The UK data centre construction market is facing mounting delivery pressures as power availability, electrical equipment procurement and specialist labour shortages threaten to delay projects despite strong investment demand.
UK data centre construction faces new grid and equipment bottlenecks
In its third-quarter 2026 construction market intelligence report, Rider Levett Bucknall (RLB) identified grid connections as a major constraint, with some developers reportedly being offered connection dates extending to 2038 or beyond. The consultancy also highlighted lead times of four to five years for large high-voltage transformers. Delays of up to two years for some medium-voltage switchgear designed for data centres were also noted.
Meanwhile, Ofgem’s consultation on proposed data centre connection reforms closed in September 2026, with a final decision still awaited as of October 24. The proposals include commitment fees and project milestones. These are intended to prevent speculative developments from occupying scarce grid capacity without demonstrating genuine progress.
As Ofgem continues its facility’s consultation process, UK technology company, Era4 has secured planning permission to build its AI data center beside Shelford Landfill in Canterbury, Kent. In contrast, Era4’s electricity supply is expected to come from a renewable power plant fueled by methane gas from the landfill. Era4’s Shelford Landfill data centre will also connect to the National Grid. This is an area of interest as the tech company continues with development.
Cable and optic fibre manufacturing presents another challenge
The challenge extends beyond grid connections. On October 22, 2026, Reuters reported that cable manufacturing capacity was emerging as another bottleneck for AI data centre construction. Cable and optical fibre company, Nexans, warns that establishing a new production line for high-voltage cross-linked polyethylene cables can take approximately three years.
These constraints reinforce the need for developers to secure critical electrical equipment and supply-chain capacity earlier. They will also need to coordinate grid infrastructure with construction programmes, and adopt procurement strategies that reflect actual manufacturing lead times.
For contractors, the implication is that delivering the UK’s expanding data centre pipeline will depend not only on workforce capacity and financial risk-sharing. It will also depend on securing the equipment and power infrastructure needed to bring completed buildings into operation.
Does the UK data centre construction market have the capacity to delivery?
As the £3.75bn DC01UK scheme in Hertfordshire — now Equinix’s flagship UK campus — pushes the data centre sector into megaproject territory, one industry voice argues that the way Britain builds these schemes, not just what it builds, now needs rethinking.
The approval of the £3.75bn DC01UK data centre in Hertfordshire — a scheme set to become one of the largest in Europe, and since acquired by U.S. operator Equinix in a deal worth close to £4bn — has been framed largely as a planning and green-belt story. Hertsmere Borough Council’s planning committee waved it through in January 2025 despite local opposition running close to two-to-one against, on the judgement that the project would do “more good than harm.” The site is due to deliver some 2 million sq ft of space and around 250MW of capacity, with construction slated to begin in 2027 and operations by 2030.
But beyond the politics of where these campuses are built lies a harder, less-discussed question: whether the UK construction market has the capacity, the balance-sheet strength and the risk appetite to deliver them at all. It is a question worth pausing on as approvals like DC01UK’s become more common — and one that James Vaughan, Associate Director at BCS Consultancy, a global consultancy specialising in the delivery of data centre and critical infrastructure projects across the UK and Europe argues the industry has not yet seriously confronted.
A market at critical mass?
Vaughan’s central observation is one of scale. “A decade ago, many UK data centre projects were measured in tens of millions of pounds,” he notes. “Today, AI campuses are routinely measured in billions.” The challenge, in his view, is no longer about simply building larger facilities, but whether the construction market has the capacity and risk appetite to deliver them.
The DC01UK scheme is his case in point. Forecast to deliver more than 300MW of IT capacity at an estimated construction value of £3.75bn and to create around 2,500 construction jobs, projects of this scale, Vaughan points out, “rival or exceed the annual turnover of many major UK contractors, creating new challenges around risk allocation, financing and delivery capacity.” A review of the UK’s largest construction firms, he adds, shows that even the sector’s biggest players operate at a scale that can make individual multi-billion-pound projects challenging to absorb onto a single balance sheet.
Compounding the pressure is timing. While such projects can in theory be split into phases or multiple contracts, developers — racing to capture AI-driven demand — are increasingly seeking concurrent delivery to accelerate time to revenue. That, Vaughan warns, places additional strain on an already-stretched contractor market, set against the sector’s widely reported shortage of skilled tradespeople.
The bond-market squeeze
If capacity is one constraint, the financial plumbing behind these projects is another. As project values climb, Vaughan notes, bond and insurance requirements are becoming both more expensive and harder to secure. The collapse of contractor ISG, combined with insurers reducing their exposure to large construction projects, has tightened the market further — with some major bond providers withdrawing from parts of the construction bond market altogether, and 37% of CECA members reporting dissatisfaction with bond availability.
“For contractors seeking to enter the data centre sector, securing appropriate bonding and insurance is becoming increasingly challenging,” Vaughan says, adding that BCS has observed even established contractors facing greater scrutiny when securing bonds for major projects. His underlying point is blunt: bond pricing reflects project risk, and as the scale and complexity of these schemes grow, the industry will have to find new ways of sharing that risk between developers, contractors, insurers and — potentially — government, if delivery is to keep pace with demand.
A lesson from the Tideway
Here Vaughan reaches for a precedent from outside the data centre world: London’s Thames Tideway Tunnel. With final costs approaching £5bn, Tideway carried risks that would have been difficult for any single contractor or insurer to absorb alone. Rather than relying solely on traditional contracting structures, risk was distributed across contractors, insurers and government-backed support mechanisms.
The most significant feature, in Vaughan’s reading, was the government’s role as an “insurer of last resort” for extreme events — a backstop that reduced exposure for contractors and insurers, improved market confidence and lowered delivery risk. With data centres now formally recognised as critical national infrastructure, he argues there is a clear opportunity to explore whether similar risk-sharing approaches could underpin the delivery of strategically important digital infrastructure.
What it means for the next wave
Vaughan’s conclusion is that data centre developments are increasingly reaching a scale more commonly associated with major national infrastructure programmes than with traditional commercial construction — and that continuing with conventional contracting models, risk allocations and financing approaches is unlikely to support the rollout of the next generation of AI infrastructure.
“Contractor balance sheets are under pressure, margins remain thin, and the bond and insurance markets are increasingly constrained,” he writes. “As project values move into the billions, these challenges will only become more pronounced.” His prescription: developers may need to embrace alliance-based models, rethink who owns which risks, and consider where government can act as a stabilising force for projects that underpin national economic and digital resilience — while becoming more visible in local and national conversations about infrastructure delivery.
For a project like DC01UK, that reframing matters. The Hertfordshire scheme cleared its political hurdle on a promise of jobs, business-rate revenue and more than £1.1bn a year in Gross Value Added. Whether it — and the wave of hyperscale campuses lining up behind it — actually gets built on time and on budget may depend less on planning committees and more on the kind of delivery and risk-sharing models Vaughan is describing.
As the UK market becomes less London-centric and looks to new technologies and partnerships to unlock electrical capacity, his closing challenge to the industry is a simple one: now is the right time to review not just what we build, but how we build it.
Leave a Reply