Like many other territories, the UK has a substantial grid connection backlog, leaving data center projects waiting many years for electrical power. The UK's energy regulator, Ofgem, has proposed a solution that may appeal to authorities elsewhere, but which could risk damaging the country's digital strategy: imposing a significant fee to remain in the queue.

According to Ofgem's consultation documents, 125 GW of large power demand projects have applied for connection to the UK's grid — three times the country's 45 GW peak demand — with data center projects making up around 73 GW of this total (see Proposed data centre connection reforms). With around 4 GW of new capacity joining the grid each year, new applications could theoretically expect to wait 30 years for power, if all these applications are genuine and viable.

In the US, a rising tide of public opposition to data centers has increased the pressure on regulators. Two US states have recently responded:

• New York introduced a 1-year moratorium (see New York's data center pause may set a global precedent) on new projects. New York State's grid queue (12 GW) amounts to 30% of the state's peak demand.
• Texas has paused all approvals while its queue is audited, and will apply additional conditions to all projects in the queue. The governor has stated that 474 GW of projects have applied for connection, five times the state's 91 GW peak demand. Although Texas added 11 GW of capacity in 2025, it has the potential to add substantially more using natural gas.

Rather than imposing a moratorium or pause, the UK's Ofgem has proposed an arguably more straightforward solution: imposing a fee to deter and remove non-viable and speculative projects from the queue. A deposit is required to enter (or stay) in the queue, is refunded if the project is completed and energized, but is forfeited if the project terminates or (potentially) misses key milestones. Consultation on this proposal is open until 16 September 2026.

The proposal

Several components of the Ofgem proposal are not fully defined, including the size of the fee and the milestone requirements. Other elements of the scheme may be subject to change, but Ofgem states that a majority of those organizations it has consulted are in favor of a commitment fee.

Connection fee

Ofgem proposes to levy a data center connection fee (DCCF) of between 2.5% and 7.5% of average data center capital expenditure on projects over 40 MW. Ofgem's suggested price range, which is between £237,000 ($322,000) and £712,000 ($966,000) per MW, would temporarily add between £23.8 million ($32.3 million) to £71.3 million ($96.8 million) to the cost of a 100 MW project.

The fee is forfeited if the project is withdrawn, is terminated by the UK's grid operator the National Energy System Operator (NESO) or fails to secure planning permission. If a project is downsized, the unused portion of the fee is retained by Ofgem. NESO or Ofgem may revise the power capacity threshold and fee level in the future.

The DCCF would apply to all projects already in the queue (except those 6 months or less from energization) unless developers opt to leave the queue during a grace period at the start of the scheme.

Projects become subject to the fee by self-declaring as data centers. Hybrid projects (such as data centers with generation capacity) are charged a fee on the data center component.

The DCCF can be secured with a credit note, guarantee or cash deposit. It is effectively added to existing cancellation fees under Ofgem's Connection and Use of System Code (CUSC), which apply to all connection applications.

Milestones

Alongside the DCCF, Ofgem proposes mandatory milestones for all data centers above 10 MW, which is lower than the DCCF threshold, to filter out any projects without sufficient financial backing or technical ability.

Developers need to choose between two pathways:

• Self-operation. The developer will operate the data center: running hyperscale functions, cloud services, managed space or its own enterprise applications.
• Lease or sale. The project will be handed over to a compute off-taker, such as a tenant, purchaser or third party.

One project can include capacity apportioned to both pathways. Developers can switch capacity between pathways during the development process, but only once. Lease or sale capacity needs to be backed within six months by a non-binding off-taker agreement, which must become binding at a specified later date before the project is connected (see below).

The proposal extends an existing series of numbered milestones within the CUSC. Data center projects must meet a new initial milestone (M0), alongside additional requirements introduced at milestones M2 and M6 (see Table 1). The deadlines for M2 and M6 are flexibly related to the end date of the project:

Table 1 Ofgem's Connection and Use of System Code: milestones and requirements

A project that fails to meet a milestone may be removed from the queue and lose its DCCF, although Ofgem has proposed grace periods and transitional arrangements.

