Europe AI data centres seek cheaper, quicker energy and land

Data centres for artificial intelligence across Europe are moving away from major cities to secure cheaper energy, land and quicker connections, according to data analysis by Reuters. This trend could lead to investment in underdeveloped areas, but developers may face opposition from residents worried about environmental impacts and limited resources.

JLL, a leading commercial real estate and services company, reports on plans for hyperscale data centres that will be located up to three times further from cities than previous projects. These centres will not need to be as close to end-users, allowing them to be built in less populated areas. The average distance of these new sites from a major hub will be around 175 kilometres, compared to 46 kilometres for projects built between 2022 and 2025.

Greenfield projects, which involve developing new land, account for 39% of the future pipeline for Europe's data centres, up from just 8% in previously built projects. This shift is largely driven by the needs of AI training facilities, which require significant amounts of electricity and water for cooling. European data centre developers estimate that the world's four largest hyperscale cloud providers will invest $725 billion in 2026, a 77% increase from $410 billion in 2025, primarily for AI computing and data centre infrastructure.

While the core markets of Frankfurt, London, Amsterdam, Paris and Dublin remain the largest and most in-demand, they are increasingly facing land shortages, planning restrictions and delays in connecting to the grid.