpenAI, raises concerns about the sustainability of such an investment.
Europe’s Response to Competitiveness
The gigafactory initiative is part of Europe’s response to the Draghi report on competitiveness, which called for bold investments and a proactive industrial policy. This plan was detailed for the first time during the AI summit held in Paris on February 11, under the umbrella of InvestAI, Europe’s €200 billion investment strategy, intended to compete with the U.S. Stargate plan worth $500 billion.
Von der Leyen described these gigafactories as a ‘public-private partnership that will enable all our scientists and companies—not just the largest—to develop the advanced, large-scale models needed for Europe to thrive in AI.’ Funding will be sourced from a new €20 billion fund derived from existing EU initiatives and contributions from member states, alongside participation from the European Investment Bank.
According to von der Leyen, each gigafactory will house 100,000 advanced AI chips, making them significantly larger than the Jupiter supercomputer currently being built in Germany. This raises concerns about financial feasibility, as U.S. companies like Meta are investing heavily in their infrastructure.
Challenges Ahead
Kevin Restivo of the data centre consultancy CBRE highlighted potential roadblocks, reporting that the gigafactories would likely face the same difficulties as private projects, including a shortage of powerful AI chips and a lack of sufficient electricity supply.
Martens further criticized the plan, stating, ‘It does not make sense to invest public funds in an AI race.’ He noted that the lifespan of such factories would be limited to about a year and a half before needing substantial updates. Additionally, advancements like China’s Deepseek model provoke questions about whether AI can be trained efficiently with fewer resources, prompting considerations toward application-focused investments.
The European Commission’s prior technical infrastructure initiative, the 2023 Chips Act, fell short of meeting its objectives regarding cutting-edge chip production, although it did stimulate investments in automotive chip manufacturing. The Commission is also proposing upgrades to 12 scientific supercomputer centers, repurposing them as AI production facilities.
Kimmo Koski, managing director of Finland’s LUMI supercomputer, commented, ‘We are unsure how these AI gigafactories will differ outside of their larger scale.’ He mentioned the emerging industrial uses, indicating it could revitalize AI use in Europe.
The growth of European chip manufacturers such as Germany’s Infineon and France’s ST Microelectronics, alongside startups including SiPearl and AxeleraAI, may benefit from this ambitious expansion.
- 0 Comments
- AI Gigafactories
- European Commission