Ofgem Ramps up Demand Reform

This consultation warrants attention from anyone involved in data centre development. The proposed reforms would impose material new costs, constrain strategic flexibility, and require developers to evidence project commitment at stages that sit uncomfortably with established market practice. Submissions close on 16 September 2026, and the outcome will directly influence long-term investment decisions and development viability for data centre projects across the UK.

Demand signals: reform that has been energised

These proposals have not come from nowhere. Ofgem published its Demand Connections Update in November 2025, flagging its intention to curate a viable queue, enable faster connections and prioritise strategic demand. NESO simultaneously launched its own call for input on the entire demand queue. In February 2026, Ofgem followed up with a formal Call for Input on Demand Connections Reform, receiving 120 responses. In June 2026, Ofgem published both a summary of those responses and a Connect update outlining further reform workstreams.

So the direction of travel has been well signposted. But this consultation crystallises Ofgem's position in a way that should cause developers and investors to pause: the regulator is now seriously considering some of the more interventionist elements of the proposed reforms. The level of specificity (for instance on fee ranges, milestone requirements and evidence thresholds) signals that Ofgem and NESO have moved beyond exploratory dialogue; they are indicating a commitment to disrupting what many in the industry would have regarded as the established approach to data centre development and deployment.

Total gridlock: the problem Ofgem is trying to solve

The consultation responds to explosive growth in the connections queue. Between November 2024 and June 2025, contracted demand surged from 41 GW to 125 GW, of which approximately 73 GW (comprising 315 projects) are data centres. For context, peak electricity demand in Great Britain in 2025 was just 45 GW. Ofgem's view is that many of these projects are speculative or non-viable, and that this is distorting network investment signals, locking up scarce capacity and delaying credible developments. In the context of total operational capacity estimated at around 2GW, just under 1GW under construction and around 11 GW of announce or permitted pipeline (Mapping the UK's data centres build-out: implications for digital sovereignty | CITP), Ofgem’s view, instinctively, feels credible. Wider connections reform measures have already accelerated around 7.8 GW of projects by an average of six years, and Ofgem expects these demand-side reforms to have a similar impact.

The Two Key Proposals

1. Data Centre Commitment Fee

Projects above 40 MW will be required to post a returnable financial commitment from connection offer acceptance until energisation. This would capture the vast majority of data centre projects in the queue, as 99% of the 73 GW queue is composed of projects above 40 MW.

The proposed fee is 2.5% to 7.5% of average capital expenditure per MW (benchmarked at £9.5 million per MW), equating to approximately £237,500 to £712,500 per MW of requested capacity. For a 100 MW project, this means securing between £23.8 million and £71.3 million at offer acceptance. The fee would be returned if the project proceeds to energisation but forfeited if the project terminates (which ties in with the second amendment).

2. Data Centre Queue Management Milestones

Developers would be required to provide evidence of genuine project progression at defined stages:

• M0.5.Dc (six months post-offer acceptance): Select a commercial pathway (self-operation or lease/sale) and provide non-binding compute offtaker evidence (e.g., heads of terms or MoU).
• M2.Dc (same timeline as existing M2: securing planning): Demonstrate procurement of long-lead electrical equipment such as transformers and switchgear.
• M6.Dc (same timeline as existing M6: agreeing construction plan): Provide evidence of financial capability (investment-grade credit rating or committed credit support), technical capability (certification against British Standard EN 50600, TIA-942, or Uptime Institute standards), and (for lease/sale projects) binding compute offtaker evidence (e.g., a master services agreement covering at least 20% of compute capacity for at least one year).

It is easy to see some of these creating timing issues in the context of a typical DC development. For instance, committing to long-lead items before planning certainty needs to be achieved: different offtakers have materially different designs, so would this be a barrier to a project intended as build-to-suit, but flexible as to which offtaker?; or providing evidence of binding compute offtaker before FID (M7) needs to be achieved.

These could cause de facto acceleration of the other (existing) milestones in order for them to be satisfied in time.

Failure to meet these milestones could result in ejection from the connections queue and associated termination liabilities/securities; to say nothing for the potential liabilities under any pre-let agreement if the position in the queue is lost.

The key bytes: scope and application

The commitment fee applies to projects at 40 MW and above connecting at both the transmission and distribution level.

Queue management milestones apply at a lower threshold of 10 MW rated IT load.

Both measures apply to existing and future projects.

All demand projects must self-declare whether they meet the statutory definition of a data centre. Misrepresentation may result in contract termination.

Alternative current: why the changes could affect developers and investors

While clearing the queue of speculative projects is something that will benefit all serious data centre investors, the detail of these proposals raises a number of significant concerns:

Powered land strategies at risk. Many industry participants acquire and hold sites with secured or in-progress grid connections as part of a "powered land" strategy, progressing relevant planning approvals and de-risking the asset before either securing a specialist data centre developer, or selling to an offtaker. A substantial upfront commitment fee, potentially running into tens of millions of pounds, could make this approach commercially unviable or prohibitively expensive. Moreover, the queue management milestones could be impossible to satisfy, making the site unmarketable.

