Can Self-Build Change the Economics of Time-to-Power for UK Data Centres?

For data-centre developers, time-to-power is increasingly an economic variable rather than simply a technical milestone.

A site that cannot be energised when required may face a very different commercial outcome from one with an equivalent amount of grid capacity available several years earlier. That makes the way a connection is physically delivered increasingly relevant to development risk, capital requirements and ultimately investability.

This is part of the broader shift discussed in our recent analysis, From Grid Connection to Deliverable Power: How UK Data-Centre Investability Is Changing: grid access still matters, but so does the credibility of the route from connection capacity to an operational site.

Ofgem's latest transmission-connections consultation introduces a potentially important change.

Published on 7 October 2026, the proposals would expand the ability of demand, generation and storage customers to construct the sole-use transmission assets connecting their site to the existing network and then transfer those assets to the relevant Transmission Owner. The consultation remains open and the framework is not yet settled policy. Ofgem's consultation sets out the proposed approach.

For UK data centres, the commercial question is therefore not simply whether self-build becomes possible more widely. It is whether greater control over connection delivery can materially change the economics of time-to-power.

What self-build would actually change

Under the traditional model, a connecting customer can be heavily dependent on the network company for the delivery of high-voltage connection infrastructure.

Ofgem's proposed framework would broaden the alternative: the connecting customer could take responsibility for delivering certain sole-use assets between its site and the existing transmission network, subject to the required standards and approvals, before transferring those assets to the Transmission Owner.

The concept itself is not entirely new. Contestable construction already exists in parts of the electricity system, but Ofgem has acknowledged that self-build of high-voltage transmission assets is currently available only in limited circumstances. Its June 2026 demand-connections reform update cited, as examples, restrictions around proximity to the existing transmission system and whether the assets could potentially be shared.

The proposed reform is intended to make that route available more widely.

That matters because infrastructure delivery and network capacity are not the same problem.

A project may have an agreed route to connection but still face a long programme for the physical infrastructure required to reach the grid. If part of that programme can instead be procured and managed directly by the developer, an element of the connection timetable potentially moves from something the project largely waits for to something it can actively deliver.

Time-to-power becomes partly an execution question

The attraction of self-build is therefore less about ownership of transmission infrastructure than about control over schedule.

Ofgem has said that stakeholder support for greater self-build reflects several expected benefits: greater control over delivery, potentially quicker or cheaper connections and reduced pressure on Transmission Owners' delivery resources.

For a data-centre development, that distinction can be commercially significant.

A shorter connection programme can affect how long development capital remains exposed before the site becomes operational. It can affect the point at which an operator or end customer can realistically commit to capacity. And it can influence whether a project is capable of meeting a particular market window rather than arriving after competing capacity.

Self-build does not automatically solve these issues. But it potentially changes who controls one part of the critical path.

That can matter even where the eventual capital cost of the connection is not dramatically different.

In a market where power availability increasingly differentiates sites, certainty of delivery may be at least as important as the headline cost of delivery.

But self-build does not create grid capacity

This is the most important limitation.

Self-build can potentially accelerate the infrastructure between a site and the transmission network. It cannot remove constraints elsewhere on that network.

Ofgem itself makes this distinction explicitly. In its June update, the regulator cautioned that any time or cost benefits in Great Britain would depend on the particular project and on whether additional work was required on the existing transmission system before the site could be energised.

That means the commercial value of self-build is likely to vary materially between projects.

Where the principal constraint is the delivery of dedicated connection infrastructure, giving a developer greater control over procurement and construction could be meaningful.

Where the real bottleneck is a major upstream reinforcement programme, self-building the final connection assets may have limited effect on the ultimate energisation date.

The distinction is important because "self-build" should not become another shorthand for "faster power".

The relevant question is which part of the connection programme is actually determining time-to-power.

The potential upside comes with transferred risk

Greater control also means greater responsibility.

Ofgem's work recognises that self-build creates questions around engineering standards, system interfaces, asset quality, adoption, defects and the allocation of cost and risk between the connecting customer, Transmission Owner and wider electricity consumers.

For developers, this means that self-build would not simply remove a delivery risk. Some of that risk would move.

A project choosing the self-build route could gain greater control over procurement and construction but would also assume more responsibility for delivering infrastructure capable of meeting the standards required for connection and eventual transfer.

That changes the capital proposition.

A credible self-build strategy is likely to depend not simply on access to funding, but on the project's ability to procure, manage and deliver high-voltage infrastructure within a regulated interface.

For investors assessing data-centre developments, that makes execution capability more relevant—not less.

The evidence on cost and timing is encouraging, but should be treated carefully

Ofgem points to Ireland, where self-build and transfer have been used more extensively for more than two decades.

Its June update cites Irish case studies in which contestably delivered connection works were approximately 20% cheaper and 11 months faster.

Those figures are notable, but they should not be treated as a forecast for Great Britain.

Network architecture, regulatory arrangements, individual connection scope and existing-system constraints differ by project and by market. Ofgem itself immediately qualifies the comparison by stating that equivalent benefits in Great Britain will depend on project circumstances and the work required on the existing network.

The more important evidence from Ireland is therefore not the precise percentage saving.

It is that giving sophisticated connecting customers greater responsibility for dedicated connection infrastructure can, under the right circumstances, alter both delivery time and cost.

A different way of thinking about connection risk

The Government has already identified timely electricity connections as the largest barrier to delivery of AI Growth Zones and has specifically committed to exploring mechanisms that allow developers to construct their own high-voltage lines and substations where doing so could accelerate connection. The policy direction is set out in Delivering AI Growth Zones.

Ofgem's October consultation begins turning that policy direction into a potential regulatory framework.

For the UK data-centre market, its significance goes beyond another connection-process reform.

It introduces the possibility that time-to-power becomes, at least partly, something that capable developers can influence through infrastructure delivery rather than simply inherit from the network programme.

That will not make every constrained site viable. Nor will it remove the importance of upstream capacity, system reinforcement or the Transmission Owner.

But where the physical connection itself is a material source of delay, the economics could change.

Self-build does not create power where the system cannot provide it. What it can potentially change is the degree of control a project has over converting available network capacity into power at the site. In an increasingly power-constrained data-centre market, that distinction could become commercially important.