Saturday, 5 September 2026
England’s data-centre market is undergoing a geographical change that could have significant consequences for industrial and development land. For years, the sector has been concentrated overwhelmingly around London, particularly Slough and the western edge of the capital. Artificial intelligence, rapidly increasing computing requirements and severe electricity constraints are now beginning to challenge that model.
London remains by far the country’s dominant data-centre market and there is little evidence that this position is about to disappear. The capital accounts for more than four fifths of existing UK data-centre capacity according to industry estimates, supported by extensive fibre infrastructure, established cloud networks and a large concentration of corporate customers. What is changing is the ability to add very large amounts of capacity in the locations where operators traditionally wanted it.
Modern data centres can require extraordinary amounts of electricity. AI facilities in particular are being planned at a scale that would have appeared exceptional only a few years ago. Projects requiring 100MW or more are becoming increasingly common, while some proposed campuses are considering several hundred megawatts or even gigawatt-scale requirements. Finding the land is rarely the greatest difficulty. Finding the electricity can be.
That distinction is beginning to transform property investment decisions. A conventional industrial development starts with questions about location, road connections, labour, occupier demand and land price. For a large data centre, one of the first questions is increasingly whether enough electricity can realistically reach the site within an acceptable timeframe. This is creating what amounts to a new category of real estate: land with credible access to large quantities of power.
The difference between such a site and an ordinary development plot can be substantial. Market research indicates that locations capable of securing 50MW or more within a relatively short period can, under the right circumstances, command multiples of conventional industrial land values. Across Europe’s major data-centre markets, the cost associated with powered development land has also increased sharply during the past five years. The premium does not simply reflect electricity consumption. It reflects scarcity.
Grid connections can take many years to secure, particularly around established data-centre clusters. Landowners that can demonstrate a realistic route to substantial power therefore possess something competitors may be unable to replicate regardless of how much land they own. This is especially visible around Slough. The area remains one of Europe’s most important data-centre concentrations and continues to attract development despite pressure on electricity infrastructure. Sites that do have a credible path to power can become more valuable precisely because so few alternatives exist.
The investment logic is consequently changing. Electricity constraints are not necessarily destroying the value of established data-centre locations. In some cases they are increasing the value of the limited sites that can still accommodate additional capacity. At the same time, the shortage is widening the industry’s search radius.
Oxfordshire is emerging as one of the clearest examples. Culham has been selected as part of the UK’s strategy to create large-scale artificial intelligence infrastructure, while the wider Thames Valley provides an established technology ecosystem with connections back towards London. The government’s criteria for new AI development locations demonstrate how dramatically site-selection priorities have changed. Potential locations are expected to show how very substantial electricity capacity could be delivered, with hundreds of megawatts potentially required over the coming years.
Reading provides another illustration. Proposals around Thames Valley Park have explored alternative on-site electricity generation while waiting for greater grid capacity. The significance for property investors extends beyond the individual development. If developers are prepared to incorporate substantial energy infrastructure simply to make a location viable, access to electricity has clearly become part of the underlying real-estate value.
The shift becomes even more interesting further north. Large AI computing facilities do not necessarily need to be located beside London’s established financial and corporate districts. Some computing tasks require extremely fast connections to users, but the training of large artificial intelligence models can tolerate greater geographical distance. That gives locations hundreds of kilometres from London an opportunity that would have been considerably harder to justify during earlier generations of data-centre development.
Leeds provides one of the strongest examples. At Skelton Grange, Microsoft is progressing plans for a major hyperscale campus on former industrial land. The company previously acquired approximately 27 acres for more than £50 million, while the wider redevelopment includes remediation, infrastructure and additional commercial development. The transaction demonstrates the potential impact on land economics. Former industrial property in a regional market can achieve a very different valuation when it becomes suitable for one of the world’s largest technology companies and has the infrastructure needed to support a major data-centre operation.
Northumberland provides an even larger example. At Cambois, near Blyth, plans are progressing for a multibillion-pound data-centre campus on land associated with former heavy industrial and energy uses. The scale of the proposed development places it among the most significant digital-infrastructure projects being pursued in Britain.
