The site is only as good as its power connection
Data centre site selection is becoming a power and infrastructure problem as operators discover that finding land is easier than finding a site that can actually be energised.
The traditional data centre site selection checklist is becoming less useful – Land, fibre connectivity, planning and proximity to established markets still matter, but for large, power-intensive facilities, none of them matter much if the site cannot secure enough electricity and this is increasingly becoming the constraint.
CBRE’s European data centre research identifies power availability and grid constraints as major limitations on new capacity – in the UK, its research points to operators considering larger campuses and locations outside established clusters where additional power can be secured which changes the economics of the site.
A piece of land is not necessarily a data centre opportunity simply because it has the right planning potential and a nearby fibre route; the critical becomes whether the electricity infrastructure can support the facility, and when that capacity can actually be delivered.
A connection on paper is not capacity in practice
This is where the problem becomes a little more complicated: Grid capacity has to be assessed in terms of timing as well as megawatts. A prospective site may have a route to a substantial connection, but if reinforcement work takes years, the theoretical capacity is of limited value to an operator trying to bring a facility online sooner, creating a new form of competition between sites.
It’s no longer just a case of which location has the best connectivity or lowest development cost – it’s which location has the most credible path from land acquisition to energised facility, for large data centre projects, this path can involve grid reinforcement, substations, transmission infrastructure, planning approvals and lengthy construction programmes – every additional dependency adds time and uncertainty.
Existing capacity has limits too
The same issue applies to existing sites: A facility with spare floor space is not automatically suitable for higher-density workloads as existing electrical distribution may have been designed around lower rack loads, cooling systems may not have enough capacity and there may be insufficient space for additional electrical or a thermal plant.
Treating available floor space is not sufficient on it’s own – power, cooling and scalability are key considerations for operators assessing both new developments as well as existing facilities because a megawatt of theoretical capacity is not the same as a megawatt of usable capacity – the infrastructure between the grid connection and the server rack determines what can actually be deployed, and when.
Time is becoming a property issue
This is where the data centre market increasingly looks like an infrastructure market – developers are not just competing for land, they are competing for grid capacity, equipment, planning approvals and construction resources and those things cannot necessarily be accelerated by simply spending more money.
CBRE’s recent European outlook highlights constrained grid capacity and longer development timelines as factors affecting new data-centre supply and for operators, this makes time-to-power a meaningful part of site value.
A site that costs more but can be energised earlier may have a very different commercial proposition from a cheaper site that requires years of infrastructure work – this is particularly important when the site is intended to support rapidly expanding compute demand. Speed to market matters.
The data centre map is being redrawn
The result is not that traditional data centre markets suddenly become irrelevant – it is power availability becoming a much stronger determinant of where new capacity can actually be built which means the site selection process needs to start much further upstream.
How much land is available, how close the fibre is and what the building could look like are all relevant questions – but how much power can we get, when can we get it, what infrastructure is required to deliver it and whether the site can support the facility we need five or ten years from now are fundamental – because these questions increasingly determine whether a data centre site is an opportunity at all.
Of course the physical location still matters, but the infrastructure underneath it matters more.
