United Kingdom

The UK is, by installed capacity, the largest data centre market in Europe and the third largest globally. More than £45 billion of private investment has been announced since July 2024 and, in September 2024, data centres were designated as Critical National Infrastructure (“CNI”), the first new CNI sector in nearly a decade. The government’s ambition is to make Britain an “AI superpower”, as set out in the AI Opportunities Action Plan of January 2025, and it has committed to at least 6 GW of AI-capable data centre capacity by 2030 under the UK Compute Roadmap. These aspirations are, however, colliding with a severely congested electricity grid, an evolving planning regime, tightening environmental obligations, and new cybersecurity requirements. What follows is a thematic analysis of the UK position across the four principal dimensions of our comparative framework.

Key contacts

Michael Corbett Partner

James Cook Partner

Daniel Mewton Partner

Simon Bartle Partner

Aaron Wu Knowledge Counsel

Securing grid access: observing the UK's capacity constraints

Grid access is the single greatest constraint on data centre investment in the UK today. The queue for new connections is long, and the rules governing how grid capacity is allocated are undergoing significant reform. For sponsors, investors, and lenders evaluating new developments, understanding when and at what cost grid connection can be secured is a fundamental first step to assess viability.

In Great Britain, customers seeking grid access can connect: firstly, at the transmission level (275kV in England & Wales; 132kV in Scotland) directly by applying to the National Energy System Operator (“NESO”); or secondly, at the distribution level by applying to the relevant Distribution Network Operator (“DNO”) covering the area within which the customer is located. Independent Distribution Network Operator (“IDNO”) and private networks also exist (outside of the general scheme).

Queue reform and prioritising grid capacity

Historically, UK grid connections operated on a “first-come, first-connected” basis. However, the surge in connection applications produced long and congested queues against an already pressured system. That experience is driving fundamental and ongoing reforms, with the UK now transitioning to a “first-ready-and-needed, first-connected” approach, under which projects are assessed on their readiness and strategic alignment, rather than solely on the order in which they applied. The ongoing queue reform process has already filtered out generation and storage projects that failed to meet the relevant criteria, with a consequent meaningful reduction of the generation queue.

The rapid increase in demand-side connection applications, which had not faced the same scrutiny during the initial filtering process, means that further steps are necessary. The Department for Energy Security and Net Zero (“DESNZ”), Ofgem, and NESO are considering a number of significant proposals, including: financial commitment requirements for data centre projects, designed to encourage withdrawals of speculative projects from the queue; reservation and reallocation by DESNZ of grid capacity for strategically important projects; and a Connections Accelerator Service providing engineering and technical support for priority projects. There is no formal statutory fast-track mechanism for data centres, though the proposed capacity reservation and Connections Accelerator Service may in practice prioritise qualifying projects. Separately, the government’s AI Growth Zones programme aims to provide pathways to timely connections for strategically important demand users located in designated Growth Zones.

Flexibility connections, charging structures, and supply options

Demand connections at the transmission level may be offered on either a firm basis (which is contractually guaranteed and not subject to curtailment), or a non-firm basis (which is conditional, and subject to curtailment during system constraints, outages, or pending reinforcement works). DESNZ, Ofgem, and NESO are currently considering whether to expand the use of flexible or non-firm capacity agreements as a mechanism for enabling projects to connect more rapidly.

Taken together, UK grid access for large demand users is becoming more structured, but also more selective. Continuing reforms are designed to prioritise projects that are both ready and strategically important. Nevertheless, those reforms will also require developers to demonstrate greater upfront commitment, and a clearer case for strategic alignment, than the previous first-come first-served regime demanded.

Power procurement: structuring supply in a mature market

The UK offers data centre operators a mature and flexible power procurement regime, with a range of plausible structures and a well-developed market in long-term renewable contracts. Yet, a principal issue for most operators is how to navigate the network charging and regulatory classification consequences that flow from each choice.