The location illustrates why England’s industrial past may become an advantage in the AI economy. Former power stations, steelworks, chemical complexes and other energy-intensive industrial locations frequently possess characteristics that are increasingly difficult to create from scratch. They can offer large areas of contiguous land, connections to high-voltage electricity networks, existing substations and infrastructure designed for industries that once consumed enormous amounts of energy. As traditional heavy industry disappeared from some of these locations, much of the underlying infrastructure remained. AI may now give that infrastructure a second economic life.
Teesside demonstrates similar potential. Large-scale data-centre proposals around established industrial areas have considered capacity measured in hundreds of megawatts, with some ambitions reaching approximately 1GW. The attraction again comes from the combination of substantial sites and an energy network originally developed to support power-intensive industry.
This creates an important new dimension for brownfield investment. A former industrial site can no longer be assessed solely according to its potential for warehouses, manufacturing or redevelopment. Investors increasingly need to understand what electrical infrastructure exists nearby, how much capacity could realistically be obtained and when that electricity could be delivered. In some cases, the value hidden beneath the site may be more important than the buildings standing on it.
The Midlands could eventually benefit from the same trend. Its central location, manufacturing heritage, extensive logistics infrastructure and comparatively lower land costs provide many of the characteristics data-centre developers require. However, evidence of a broad Midlands data-centre cluster remains less developed than the projects emerging around Oxfordshire, Leeds, Teesside and Northumberland. The regional opportunity should therefore not be interpreted as a uniform migration away from London. England is more likely to develop several different data-centre markets serving different requirements.
London and its surrounding areas can remain dominant for cloud computing, financial services and applications where connectivity and proximity to established digital infrastructure are critical. Large AI campuses, by contrast, can increasingly follow electricity. This distinction could fundamentally alter the geography of technology property investment. Locations that were once considered too distant from London may become viable if they can offer substantial power earlier and more reliably than sites in the South East.
However, the growing value attached to electricity also creates significant investment risk. Britain has accumulated a large pipeline of proposed data centres, but not every project with a grid application will ultimately be built. A requested connection is not the same as available power, just as a development concept is not equivalent to planning permission.
Investors therefore need to understand precisely what sits behind claims that a site is powered. There is a major difference between submitting a connection request, receiving an offer, funding the necessary network reinforcement and obtaining electricity on a commercially usable date. Those stages can represent very different levels of development certainty and therefore very different land values.
This is likely to become increasingly important as grid authorities attempt to distinguish credible projects from speculative applications. Large amounts of proposed electricity demand can otherwise occupy positions within connection queues even when the underlying developments have little prospect of proceeding.
For property investors, due diligence consequently needs to extend beyond conventional title, planning, contamination and construction analysis. Grid capacity, connection agreements, reinforcement requirements, delivery dates and the financial obligations attached to electricity infrastructure are becoming fundamental components of development value.
Government policy is reinforcing the change. Britain’s programme for large AI development zones is attempting to bring together land, planning, electricity and technology investment in locations capable of supporting substantial computing capacity. Yet designation alone will not create successful data-centre markets. The locations that ultimately attract investment will be those capable of turning political ambition and theoretical electricity capacity into infrastructure that operators can actually use.
This is why England’s emerging data-centre geography should be viewed as a property investment map as much as a technology map. Slough demonstrates the premium attached to scarce power inside an established cluster. Oxfordshire and the Thames Valley illustrate the expansion of the London technology ecosystem into locations with greater development potential. Leeds shows how former industrial land can attract hyperscale investment, while Teesside and Northumberland demonstrate how Britain’s legacy energy infrastructure could support a new generation of enormous AI campuses.
The common factor is not cheap land. It is electricity. As computing requirements continue to increase, the most valuable development sites may increasingly be those where investors can answer three questions with certainty: how much power is available, when it can be delivered and whether that capacity is genuinely secured.
That could have profound consequences for industrial land values across England. Sites previously valued according to warehouses, factories or conventional redevelopment potential may command entirely different prices when their electrical infrastructure is recognised. The next generation of England’s data-centre market may therefore be determined less by where technology companies would ideally like to locate and more by where the electricity system allows them to build.
For property investors searching for the next data-centre location, following the power network may increasingly prove more important than following the motorway map.
Source: © CIJ.World UK Research & Analysis Team
