Advertisement
The Gulf conflict has disrupted global supply chains, pushed...
REITs have changed the way commercial real estate is owned a...
What does it take to preserve a real estate legacy while bui...
What really powers the cloud? Behind every Google search, A...
A lot of what defines a home isn’t visible at handover. I...
European data-centre developers are increasingly looking beyond major cities as the rapid growth of artificial intelligence drives demand for large amounts of power, land and cooling capacity. JLL data shared with Reuters shows that hyperscale data centres planned for 2026-2028 will be located an average of 175 km from major urban hubs, compared with 46 km for projects delivered between 2022 and 2025. Greenfield sites account for 39% of the future pipeline, as developers seek lower land costs and faster power connections while facing shortages, planning restrictions and rising infrastructure costs in established markets.
European developers of artificial intelligence data centres are increasingly moving away from major cities to locations where land and electricity are cheaper and grid connections can be secured faster.
Data from JLL shared with Reuters shows that hyperscale data centres expected to come online between 2026 and 2028 will be located an average of 175 km from major urban hubs. This compares with an average distance of 46 km for projects delivered between 2022 and 2025.
The shift is being driven largely by AI training facilities, which require substantial amounts of electricity and water for cooling. As land and available power become increasingly limited in established markets such as London and Frankfurt, developers are looking at industrial, edge-of-city and rural locations.
Greenfield projects account for 39% of Europe's future data-centre pipeline, compared with 8% of projects already delivered. At the same time, the share of pipeline projects located in inner-city areas is expected to fall to 5% from 13%.
JLL's head of data centres for Europe, the Middle East and Africa, Assad Noori, said the availability of sufficient power was increasingly determining where data centres are developed, rather than simply locating facilities close to demand.
JLL estimates that the world's four largest hyperscale cloud providers will spend USD 725 billion in 2026, up from USD 410 billion in 2025. Most of this spending is expected to be directed towards AI computing and data-centre infrastructure. The company also estimates that AI workloads could account for around half of global data-centre capacity by 2030.
The trend is also visible in proposed large-scale projects across Europe. DC Byte data cited by Reuters shows that of nine proposed data centres with capacity of more than one gigawatt, only one is planned near a major city, Paris. The others are spread across locations ranging from rural Spain to northern Sweden.
Major markets including Frankfurt, London, Amsterdam, Paris and Dublin continue to attract demand, particularly from enterprises. However, these markets are facing limited land availability, planning restrictions and longer waits for grid connections.
JLL's president of EMEA data centres, Martin Jensen, said the established markets would remain important because enterprise demand was expected to continue, but AI infrastructure required a different scale of power and land.
The cost difference between locations is also becoming an important factor. JLL estimates that powered land costs an average of EUR 2.36 million per megawatt of IT load in Europe's core data-centre markets. The average falls to EUR 978,000 in secondary cities such as Copenhagen, Warsaw and Milan, while tertiary locations such as Bordeaux can have costs of around EUR 512,000 per megawatt, with some sites available for as little as EUR 200,000.
Amsterdam is currently the most expensive market at about EUR 2.7 million per megawatt, followed by London at EUR 2.6 million and Frankfurt at EUR 2.5 million.
Rupert Duckworth, associate director for EMEA data-centre advisory at Savills, said London's earlier growth in cloud infrastructure had increased competition for available land from other asset classes, pushing up land prices. He also pointed to power constraints in key cloud locations across the market.
The movement towards less traditional locations could bring investment and employment opportunities to areas outside Europe's established technology and commercial centres. However, large data-centre projects can also face resistance from local communities over concerns about land use, natural habitats, electricity availability and water consumption.
For Europe's real estate and infrastructure markets, the expansion of AI data centres is therefore creating demand for sites where land, power and connectivity can be secured together. This is likely to make access to reliable electricity an increasingly important factor in determining the next generation of large data-centre locations.
Source PTI
Advertisement
Facebook
Whatsapp
