Europe's AI Data Centres Move 175km From Cities as Power Becomes the Deciding Factor

Data centres planned for delivery in Europe between 2026 and 2028 will sit an average of 175 kilometres from major urban hubs, more than three times further than the 46-kilometre average for projects delivered between 2022 and 2025, according to data from JLL shared with Reuters. Greenfield projects now account for 39 percent of Europe's future data centre pipeline, up from just 8 percent of already-delivered projects, while the share of projects sited in inner-city locations is expected to fall to 5 percent from 13 percent.

 

Why Power Availability Has Overtaken Proximity to Demand as the Deciding Factor

 

JLL's head of data centres for Europe, the Middle East and Africa, Assad Noori, described the underlying shift directly: "the determining factor is increasingly where sufficient power can be secured, rather than simply where demand exists," summarising the change as data centres being "brought to where the power is, not the other way around." That reversal matters because data centres have traditionally been sited close to major population centres specifically to minimise latency for end users accessing cloud services, but AI training workloads, which don't require the same real-time responsiveness as consumer-facing cloud applications, don't need that same proximity, freeing developers to prioritise power availability over urban proximity entirely.

That shift aligns with power grid capacity constraints becoming an increasingly binding limitation on data centre development across multiple markets covered elsewhere in recent reporting, from Stella Power's onsite gas generation strategy explicitly designed to bypass grid interconnection queues, to Envision's fully self-contained renewable-powered AI campus in Inner Mongolia, both representing different responses to the same underlying problem JLL's data illustrates at a European market-wide scale: available grid power, not available land or proximity to customers, has become the binding constraint on where new AI infrastructure can actually be built.

 

UPCOMING WEBINAR

Moving Sustainable Procurement Beyond Cost to Strategic Value Creation

Online, 09 September 2026 · Balancing cost, managing supplier risk and building long-term value

 

What the Greenfield Shift Reveals About a Genuine Tension

 

The jump from 8 percent to 39 percent greenfield development, building on previously undeveloped land rather than converting existing industrial or urban sites, signals a meaningful change in how data centre developers are approaching site selection, one carrying a specific and acknowledged tension. Reuters notes this trend can mean investment flowing into underprivileged areas where governments are actively trying to stimulate jobs and economic growth, a genuine potential economic development benefit for regions that have historically attracted less large-scale infrastructure investment than established urban centres.

At the same time, the same reporting notes developers risk opposition from residents concerned about dwindling natural habitats and competition for power and water resources, precisely the kind of concern that has driven community opposition to data centre projects documented elsewhere, including the $156 billion in blocked or delayed US data centre projects tracked by Data Center Watch covered in earlier reporting this batch. That greenfield development inherently converts previously undeveloped land, land that may support existing habitat or agricultural use, into industrial infrastructure, meaning the economic development benefit for a given region comes paired with genuine environmental and community trade-offs that weren't as pronounced when data centres were built primarily on already-developed urban or industrial sites.

 

Why the Scale of Investment Makes This Shift Consequential

 

JLL estimates the world's four largest hyperscale cloud providers will spend $725 billion in 2026 on AI computing and data centre infrastructure, up 77 percent from $410 billion in 2025, with AI workloads projected to account for roughly half of global data centre capacity by 2030. That scale of capital deployment means the siting pattern JLL describes isn't a marginal shift affecting a small number of projects, but a structural change affecting where a substantial share of hundreds of billions of dollars in infrastructure investment will physically land over the coming years, with corresponding implications for which regions experience the associated economic activity, construction employment, and long-term tax revenue, alongside whichever environmental and resource trade-offs accompany that development.

Data from DC Byte, tracking early-stage projects, reinforces the geographic pattern concretely: of nine proposed gigawatt-plus capacity data centres across Europe, only one is planned near a major city, Paris, with the remainder spread across locations ranging from rural Spain to northern Sweden, illustrating that this shift extends to the very largest planned facilities, not only smaller or more marginal projects.

 

RELATED EVENT

Reuters Sustainability Europe 2026

London, 20–21 October 2026 · 400+ sustainability, finance and procurement leaders on regulatory uncertainty and Scope 3 data

 

Why Core Urban Markets Are Bifurcating Rather Than Simply Declining

 

Despite this broader relocation trend, JLL's Martin Jensen stated that "Europe's core markets will remain critical because enterprise demand isn't going anywhere," pointing to continued demand in established hubs including Frankfurt, London, Amsterdam, Paris and Dublin even as those markets face land shortages, planning restrictions and lengthy grid connection waiting times. That combination, continued underlying demand paired with genuine physical and regulatory capacity constraints, suggests Europe's data centre market is bifurcating rather than uniformly shifting away from established hubs: AI training infrastructure requiring vast power and land is increasingly moving to less traditional, power-rich locations, while latency-sensitive enterprise and cloud services demand continues concentrating in the established urban markets that can still support it despite growing constraints.

Whether the growing pipeline of rural and edge-of-city AI training facilities delivers on its stated potential to bring genuine economic development to underprivileged regions without triggering the kind of sustained community opposition that has already stalled billions of dollars in data centre projects elsewhere, and whether Europe's established urban data centre hubs manage to resolve their land and grid connection constraints sufficiently to retain their continued enterprise demand, will determine how this bifurcated development pattern ultimately shapes the distribution of AI infrastructure investment across the continent over the coming years.