Europe’s slow and steady approach to AI could be its edge

Europe is often viewed as lagging behind the U.S. and China in the race to scale artificial intelligence (AI) technology, primarily due to its fragmented markets. However, this very fragmentation might offer Europe a unique advantage in establishing robust infrastructure necessary for the AI revolution. Pankaj Sachdeva, a senior partner at McKinsey, highlighted that the world is on the brink of a massive expansion of data center capacity, with an estimated investment of up to $7 trillion required by 2030. While the U.S. is projected to dominate this expansion, Europe is expected to significantly grow its data center capacity, nearly doubling what it currently has. Despite these promising projections, Europe faces substantial challenges, especially concerning energy access and regulatory hurdles. According to experts, electricity availability is a critical bottleneck. Regions like the Nordics and Spain are becoming increasingly attractive for data center development due to their renewable energy surplus, whereas Germany and the U.K. are grappling with energy supply constraints that could deter investment. Interestingly, Italy is emerging as a favorable location due to its relatively shorter grid connection times compared to the European average. However, countries like Germany, the U.K., Ireland, and the Netherlands are struggling with grid capacity, leading to a moratorium on new connections in some areas. While the differences between European countries are pronounced, experts believe that Europe will find it difficult to catch up with the U.S. in the short term, where deregulation and extensive investment allow for rapid infrastructure development. Currently, the U.S. boasts approximately 5,400 data centers, compared to just 200 to 300 in most European countries. This disparity is prompting a shift in investment away from traditional markets like Frankfurt, London, Amsterdam, Paris, and Dublin, towards areas where resources are more stable. In the U.K., for instance, the government has begun overriding local decisions to expedite the approval process for data centers, recognizing their critical role in the economy. The International Energy Agency predicts that energy consumption from data centers could surge to 1,000 terawatt-hours (TWh) by 2026, largely driven by AI demands. As electricity remains the largest operational cost for data centers, Europe's high energy prices, exacerbated by geopolitical tensions, pose a significant challenge. To tackle issues related to grid congestion, reforms are underway in the U.K. that would prioritize project readiness in power connection queues, thereby allowing completed projects to gain faster access to energy resources. These changes are indicative of an evolving landscape where energy procurement is being re-evaluated to accommodate the growing demand for data centers. While Europe may not lead in developing facilities for large-scale AI training, it holds promise in creating smaller, cloud-oriented facilities that focus on AI inference—an area projected to account for 70% of AI demand. Investors are increasingly cautious, focusing on securing long-term customers before breaking ground on new projects. This cautious approach aims to avoid the pitfalls of speculative developments that have plagued the past. While European regulations mandate transparency in energy and water usage, they may ultimately serve to integrate data centers into local communities, fostering sustainable practices. In summary, Europe's deliberate pace in expanding its AI infrastructure may not only mitigate risks associated with speculative investments but also promote innovative solutions tailored to meet the evolving needs of the AI landscape. As Europe navigates these complexities, the potential for long-term value creation for both investors and society remains significant.