Selecting procurement models: structural considerations

Data centre operators in the UK can obtain electricity from the public network under a standard supply contract with a licensed electricity supplier, source electricity directly from one or more grid-connected generators via a bilateral power purchase agreement (“PPA”), or procure power from on-site generation assets or via private wire arrangements from facilities which do not supply them through the grid.

In practice, most operators that rely on the grid go further than standard supply contracts, either by entering into PPAs (“sleeved” by a licensed supplier) to gain price certainty and to ensure green benefits, or as a financial hedging mechanism. Operators opting for virtual PPAs may fall within the definition of a “commodity derivative” under the UK Markets in Financial Instruments framework, meaning that actions in relation to the virtual PPAs may constitute a “regulated activity”, requiring certain authorisations or exemptions. It is, however, possible to structure virtual PPAs so that they fall within the available exemptions and are not considered commodity derivatives.

In January 2026, the government announced a call for evidence on corporate PPAs, in which it sought views from market participants as to how the UK’s corporate PPA market can be developed and improved in order to help industrial and commercial offtakers (including data centre operators) secure stable, competitively priced low-carbon electricity. The government has suggested that its strategy in this area is likely to be informed by international best practice in order to improve competitiveness, increase efficiency, and remove barriers to growth.

There are no material legal restrictions on developing on-site generation assets, although the usual planning consents and environmental permits apply.

Managing network charges: the private wire opportunity

Where power is delivered through the grid, licensed suppliers bear a range of charges: Transmission Network Use of System (“TNUoS”) charges or Distribution Use of System (“DUoS”) charges (as applicable), Balancing Services Use of System (“BSUoS”) charges, and Supplier Obligation levies (including policy costs passed through under the Renewables Obligation and Contracts for Difference (“CfD”) schemes), which they typically pass through to their customers as “non-commodity” components of the supply bill. The Climate Change Levy is also payable on supplies of energy in full unless an exemption or discount applies: renewable electricity no longer attracts a blanket exemption, though a reduced rate may be available in certain circumstances. A principal advantage of using private wire arrangements is that TNUoS, DUoS, and BSUoS charges are not typically payable by customers.

Renewable energy procurement and the UK certificate regime

Data centre operators have become a driving force behind demand in the UK’s corporate PPA market for renewable or low carbon electricity, and often require certificates to be provided by the counterparty to the relevant PPA to demonstrate these low carbon or renewable credentials.

The principal renewable energy certificate scheme is the Renewable Energy Guarantee of Origin (“REGO”) regime, administered by Ofgem under the Electricity (Guarantees of Origin of Electricity Produced from Renewable Energy Sources) Regulations 2003 (as amended), under which one REGO is issued per MWh of eligible renewable electricity generated. There is currently no mandatory additionality requirement, meaning REGOs may be issued in respect of existing as well as new generation, though offtakers may impose contractual additionality criteria in their PPAs (such as requiring that the relevant generation facility achieved grid connection after a specified date). Annual matching remains the standard, but Ofgem and DESNZ have both expressed interest in developing more granular – potentially hourly – matching standards and have consulted on the future architecture of the scheme: a direction of travel that several large technology companies – including Google and Microsoft – have actively advocated. UK REGOs are no longer mutually recognised with EU Guarantees of Origin: the UK is not a member of the European Energy Certificate System, meaning UK REGOs cannot be used to demonstrate renewable origin for EU regulatory purposes, and EU Guarantees of Origin cannot be imported into the UK REGO registry. No bilateral recognition arrangement is currently in place, although discussions remain ongoing.

Large nuclear and SMR developments

In relation to nuclear, there is no regulatory prohibition on bilateral PPAs between data centre operators and nuclear generators. However, in practice, the existing fleet has historically sold output through the wholesale market. Hinkley Point C – the 3.2 GW nuclear power station currently under construction in Somerset – is supported under a bespoke CfD arrangement with a UK government-owned counterparty whose terms limit the scope for direct PPAs during the CfD term. Corporate PPAs with small modular reactor (“SMR”) projects are areas of active commercial interest, with developers including Rolls-Royce SMR and GE Hitachi highly engaged in the UK market.

