High energy prices could derail Europe’s AI race with U.S. and China

As Europe strives to establish itself as a frontrunner in artificial intelligence, experts warn that skyrocketing energy prices could hinder these aspirations. The continent is eager to advance in the AI sector by enhancing computing capacity and developing the essential infrastructure for this technology. However, the energy-intensive nature of data centers makes them especially vulnerable to rising costs, which have surged in light of geopolitical tensions, notably the U.S.-Iran conflict. Analysts predict that data center projects may begin relocating to regions within Europe that offer more affordable energy. This shift could create significant disparities across the continent. Michael Brown, a global investment strategist at Franklin Templeton, emphasized, "The differences in energy costs globally are becoming increasingly pronounced. For energy-intensive investments, the trend will be to seek out the cheapest energy sources. If I were to invest in a $7 billion data center, my choice would likely be the U.S. or China." The recent Iran crisis has sparked a renewed focus on electrifying the economy, according to Olivier Darmouni, an associate professor at HEC Paris who specializes in energy transition. He pointed out that the rapid expansion of data centers could drive regional electricity prices up by 20-40% in areas like Texas and Virginia in the U.S. and in key locations in the U.K. and France. Darmouni emphasized the need to view the energy system as integral to economic sovereignty, stating, "We cannot achieve affordability, competitiveness for European companies, and technological leadership in AI without addressing our energy system." According to the International Energy Agency, energy prices for industries in Europe were, on average, about double those in the U.S. last year, and 50% higher than in China and India. Data centers currently account for 2% of global electricity consumption, a rise from 1.7% in 2024, as reported by the International Data Center Authority. The report also noted that community and political resistance against data centers tends to increase when their electricity usage exceeds 5% of a nation’s total consumption, a threshold the U.S. is nearing at 6%. Chris Seiple, vice chairman of Wood Mackenzie’s power and renewables division, identified three primary reasons for Europe’s lag in data center development: high energy costs, the geographic positioning of data center developers, and the time required to establish the necessary infrastructure. "These factors make it more challenging to develop data centers in Europe," he remarked. While Europe aims to enhance its computing capacity and expand data center facilities, Darmouni cautioned about the continent's commitment to achieving technological leadership in AI. He noted, "The scale of U.S. data center investments compared to Europe is staggering—it's like comparing 1 to 100. If Europe wants to catch up, it will require substantial investment." Vladimir Prodanovic, a principal program manager at Nvidia, highlighted the struggles faced by central European nations like Germany and the U.K. due to high electricity costs. For instance, in May, the average electricity price per megawatt in the U.K. was $111.65, compared to $28 in the U.S. Despite plans to increase data center capacity, OpenAI recently paused its Stargate project in the U.K., citing energy costs and regulatory challenges as significant factors. Darmouni predicts that AI models may eventually factor energy costs into their pricing structures, potentially leading to price discrimination for AI services based on energy expenses. In contrast, the Nordics and France are seen as prime candidates for AI investment due to their lower electricity prices and diverse energy sources. For instance, Nvidia's Prodanovic pointed out that Norway is currently a favored location for major AI companies, with Microsoft investing heavily in the region. As Darmouni noted, France enjoys a significant advantage in energy pricing, largely due to its robust nuclear energy sector. However, he stressed that lower electricity prices alone are not sufficient; a country’s ambition to develop new power sources is equally critical. He advocated for greater integration of energy systems across national borders to stabilize energy prices throughout Europe, despite the geographical challenges faced by regions like the U.K. and parts of Scandinavia. Finally, the cost of securing data center capacity in Europe’s top five markets—Frankfurt, London, Amsterdam, Paris, and Dublin—is projected to increase by 12% by 2026, according to research from real estate investment firm CBRE. The overall landscape for AI in Europe remains uncertain as energy costs continue to rise.