How data centres can chase renewable energy across Europe

In 2022, demand for electricity from data centres amounted to 460 terawatt hours, representing 2% of the world’s total electricity demand. This number is climbing fast, because of energy consumption related to artificial intelligence. The situation is especially acute in some parts of Europe – in Ireland, the share of data centres in electricity consumption could increase from an already massive 21% in 2023 to 32% in 2026. In Denmark, electricity use by data centres could increase sixfold by 2030. Hyperscale data centres – large, highly efficient facilities with hundreds of server nodes that are geographically dispersed and managed collectively – drive a staggering growth in computing power demand. The carbon footprint of data centres is becoming a significant concern. 

Electricity grids can move clean electricity from the best wind and solar energy generation sites to demand centres and help balance varied electricity generation across different areas. But what if the source of demand could be relocated to the site of renewable electricity generation instead? Few demand sources are this flexible, but data centres are in some ways unique: many computation jobs, such as offline data processing, aren’t time- or location-critical and can be scheduled flexibly. 

There is a big demand for new data-centre infrastructure that has not been sited yet. Leveraging the flexibility of data centres can deliver many of the benefits associated with grid expansion, but without the challenges of NIMBYism and the long waiting times to build new power lines. Transporting data over long distances is cheaper than transporting the amount of electricity that it would have taken to process this data at the destination. By shifting this computational load between data-centre locations and from one time period to another, computing can take place where and when green electricity is available, enabling significant savings in both energy costs and emissions.

Companies can leverage three geography- and weather-dependent features to effectively harness this flexibility. The first feature is geographic variation in the capacity of local renewables, ie the volume of electricity a certain solar or wind installation can produce in an average year. Renewable energy availability varies significantly across Europe: much more sun in southern regions, while wind generally peaks in northern countries. By leveraging flexibility, maximum use can be made of the renewable-energy resources available across Europe.

The second feature relates to significant variance in weather patterns across Europe. For example, the wind generation correlation between Denmark and Portugal is almost zero. This means that when wind output is low in one location, it might be high in another. A distance of 300-400 kilometres is sufficient for two wind turbines to have very different generation patterns. Data centres can exploit this by dynamically shifting computing tasks to locations with greater wind-power availability, thereby reducing reliance on grid-based electricity or on energy storage during periods of low local wind generation.

The third feature involves time lags in solar radiation peaks caused by Earth’s rotation. Solar generation peaks in eastern and western Europe occur several hours apart, allowing computing workloads to align with these temporal shifts and thereby cut the costs associated with expensive energy storage. For example, a data centre in Greece might process computations required in Portugal during one part of the day, while the Portuguese data centre handles tasks for Greece a few hours later. This synchronisation effectively ensures that loads follow the sun.

Leveraging flexibility can mitigate one of the most significant challenges related to renewable energy: constrained grid infrastructure. Expanding Europe’s electricity grid to support increased renewable-energy generation demands substantial investment in power transmission lines. For example, it has been estimated that up to 76 gigawatts of new demand in the United States – equivalent to 10% of its peak electricity load – could be integrated with minimal reduction in production if the load is flexible and can be temporarily reduced during system peaks.

Similar system efficiency gains could be made in Europe, reducing the need for costly grid reinforcements while accelerating the integration of renewable energy. Transferring data loads to different locations presents a big opportunity, offering a complementary solution to grid reinforcement by enabling virtual transfers of power loads, effectively bypassing physical constraints in electricity transmission.

Supportive policies and economic levers that incentivise data centres to dynamically adjust their demand can help integrate this load-shifting potential into energy markets. Regulators, transmission system operators and industry should work together to foster an environment in which flexible demand from data centres becomes a significant component of the energy system, helping to stabilise electricity prices, optimise renewable energy use and accelerate decarbonisation.