AI superpower ambitions, fiscal incentives, and planning consent

The UK has aspirations to become a global hub for AI infrastructure. The AI Opportunities Action Plan, published in January 2025 with all 50 recommendations accepted by the government, gives formal expression to that ambition. The investment case for data centre sponsors turns, in practice, on whether planning consent can be secured in reasonable time. This position has improved materially over the past 18 months, although important risks remain.

Fiscal incentives and AI Growth Zones

On fiscal incentives, the permanent full expensing regime (in force from April 2023) allows a 100% first-year deduction for qualifying main-pool plant and machinery, with up to 80-90% of new-build data centre expenditure potentially eligible. Special rate pool assets also attract a 50% first-year allowance. Five AI Growth Zones – at Culham, Blyth Cobalt Park, North Wales, South Wales, and a fifth currently under review – each target at least 500 MW of capacity by 2030, with bespoke electricity price discounts for constrained regions, and further designation expected on a rolling basis.

Critical National Infrastructure designation, cybersecurity obligations, and foreign investment controls

That changes with the Cyber Security and Resilience Bill, introduced November 2025 and expected to receive Royal Assent in 2026, which will designate data centres as essential services under amended NIS Regulations: applying to colocation and other non-enterprise facilities at or above 1 MW of rated IT load, and to enterprise facilities at or above 10 MW, with penalties of up to £17 million or 4% of worldwide turnover. Two features warrant particular attention. The first is the incident-notification cadence – an initial notification within 24 hours and a full report within 72 hours – which broadly tracks NIS2 but diverges from it in definitional detail, leaving operators with both UK and EU sites to run parallel rather than unified reporting playbooks. The second is the Bill's supply-chain dimension: operators will need to embed security testing, audit, and incident-cooperation duties into supplier and colocation contracts.

Alongside this, the National Security and Investment Act 2021 designates “data infrastructure” as one of 17 mandatory notification sectors. In the latest reporting year, the sector accounted for 14% of all notifications, making it the joint fifth most common notification basis. Foreign investment in data infrastructure is not prohibited, but CNI designation has heightened government scrutiny of cross-border acquisitions. UK GDPR imposes no data localisation requirement, although sector-specific rules—including NHS frameworks, FCA/PRA outsourcing and operational resilience rules in the financial sector, and defence networks—support or reinforce demand for UK-sited capacity.

Planning pathways: NSIP reform, fast-track routes, and judicial
review risk

The two most consequential recent planning developments concern the NSIP regime and environmental challenge risk. From January 2026, data centres became eligible projects under section 35 of the Planning Act 2008, creating an opt-in route to the NSIP process, through which a Development Consent Order can consolidate planning permission, compulsory acquisition powers, and ancillary authorisations into a single instrument. The thresholds triggering mandatory NSIP treatment are to be set in a forthcoming National Policy Statement for Data Centres, with no consultation having commenced as at the date of this publication. The Planning and Infrastructure Act 2025 supports the introduction of a new fast-track route for certain NSIPs, intended to shorten the acceptance-to-decision stage from an average of around 17 months under the standard process to a non-statutory target of up to 12 months for eligible projects.

Sponsors of large-scale projects should weigh the NSIP route with care: although the consolidated consent mechanism offers genuine advantages, uncertainty as to the threshold and the nascent National Policy Statement leave significant questions unresolved. The Foxglove-led judicial review of January 2026 – in which the government conceded that a 90 MW facility had been consented without adequate EIA screening – marks the first successful environmental challenge to a UK data centre consent and indicates that EIA compliance may attract heightened scrutiny. Below the NSIP threshold, the revised National Planning Policy Framework of December 2024 requires planning policies to “pay particular regard” to data centre needs, and Green Belt policy has been materially relaxed through the “grey belt” concept: three recovered appeals determined by the Secretary of State in 2024-25 each approved hyperscale campuses on Green Belt land, though one (Woodlands Park) remains subject to a live judicial review.

