If Europe builds the gigafactories, will an AI industry come?

Whether Europe’s construction of AI gigafactories will lead to a thriving AI industry is uncertain and depends on several interlocking factors, not just the presence of advanced infrastructure. The European Commission, under Ursula von der Leyen, has launched the InvestAI initiative, committing €20 billion to build four gigafactories, each equipped with 100,000 cutting-edge chips. This is part of a broader €200 billion public-private partnership aimed at making Europe an “AI continent” by providing scientists and companies with the computational power to develop advanced AI models. The plan responds to the Draghi report’s call for bold investments to boost competitiveness, positioning Europe against the U.S. (with its $500 billion Stargate plan) and China’s state-driven AI efforts.

On one hand, the gigafactories could lay the groundwork for an AI industry. They promise unparalleled access to computing resources, which could empower startups like France’s Mistral or Finland’s Silo AI (acquired by AMD for $665 million) to scale innovations. Kimmo Koski of Finland’s LUMI supercomputer sees potential in “pushing industry use further,” suggesting these facilities could bridge scientific research (e.g., climate modeling) and commercial applications. Europe’s strict AI safety and data protection rules could also attract firms seeking “trustworthy AI,” differentiating it from less-regulated markets like the U.S.

However, challenges cast doubt on this vision. Experts like Bertin Martens of Bruegel highlight a “chicken and egg problem”: even if the factories are built and models trained, what happens next? Without a clear ecosystem—demand, skilled talent, and downstream applications—the investment risks becoming a white elephant. Securing enough Nvidia chips is a hurdle, given U.S. export restrictions under Biden (whose future under Trump is unclear), and Europe’s energy grid may struggle to power these facilities, as noted by CBRE’s Kevin Restivo. The 2023 Chips Act’s failure to deliver cutting-edge chip manufacturing suggests execution could falter again. Meanwhile, U.S. projects like Meta’s $10 billion, 1.3 million-GPU facility dwarf Europe’s scale, raising questions about competitiveness.

The counterargument is that Europe’s collaborative approach—pooling resources across member states and involving the European Investment Bank—might foster a unique niche. Local chipmakers like Infineon or startups like SiPearl could benefit, and sectors like healthcare or manufacturing might leverage federated data for breakthroughs. Yet, skeptics argue that €20 billion pales against global spending, and without a robust talent pipeline or market pull, the AI industry might not materialize as hoped.

In short, building gigafactories could catalyze an AI industry if Europe overcomes supply chain, energy, and utilization challenges while capitalizing on its regulatory edge. But there’s no guarantee—success hinges on execution and demand, not just infrastructure. The jury’s still out.