Discussion

The proposal has two major effects: the need for additional funds or credit at the start of the project and a requirement to meet development milestones and share details to with Ofgem during project development.

Ofgem reports that it examined comparable commitment fees in Spain, Texas (US) and Georgia (US), before proposing one at a higher level to suit the UK's specific market context. to deter owners from banking powered land with speculative applications, and developers making multiple applications for a single project across different locations. The largest project known to be in the queue, at 1,500 MW, would be required to pay a fee of between £356.25 million ($482.15 million) and £1.07 billion ($1.448 billion).

Since May, Spain has been charging a monthly fee around €800/MW to remain in the queue, Texas is proposing a fee of $50,000MW plus $50,000 in security, while Georgia has no public commitment fee.

Industry responses to the proposal have included the following observations and concerns:

• Immediate uncertainty. The Ofgem proposal will have an unpredictable impact on the grid queue. Many projects will withdraw from the queue immediately, but the number of projects that will withdraw remains uncertain ahead of the new regime. The queue could remain significantly oversubscribed, or there might be significant available capacity. This is the risk for developers: they cannot know in advance what improvement in connection time their DCCF will buy.
• Viable projects may be culled. The DCCF may exacerbate the problem it is designed to solve. By requiring additional financial requirements before final investment decision or planning consents, it could render some otherwise viable projects uneconomic. The level of risk is highly uncertain as the DCCF terms are subject to future change by NESO and Ofgem.
• Divergence from national strategy. Ofgem's approach is driven by its own priorities and may not align with broader national priorities. A high DCCF will favor the best-funded projects, but this may have unintended consequences. For example, it could allow well-funded US hyperscalers to undermine sovereign infrastructure providers, and force innovative projects to relocate outside the UK. Retained DCCF payments might contribute to nationally required grid upgrades, but are more likely to be used to insulate consumers from electricity price increases.
• Additional evidence more of a concern than DCCF fee. Large operators have commented that the DCCF fee level may be acceptable, but obtaining the necessary letter of credit will require guarantees against complex risks. This letter would, in effect, become a financial instrument that does not yet exist.
• Risk of speculative off-takers. Lease or sale pathway projects will be required to sign a non-binding off-taker agreement at the M0 milestone, sooner than is currently the case, and then will be required to sign a binding agreement at M6. This creates the possibility of speculative off-takers signing an agreement at M0, then attempting to extort favorable terms at M6 from a developer that needs to sign a binding contract.
• Queue shift. The queue of speculative projects may simply move to other jurisdictions. Projects may opt to remain at Gate 1 (a holding stage before entering the queue) or may opt to prioritise planning approval first, shifting the backlog to earlier in the process.
• No consideration for phased projects. Ofgem estimates that its DCCF represents around 3-9 months of a data center's expected return, based on the whole project's internal rate of return. However, data center projects are generally scoped and delivered in phases, meaning the project revenues would take longer to recover the fee.
• Contingent milestone achievement. Some milestones may fall outside of a developer's control, and the information they require might be bound by partner confidential agreements.
• NESO assessment. NESO will be required to assess financial and technical capability, along with the credibility of partners (including off-takers). It is unclear how the grid operator will implement this process.
• Behind the meter (BTM) power. Does the 40 MW apply to the power imported or the computing capacity? For example, if a project with 100 MW computing capacity has 70 MW BTM power — and imports only 30 MW — will it pay DCCF on 100 MW, or not pay it at all?
• Impact on other energy networks. If projects opt to connect to the gas network, will there be a similar DCCF to prevent speculative applications for gas capacity?
• Undefined matters. Ofgem has not specified how interest accrued on DCCF funds would be handled, or what form an appeals process might take.

Uptime Intelligence advises all developers with UK data center projects or plans to engage with the Ofgem consultation before the 16 September deadline, and developers outside the UK to engage with any grid queue management reforms that emerge in their own territories.