Environmental obligations, efficiency, and disclosure

The most striking structural feature of the UK’s environmental framework for data centres is what it lacks: unlike the EU, which imposes mandatory power usage effectiveness and energy performance reporting on data centres through Article 12 of the recast Energy Efficiency Directive, the UK has no equivalent regime. This is a significant point of divergence for global operators and investors. The UK framework is instead built upon a combination of voluntary participation, sectoral fiscal incentives, as well as general environmental regulation and licensing, overlaid by an evolving and diverging corporate sustainability disclosure regime that may continue to tighten through to 2028 and beyond.

Energy, emissions, and water: operational constraints and incentive structures

The primary energy efficiency mechanism is the Climate Change Agreement for data centres, administered through techUK, in which over 170 sites participate. Participation delivers an approximately 93% discount on the Climate Change Levy against a 14.5% efficiency improvement target measured from a 2022 baseline to 2030. While the Climate Change Agreement applies on a voluntary basis incentivised by the discount, the UK also has a general energy efficiency regime – the Energy Savings Opportunity Scheme – which applies to any undertakings with an annual turnover in excess of £44 million and an annual balance sheet total in excess of £38 million. In-scope undertakings are required to collect information relating to their energy consumption, carry out energy audits and implement action plans to improve energy efficiency.

While the UK ETS does not traditionally apply to the data centre sector, it does apply to power generation activity, which means that installations ancillary to data centres with thermal combustion capacity at or above 20 MW (a threshold typically triggered by aggregate backup diesel energy generation of around 7.5 MW) may nonetheless fall within scope of the UK ETS, which carries significant potential cost implications.

Water is an emerging operational constraint: generally, abstraction exceeding 20 m3 per day requires an Environment Agency licence under the Water Resources Act 1991. Moreover, the first AI Growth Zone at Culham sits within a “seriously water stressed” catchment, a consideration that sponsors of large AI facilities in that zone must address from the outset. Additionally, an Ofwat consultation in August 2025 relating to water efficiency and wholesale charges for business customers has indicated that the government intends to revise the current charging structure for wholesale water supply which currently discounts charges for large-scale water users, a policy direction affirmed by the more recent Water White Paper released in January 2026. This suggests a high likelihood of increased costs for water usage in the foreseeable future, intended to drive a focus on water efficiency.

Alongside these requirements, the UK heat network and waste frameworks contain further latent obligations. The Heat Networks (Market Framework) (Great Britain) Regulations 2025, in force from 27 January 2026, regulate the supply of waste heat, including by data centres. Primary legislation, notably the Energy Act 2023, will also enable future heat network zoning that could require large non-domestic buildings to supply waste heat into local networks. No such supply obligation currently applies. Operators may also be subject to the waste duty of care under the Environmental Protection Act 1990 in respect of certain electrical equipment once it becomes waste. No data centre-specific circular economy obligations currently exist, although the Environment Act 2021 carries enabling powers for extended producer responsibility schemes that could eventually reach IT hardware.

The trajectory of sustainability disclosure

While the UK imposes no sector-specific sustainability disclosure requirements on data centres, UK requirements are likely to compel increasingly detailed and demanding disclosure over time. UK Sustainability Reporting Standards S1 and S2, adopted on 25 February 2026 and based on the International Sustainability Standards Board’s International Financial Reporting Standards S1/S2 framework, adopt a single materiality approach. They also encourage disclosure in accordance with the SASB standards, including the Technology & Communications standards. This diverges from the Corporate Sustainability Reporting Directive’s double materiality approach, an important difference to note for operators with reporting obligations in both jurisdictions.

