Datacentre connections: babies and bathwater?

Ofgem is consulting on reforms designed to stop speculative data-centre projects clogging up the electricity connections queue.

The problem is real. A firm grid connection is increasingly valuable, and scarce network capacity should not be tied up for years by projects with little prospect of being built.

But the proposed solution risks creating a different problem.

At the centre of Ofgem’s proposals is a new Data Centre Commitment Fee. For projects of 40 MW or more, developers would need to provide financial security from the point they accept a Gate 2 connection offer until energisation. The proposed fee is substantial: £237,500 to £712,500 per MW. For a 500 MW development, that equates to roughly £120 million to £360 million of security. This does not necessarily mean setting aside that amount in cash. Developers could use established forms of credit support. But they would still need the balance sheet, banking capacity or financial backing to stand behind a very large obligation. And the exposure is not simply a financing cost: the fee can be forfeited if the project terminates or fails to meet relevant requirements.

Some meaningful risk of forfeiture is necessary if the fee is to change behaviour. The question is whether the scale and timing of that exposure are proportionate to what developers can reasonably know and control at that stage of the project.

That matters because the electricity network and the data-centre market operate on very different clocks. A connection may need to be secured many years before energisation, particularly where substantial transmission reinforcement is required. Hyperscale customers typically make binding capacity decisions much closer to delivery, . Published evidence is imperfect, but points towards customer commitments often being made only 2-3 years ahead. A developer may therefore need to accept a connection offer, and take on hundreds of millions of pounds of contingent exposure, years before the eventual customer is ready to make the commitment that would underpin the project.

That creates a risk that the proposed fee tests access to a large balance sheet as much as the quality of the underlying project. This matters because the data-centre development market contains a range of specialised businesses. Some developers secure land, progress planning and develop a credible grid proposition before partnering with or selling to a hyperscaler or colocation provider. That activity is different from simply acquiring a grid position in the hope that it becomes valuable.

Ofgem is right to want to discourage the latter. But a very large early financial hurdle could also deter the former. The consequence could be a market increasingly tilted towards hyperscalers, major colocation businesses and the largest diversified developers. That would not necessarily be benign. Specialist developers may have comparative advantages in land assembly, planning, environmental assessment, network development and local stakeholder engagement. If fewer of them are willing or able to originate GB sites, hyperscalers could ultimately have fewer mature, investable propositions from which to choose. Given that data-centre capital is internationally mobile, the downside could be fewer genuinely investable GB projects, not simply fewer speculative applications.

There may be a better way to distinguish the two.

Rather than attempting to eliminate development risk at the outset, the connections regime could ask whether a project has a credible pathway to delivery, with the evidence required becoming progressively stronger as the project matures. Outline planning permission could form part of that early test. It would not prove that a project will be built, but obtaining it requires meaningful expenditure and development work while preserving flexibility over detailed design. Combined with land rights and a meaningful but more modest financial commitment, it could provide useful evidence that a project is serious. And the test could then strengthen over time. Developers could then be required to show progress through reserved matters or more detailed approvals, with financial security rising as the project matures and the connection becomes more valuable. Closer to construction, stronger evidence of financing, procurement and investment commitment would become reasonable.

The challenge, therefore, is not whether Ofgem should make speculative queue positions harder to hold. It should. The challenge is how to do so without screening out credible development models alongside speculative ones. A good regime should reward evidence of genuine progress, require financial commitment that grows with project maturity and remain neutral about the particular commercial route through which a viable project reaches delivery. That is the baby-and-bathwater problem Ofgem now needs to solve.