Europe’s artificial intelligence efforts are hitting an electrifying obstacle. As the region competes to attract the most promising models, data centers and computing infrastructure, another vital resource is power, and the ability to deliver it reliably at scale.
The International Energy Agency (IAE) estimates that data centers consumed 415 terawatt-hours of electricity, about 1.5% of global electricity consumption, in 2024. It has grown at 12% per year over the last five years.
Worldwide, data-center electricity demand is expected to roughly double by 2030 to 945 TWh, with AI workloads driving much of the increase, per the IAE.
The urgency of the challenge is compounded by the fact that data centers can typically be built two to three times faster than the electricity networks that feed them. And the EU, to compete in this paradigm, wants to boost its computing capacity – while staying true to its sustainability pillars.
The European Commission recognizes the need for data center capacity “as digitalisation accelerates”. And it aims to triple its data center capacity by 2035 via energy-efficient and sustainable frameworks.
Europe’s AI expansion meets the grid
The EU’s proposed Cloud and AI Development Act would at least triple EU data-center capacity within five to seven years while specifically citing energy availability, land-use restrictions, financing and permitting as constraints on expansion.
ENTSO-E, the association representing Europe’s transmission system operators, has projected that electricity demand from data centers in its jurisdiction could rise by more than 50% from 2025 to 2030, increasing from about 87 TWh to 134 TWh.
It warns that data centers are becoming key electricity consumers and that their growth should be factored into electricity planning, due to their size and concentration and the potential for demand-side flexibility.
Data centers accounted for 23% of metered electricity consumption in Ireland alone in 2025, up from 5% in 2015, according to the Central Statistics Office of Ireland.
But that does not mean Europe is running out of electricity. Rather, it means the location of new AI infrastructure increasingly matters.
The IEA recommends prioritizing data centers in areas with spare grid capacity, improving grid-connection queues and using onsite batteries and other assets to reduce pressure during constrained periods.
Finland offers a glimpse of the next AI map
Google’s latest Finnish investment makes the connection between computing and electricity unusually explicit.
The company announced a €13 billion investment in AI infrastructure in Finland, including three new data centers and a 22-year agreement to purchase up to 50% of the output of Finnish energy company Fortum’s Loviisa nuclear plant, located in the country’s southern shore, that houses two Soviet-designed pressurized water reactors.
The project also includes grid, renewable and battery-storage upgrades.
Finland’s abundant nuclear and renewable electricity, cold climate and existing digital infrastructure make it a good fit for energy-intensive data centers.
The deal has raised political concerns about electricity prices in Finland, reflecting the tension between the goals of attracting technology and energy-intensive industries and maintaining affordable and reliable electricity for all consumers.
Nuclear, renewables, or both?
There is unlikely to be a single power source for Europe’s AI buildout. The IEA projects that renewables and nuclear power will supply about 85% of the additional electricity needed for data centers in Europe by 2030.
Nuclear power offers reliable power in Europe, while wind, solar and hydropower can supply large amounts of flexible, low-carbon electricity if strengthened by the right transmission, storage and system adequacy measures.
That mix also exposes Europe’s central challenge: generation alone is not enough. Power has to be available where data centers are being built and delivered through a grid capable of handling large, concentrated loads.
The European Commission has already acknowledged that limited energy access and long permitting procedures are holding back cloud and AI infrastructure. Its August 2026 study also identified geographically concentrated computing capacity and energy-grid constraints among the major obstacles to Europe’s AI expansion.
Electricity is becoming part of Europe’s AI strategy
This context changes the question Europe should be asking. It is no longer simply whether Europe can build more AI capacity; it is whether Europe can build that capacity where affordable, reliable electricity is available.
That could reshape the region’s technology map. Countries with strong nuclear fleets, abundant renewables, available grid capacity and faster permitting may become more attractive destinations for AI infrastructure than traditional technology hubs.
So, the next major AI hub may not be determined only by talent, capital or chips. It may be determined by megawatts.
Featured image: Alex Shuper via Unsplash+