Strategic flexibility constrained. The M0.5.Dc milestone forces developers to declare within six months of offer acceptance where they intend to self-operate or pursue a lease/sale model. In practice, development strategies evolve as market conditions, offtaker interest and financing arrangements crystallise. Being locked into a pathway at such an early stage is a heavy fetter on the strategic flexibility that is essential for successful data centre development. Data centre projects take several years and the market can change rapidly in that time.

The offtaker commitment paradox. Perhaps the most significant tension in the proposals is the requirement to evidence offtaker demand at early milestones (non-binding at M0.5.Dc; binding at M6.Dc). In practice, the market works the other way around: compute offtakers (particularly hyperscalers) typically will not commit with a developer until a developer has significantly derisked a project - usually by securing a firm power connection. Requiring developers to demonstrate offtaker commitment before they have a firm energisation date puts the cart before the horse and risks placing undue negotiating power in the hands of a small number of large offtakers.

Impact on smaller and mid-market developers. If the financial commitment thresholds and offtaker evidence requirements are crafted stringently, these reforms have the potential to reduce the likelihood of progression for projects that cannot easily secure early-stage commitments from hyperscale customers. The proposals risk creating a two-tier market that favours well-capitalised, vertically integrated players with existing hyperscaler relationships, while crowding out the broader developer community. This could have the ironic consequence of making projects outside of core areas like Slough even harder, despite the Government’s previously stated intention to encourage data centre development in more remote areas.

Disruption to existing deals. These proposals introduce new risk factors to projects that have already contracted power on the basis of existing connection arrangements (for example, deals conditional on securing a connection offer or a Gate 2 offer). The additional requirements are unlikely to have been contemplated or allocated for in those contracts. This echoes the disruption experienced when the Gate 2 process was first introduced under the TMO4+ reforms, and developers should review existing contractual arrangements carefully.

Knock-on impacts on energy innovation. The proposals do not operate in a vacuum. If data centre developers' projects face increased risk of queue ejection or delay, that exposure inevitably flows through to the energy providers and investors backing non-grid and bridging energy infrastructure solutions designed to meet data centre demand (including private wire, on-site generation, and behind-the-meter arrangements. Government, Ofgem, and NESO should generally be supportive of this innovation on the supply side of the powering data centres puzzle). Reforms that make it significantly more costly, complex or risky for DC developers to progress their projects risk drying up the very energy innovation pipeline that is needed to complement grid connections and deliver capacity at pace.

Conservative deal structuring ahead. Any degree of regulation carries uncertainty until it is finalised, and in the short term this may dampen investor appetite. The proposed commitment fee represents an additional stranded cost that will need to be embedded into investors' risk profiles, potentially increasing the overall cost of data centre development in the UK. Until the detail of the final rules is known, market participants will (as with the Gate 2 experience) be forced to take a very conservative approach to deal structuring and risk allocation, which itself has a chilling effect on transaction activity and project progression.

A deliberate shock? But at what cost?

There is a case to be made that these proposals represent a deliberate and intentional signal from Ofgem and NESO that the "old world" approach to securing and holding grid connections (characterised by speculative applications, large capacity reservations, and extended development timelines) is no longer acceptable. The scale of the queue problem is real and doing nothing is not an option.

However, the Government has simultaneously recognised data centres as critical national infrastructure, designated AI Growth Zones with access to the Connections Accelerator Service and used the Planning and Infrastructure Act 2025 to give Ofgem and the Secretary of State new three-year powers to amend licences, codes and connection agreements to accelerate strategically important demand projects.

There is a real tension between the Government's stated ambition to position the UK as a global leader in AI infrastructure investment, with the value of that investment spread throughout the UK and reforms that (if not carefully calibrated) risk undermining or delaying the very pathways to delivery that make the UK attractive to international capital and limiting investible projects to those in core regions.

The devil, as always, will be in the detail. The consultation process is an opportunity for the industry to engage constructively and ensure that sensible, proportionate policy prevails: policy that clears the queue of genuinely speculative projects without inadvertently hindering the investment pipeline and the Government's own AI ambitions for projects that will deliver the capacity the UK needs.

Next Steps

Ofgem intends to take decisions on these proposals later in 2026, with implementation through changes to the CUSC, licences and contractual arrangements. The consultation response deadline is 16 September 2026.

Responses should be submitted to connections@ofgem.gov.uk.

We would welcome the opportunity to discuss these proposals with you and explore their implications for your projects.

Please do not hesitate to contact us to arrange a call. Statistics stated in this bulletin but where sources have not otherwise been cited are drawn from information in Ofgem’s publications.