12 September 2026
TL;DR
Curate is a new system designed to improve demand management specifically for data centres. It evaluates whether developers are serious and capable of delivering their projects, and it can weed out those who aren’t up to the task.
Greater visibility and financeable, transferable slot reservations address the root causes of delay, rather than relying on retrospective fees or complex milestones.
Recognising the constraints of the power supply chain.
"Industry engagement undertaken during policy development indicated that major electrical equipment, particularly transformers and switchgear, often needs to be ordered well in advance of construction due to manufacturing and delivery lead times.”
Access to power remains the decisive factor, even when everything else is ready. Curate recognises the power supply chain as fundamental and fragmented. The devil is in the details.
We support Curate’s aim to give NESO and distribution network operators a reliable way to separate committed projects from speculative ones. Our response sets out how procurement and financial capability evidence can be both more robust and easier to verify, grounded in what we see daily in OEM capacity and lead times.
Financeable capacity reservations should be recognised, alongside invoices and signed contracts, as evidence that a developer has secured access to scarce equipment. This gives NESO and DNOs an additional, objective basis for assessment, verifiable against a manufacturer’s production schedule.
How we responded.
We deliberately did not answer on fee value and calibration, credit ratings, or technical certification standards. Those are questions for developers and financiers, and we would be speculating.
Q29. Do you agree with the proposed queue management milestone measures? Yes, in principle.
M2.Dc evidence should reflect that variability, so the framework doesn't penalise viable projects whose critical path relies on longer-lead equipment.
Q30. Which option for binding compute offtake evidence, and do you agree with 20% for at least one year? Option 2, with flexibility for phased builds.
We need a workable starting point that sets a meaningful bar. As drafted, only near-full, pre-let, hyperscaler-anchored projects would meet M6.Dc.
Q31. Should the offtake threshold differ by business model? Yes, and design it differently rather than just setting it higher or lower.
Maintain the deterrent for speculative projects without penalising the common colocation model for its slower ramp.
Q32. Should the long-lead requirement cover transformers and switchgear only? No. The scope should be broader.
Limiting the scope creates an evidentiary gap for projects whose critical path depends on other equipment.
Q33. Voltage-linked ranges, or case-by-case assessment? Case-by-case.
NESO and DNOs should receive guidance mapping expected evidence to equipment class and lead-time range, and that where the evidence is a capacity reservation, it can be verified directly against the manufacturer's confirmed production schedule.
Q40. What other financial and technical capability tests could be used? An additional evidence route.
Recognise a financially committed capacity reservation against a verified manufacturing slot, structured so capital deploys in line with the project's draw schedule rather than as a single upfront payment, as evidence of both financial and technical capability. This is verifiable against the manufacturer's production records and gives NESO and the DNOs an objective basis for assessment.
Q42. Do you agree with the proposed timing of each evidence requirement? Partially.
Aligning with existing NESO and DNO milestones is logical. Ofgem should test M2.Dc timing specifically against the longest lead-time categories, such as large GSUs and EHV transformers, to ensure committed projects have a realistic route to producing binding procurement evidence within the window, rather than being caught out by manufacturers’ commercial processes.
What happens next?
Curate is a consultation, not a decision. It closes on 16 September, with decisions to follow later in the year, and the draft grace periods run to six months. The window between now and a live obligation is measurable and short.
Three questions remain genuinely open, and these are where a developer, EPC, or connection provider can influence the outcome:
Q32 - whether the long-lead requirement should cover transformers and switchgear only.
Q33 - voltage-linked ranges, or project-specific assessment.
Q42 - whether the proposed timings are right.
If you're building in Great Britain and your critical path needs HV equipment, answering these three questions is worth twenty minutes of your time before the 16th.
