Energy as the gatekeeper

Energy is now the primary delivery risk for UK data centre development. While the sector’s growth is anchored in cloud adoption and AI-led compute demand, a scheme’s viability is increasingly determined by whether power can be secured at the right scale, in the right location, and within an investable timeframe. This has brought grid capacity, connection timescales and regulatory change to the forefront of site selection and planning strategy, particularly in London and the South East, where demand is concentrated and existing infrastructure is most constrained.

At a high level, the UK electricity system is split between transmission and distribution networks, coordinated by the National Energy System Operator (NESO). Transmission networks move power at high voltage and are operated by three Transmission Owners: National Grid Electricity Transmission (England and Wales), Scottish Power Transmission and SSE Transmission (Scotland). Power then feeds into lower-voltage distribution networks managed by regional Distribution Network Operators (DNOs). NESO, renationalised in September 2024 from the former ESO, sits above these operators, coordinating system planning and, critically for large loads, administering the transmission connection process and queue management.

The tightening of capacity is the product of both supply-side and demand-side shifts. On the demand side, electrification of transport and heating is weighting on baseline electricity requirements, while data centres represent a step change in large, continuous loads. On the supply side, increasing volumes of renewable generation are connecting at lower voltages, changing historic one-directional power flows and increasing the need for reinforcement on local networks. The result is a ‘perfect storm’, where grid upgrades are required not only to accommodate rapidly rising demand, but also to adapt the network itself to a fundamentally different pattern of electricity generation and power flows.

This pressure is most visible in the connection queue. By the time the application window closed in March 2025, NESO had identified around 125 GW of connection requests, around 1.7 times the UK's total grid capacity of approximately 75 GW. Around 140 data centre projects alone were seeking transmission connections, representing approximately 50 GW of requested capacity. While not all projects will proceed, the scale of oversubscription has reinforced the need to distinguish credible demand from speculative applications. Reflecting this, NESO's connection reforms have allocated around 11 GW of demand connections due to be connected before 2030, providing a more realistic indication of the capacity likely to be delivered over the remainder of the decade.

Cost of power and competitiveness gap

Cost is the other side of the power access equation. For data centres, electricity pricing is driven not only by wholesale power markets but also by a growing layer of regulated and policy-related charges, including transmission and distribution network charges and system-balancing costs. This matters because the UK’s non-domestic electricity prices remain elevated by recent historical standards, even as wholesale markets have eased since the 2022 energy crisis.

In a European context, the UK’s cost position weakens its relative location appeal for power-intensive data centres. Wholesale electricity prices across Europe have diverged sharply since 2021, with Nordic and Iberian markets benefiting from stronger renewable penetration and interconnection, translating into materially lower average prices (Fig 4). While wholesale prices are only one component of the delivered tariff, the UK consistently sits towards the higher end of the European range. For developers and occupiers running high-utilisation compute, these differentials feed straight into total cost of ownership, making power cost, alongside latency and connectivity, a decisive factor in location selection.

As connection reform favours ‘ready and needed’ projects, power-cost credibility is becoming a competitive differentiator. Actual electricity outcomes vary widely by site, driven by regional network charges, voltage levels, and contract structures, including the balance between fixed and pass-through costs. In practice, this means that securing capacity alone is no longer sufficient. As the bar rises on deliverability, investors and operators increasingly expect schemes to demonstrate not just a viable connection pathway, but a clear, financeable strategy to manage long-term power costs. Those able to evidence both are materially better positioned to progress in a constrained and more selective UK grid environment.

From queue to strategy: reforming access and reshaping geography

In response, the UK is moving away from a pure ‘first come, first served’ model and towards prioritising projects that are demonstrably deliverable. Connection reform began with consultations in 2023 and was implemented from January 2025, alongside a temporary pause on new transmission applications (with demand-side windows closing in March 2025). Under the reformed approach, connection rights are more tightly linked to real-world deliverability, most notably through strengthened evidence requirements around land control (including defined site boundaries and, at a minimum, multi-year option agreements). Queue ordering is also increasingly influenced by planning status and demonstrable need, with a two-gate framework separating projects that meet the new criteria (Gate 2) from those placed in a holding position without a firm connection date or price (Gate 1).

The practical effect is to raise the bar for data centre developers: power strategy must now be built around secured land, a credible consenting pathway and robust, defensible load assumptions.

Alongside NESO's queue reordering, Ofgem, DESNZ and NESO are implementing a broader Demand Connections Reform programme built around three pillars: Curate, Plan and Connect. Together, these reforms aim to remove speculative projects from the queue, prioritise strategically important demand, including data centres, and accelerate the delivery of viable connections. This is expected to translate into stronger queue entry and progression requirements, including clearer evidence of planning progress, financial backing and land rights, alongside mechanisms that discourage non-credible applications. A further proposal under consideration is requiring applicants to demonstrate committed end-user demand before securing capacity, helping ensure that grid access is allocated to genuine projects rather than land and power speculation. With application windows unlikely to reopen before late 2026, the current pause provides developers with an opportunity to complete technical due diligence, refine power requirements and strengthen commercial arrangements ahead of the next allocation round.

These reforms also sharpen the geographic dimension of UK data centre delivery. In England and Wales, distribution connections can, in some cases, support medium-scale loads, but practical limits mean the largest campuses often require transmission-level access. In Scotland, where transmission responsibilities extend further down the voltage hierarchy, large data centres are more directly dependent on transmission applications, making the current closure of the transmission window particularly constraining.

Looking ahead, further prioritisation is expected to focus on areas where hyperscalers already cluster (often referred to as ‘availability zones’) and on government-designated growth locations, including emerging AI growth zones. This suggests that ‘where to build’ is increasingly intertwined with national infrastructure planning: developers may need to align site choices not only with fibre and latency, but with where grid reinforcement is planned, where policy support is clearest, and where delivery can be evidenced earliest.

To support more proactive decision-making, NESO is also developing a multi-tier spatial energy planning framework spanning local-area energy plans, regional energy spatial plans, and a national strategic plan. The intent is to model the system to 2050 and coordinate reinforcement ahead of need, rather than reacting once connection requests arrive. For data centres, this matters because it can improve transparency around where capacity will emerge and where constraints are likely to persist, information that can be used earlier in the development cycle to reduce aborted site costs, inform master planning and engage planning authorities with a clearer evidence base on infrastructure feasibility.

>>  For more information, please contact our EMEA Data Centre Advisory team

_{Read the articles within Spotlight: UK Data Centre Building Capacity below.}