Summary

  • Ofgem’s queue rose from 41 GW of contracted demand in November 2024 to 125 GW in June 2025; it attributes roughly 73 GW across about 315 projects to data centres.
  • Its closed consultation proposes a returnable commitment security and milestone tests for large projects. The rules remain proposals, and Ofgem had not published its decision by 23 October.

The grid queue has become a place to hold an option. A developer can secure a connection position while planning permission, customers, equipment and finance catch up. That position matters because a data centre without power is an expensive building-shaped ambition, while a credible route to power can change what land and capital are worth.

Ofgem’s own figures show the size of the bet. Contracted demand in the queue climbed from 41 GW in November 2024 to 125 GW by June 2025. Around 73 GW belonged to 315 data-centre projects. Great Britain’s peak electricity demand in 2025 was 45 GW. The 125 GW records requested connection commitments; actual consumption depends on which projects reach energisation. Accumulated requests shape network investment and can leave viable projects waiting behind positions that never energise.

The regulator says a key reason is that securing and keeping a demand connection position is relatively inexpensive. Ofgem also found that at least 9 GW of transmission projects changed their requested technology from batteries to data centres between May 2024 and August 2025. Those category changes show how strongly queue rules can influence what developers ask the system to reserve.

Ofgem published the consultation page and its July consultation document. The proposal pairs two tests: money at risk and evidence that a project is advancing.

The proposed commitment fee would apply at 40 MW or above to projects connecting to transmission or to distribution networks where they trigger a Transmission Entry Assessment. Developers would secure the full proposed amount from accepting a Gate 2 offer until energisation. Ofgem set a consultation range of 2.5% to 7.5% of its average data-centre capital-cost assumption: about £237,500 to £712,500 per MW. At 40 MW, that arithmetic produces £9.5 million to £28.5 million in secured commitment. The money is returnable if the project meets the requirements and proceeds; it can be forfeited if the project terminates or fails them.

The security ties money to project performance over time.

The second test asks what sits behind the capacity request. Proposed milestones cover a credible compute customer or end user, procurement of long-lead electrical equipment, and the developer’s financial and technical ability to build. A self-operated facility could show that the developer or a related company intends to operate it. A project intended for lease or sale would need a credible route to a tenant, purchaser or third-party operator, with binding customer evidence later in its development. The evidence follows the business model: an operator can document its own build plan, while a landlord needs a route to a tenant or buyer.

There is a sound operational reason to ask. A queue is a plan for future infrastructure, and a paper project can distort that plan. The danger sits in the test itself. A developer may have a genuine site and customer but still face planning delay, unavailable switchgear, financing conditions or a network schedule it cannot move. Operators will need to separate project-controlled delays from planning, equipment and network delays.

Ofgem proposed milestones tied to the time remaining before a project’s connection date, with longer schedules for projects farther out. It also proposed separate evidence pathways and a one-time change between them. Those details make the package more than a flat deposit. They also make implementation decisive: administrators must assess changing evidence, explain failed milestones, apply grace periods consistently and distinguish a late project from a dead one.

The consultation closed on 17 September. Ofgem’s page still listed it as awaiting a decision on 23 October. Until the final rules appear, the proposal had charged no developer and released no queue capacity.

The test for the finished policy is straightforward. Publish how much capacity leaves the queue, how often securities are forfeited, which milestones projects miss, and whether cleared capacity advances projects with stronger evidence. A fee can make a queue more expensive. Only a visible trail from evidence to decision can show that it has made the queue more truthful.