Under Financial Conduct Authority (“FCA”) proposals, mandatory climate disclosure under UK Sustainability Reporting Standard S2 (Climate-related Disclosures) is expected for listed companies from accounting periods beginning 1 January 2027, with the UK government intending to consult later this year on introducing similar disclosure requirements for large unlisted companies. The Streamlined Energy and Carbon Reporting framework already applies to quoted and large unquoted companies, while the parallel energy use reporting obligations under Phase 4 of the Energy Savings Opportunity Scheme fall due by December 2027.

The UK imposes no sector-specific sustainability disclosure requirements on data centres. As we have seen in previous chapters, the EU adopts a different approach to the data centre sector. For data centre sponsors and investors operating across both markets, this asymmetry creates an immediate practical question: whether to voluntarily align UK reporting with EU standards, or to maintain distinct frameworks. As the UK’s disclosure regime matures and diverges further from EU requirements, that issue may become harder to defer.

Site selection and real estate structuring

The real estate workstream of developing a data centre may determine the timetable for a UK data centre project, with the acquisition of appropriate land one of the primary considerations.

Securing development land

Selecting an appropriate site for a data centre project is not an easy task. As discussed in the foregoing sections, the site must fit a number of physical and regulatory constraints that are largely outside a developer’s control: firstly, proximity to a suitable grid offer, substation with available capacity or some other form of power; secondly, sufficient cooling water for operating the data centre; and thirdly, a land area large enough to house the scale of the planned project, and any potential expansion plans in the future. Consideration as to whether planning consent would be granted for the area is also required. In general, suitable land for data centre development in the UK is scarce, with competing demands from housing, energy infrastructure, farming, and other industry. Brownfield and other former industrial sites may be fruitful ground for developing data centres, however they may bring heightened risk of environmental remediation work requirements, which will require additional due diligence and potentially increase costs.

Freehold, leasehold, and long-term control

Many large data centre sites are held as freehold properties, meaning the developer owns both the building and the land it stands on, for an unlimited period of time. There is no landlord and no expiry date. The primary alternative way to hold land in England and Wales is through leasehold structures, where the proprietary right is derived from a lease, which grants exclusive possession of a property for a defined term, subject to conditions. Ultimate ownership remain with the owner of the freehold, and, unless extended, the property reverts to the freeholder when the lease expires.

The choice between freehold acquisition and leasehold structures is rarely a pure legal question: it turns on the developer’s capital strategy, the landowner’s willingness to dispose, the availability of debt finance against each tenure, and the long-term operational flexibility required. Freehold ownership maximises operational autonomy, provides security of tenure and preserves residual asset value for developers, which is significant given typical 25 to 30 year operational lives for hyperscale facilities. Long leasehold structures (commonly 125 years or longer, at peppercorn or geared rents) are nevertheless prevalent where the underlying landowner is a public body, a strategic landowner, or an institution unwilling to dispose of the freehold; in such cases, particular attention should be paid to provisions of the lease so it provides the desired flexibility for the data centre project and the necessary security of tenure.

Securing easements and managing restrictive covenants

Data centre sites depend on a network of third-party rights over neighbouring land for power, water, drainage, fibre, and vehicular access, all of which are essential for their daily operations. To enact these rights, careful consideration of easements may be needed. An easement, in English law, is a non-possessory property right giving its holder a defined right of use over land owned by another person. The land subject to the easement remains fully owned by the landowner, but the use of this is limited by the teams of the easement. Typical examples include a right to use water pipes or drains, use gas pipes or electricity wires, or pass over a path, road, or way.

Developers should therefore ensure that any selected site has the necessary easements and other access rights required for the data centre to function and, if not, how likely it would be for neighbouring landowners to grant the necessary access rights. Where a site is being acquired out of a larger estate that will become the neighbouring land, the developer should ensure that the transfer of the site grants these rights. In such circumstances, the developer should also consider whether it needs to restrict the use of the neighbouring land – for example, to prevent uses that could disrupt its operation as a data centre – and, if so, seek to include restrictive covenants in the transfer.