Nuclear energy is entering a critical deployment phase in the UK. Driven by energy security imperatives, industrial decarbonisation obligations, and the emergence of large-scale firm-power-demanding loads — both digital and industrial — the conditions for a new era of nuclear build are more favourable than at any point in the past three decades. With the latest regulatory developments, Small Modular Reactors (SMRs) and other advanced nuclear technologies are emerging as a promising solution to meet the energy demands of data centres in the UK, offering low-carbon, reliable power while addressing sustainability challenges. This article by Tania Arora and James Wyatt, Partners at Baker McKenzie look at the recent UK nuclear regulatory developments through the lens of digital infrastructure market players.
Advanced nuclear solutions for data centers
Data centre developers (and in particular, hyperscalers) require firm, low-carbon, 24/7 baseload power at a scale that renewables and storage systems alone cannot reliably deliver.
While long term power purchase agreements help manage costs, they do not resolve the challenge of intermittency of renewables. Battery storage helps alleviate some of the challenges, and its technologies are continually improving. Grid-scale battery storage is well-suited to smoothing intra-day variability – this is already a commercially viable and fast growing market. However, despite these advancements, currently storage cannot bridge multi-day and multi-week generation gaps at the scale required. Long-duration storage technologies — flow batteries, compressed air, green hydrogen round-trip — are being developed but are not commercially deployable at the required scale within project planning horizons. For example, the storage requirement to bridge a 500MW industrial or digital campus through a five-day dunkelflaute is approximately 60GWh. At current LCOS (Levelised Cost of Storage) for grid-scale lithium-ion of approximately £100–120/MWh, the capital cost of that storage alone approaches £6 –7 billion — before any generation asset.
Furthermore, data centre operators procuring renewables at scale remain dependent on the grid for delivery. Grid reinforcement queues across Europe now extend seven to twelve years in many locations. A campus relying on grid-delivered low-carbon power faces the same transmission constraints as any other large load — regardless of how many PPAs it holds. The UK’s new AI Growth Zone policy framework (described below) could compress connection timelines for designated projects, but it does not eliminate grid dependency for renewables-only campuses, and it does not resolve the fundamental capacity constraint on the transmission system.
Nuclear is the only technology that resolves all of the above constraints simultaneously:
- Capacity factor: consistently above 90%, compared to 25–40% for wind and solar. This is a physical characteristic of fuel-based generation that intermittent sources cannot replicate.
- 24/7 availability: no intermittency exposure — power is available regardless of weather, season, or grid conditions.
- Load-following: newer SMR designs (including the BWRX-300) offer genuine load-following capability, enabling output adjustment to match variable campus demand profiles across multiple simultaneous offtakers.
- Modular scaling: unit sizing from 300MW to 1.1GW+ allows phased capacity addition aligned with campus or industrial load growth, reducing upfront capital commitment and allowing demand validation before full fleet deployment.
- Land footprint: an SMR delivering 300MW requires orders of magnitude less land than an equivalent wind or solar installation, enabling genuine industrial-site or urban-adjacent deployment.
Key challenges for SMR deployment in the data centre sector
While the benefits of SMRs (and other advanced nuclear technologies) and their use case for data centers are highly attractive, a number of challenges remain to be addressed before SMRs are become reality in the data centre industry.
Firstly, SMRs are as yet unproven at scale. While a number of SMR designs are progressing through the licensing process in a number of jurisdictions, the uncertainties and risks associated with first of a kind (FOAK) projects make the case for SMRs more complex when compared to other power generation sources. FOAK nuclear construction carries schedule, cost overrun and vendor performance risks that no data centre operator or hyperscaler can absorb on its balance sheet.
Furthermore, construction risk is also the binding constraint on private financing. FOAK nuclear projects are structurally late and over budget across every jurisdiction and every design. No private financing structure closes without explicit, calibrated sovereign capital participation in the construction overrun band.
Finally, most jurisdictions lack a regulatory framework adapted to SMR characteristics. Traditional regulatory regimes developed for conventional gigawatt scale reactors that focus on site-specific licensing, large-scale construction oversight and complex emergency planning are not wholly appropriate to SMRs where e.g. development risk, operating profiles and different safety approaches (e.g. through passive safety systems) are different.
As is the case with many new technologies before they are deployed at scale, FOAK SMR projects are likely to become commercially viable only with some level of state support.
The UK’s approach to support
The UK’s nuclear policy history is dominated by two revenue mechanisms: the Contract for Differences and the Regulated Asset Base. Both were designed to solve a specific problem — merchant price risk and political exposure to wholesale electricity markets. In a data centre-anchored SMR structure, that problem does not arise.
Long-term availability-based or capacity-based payments deriving ultimately from a data centre operator or hyperscaler tenant replace the revenue stabilisation function of CfD and RAB. Senior lenders assess counterparty creditworthiness, not wholesale price volatility. The data centre operator or hyperscaler performs the role that government performed under Hinkley’s CfD and under Sizewell C’s RAB — without the public balance sheet exposure, and with a counterparty that commercial lenders better understand.
This reframes the UK discussion. The question for government is not how to support the revenue model. The question is how to support the delivery model — specifically, how sovereign capital participates in construction-phase risk in a way that is structured, bounded and politically defensible.
UK regulatory developments tackling some of the challenges of SMR deployment for data centres
Recent regulatory developments signal that the UK Government has decided it wants to facilitate transformational data centre / nuclear infrastructure build-out, and has begun aligning its planning, grid and nuclear regulatory machinery accordingly.
Delivering AI Growth Zones
The AI Growth Zones framework (published 13 November 2025) identifies specific geographic zones where the UK Government will actively intervene to accelerate delivery of AI infrastructure. The intervention comprises: reallocation of freed grid capacity to designated AI projects rather than returning it to the general connection queue; capacity reservation mechanisms at substations serving identified geographic zones; compression of planning timescales for Nationally Significant Infrastructure Projects within those zones; and streamlined engagement between NESO, DNOs and project developers to reduce the timeline from connection offer to energisation.
Advanced Nuclear Framework
The Advanced Nuclear framework (published 4 February 2026) is a formal framework for the deployment of advanced nuclear technologies in the UK. This framework includes a number of provisions relevant to privately-anchored data centre projects. First, it establishes a pipeline coordination mechanism — managed through Great British Energy – Nuclear — that allows privately led projects to engage with the regulatory pathway in a structured and pre-competitive way before formal Generic Design Assessment (GDA) entry. Second, it signals government appetite for privately-financed advanced nuclear, explicitly anticipating structures where the revenue anchor is a private offtaker rather than a government-supported price mechanism (CfD or RAB). Third, it identifies the UK-US cooperation frameworks — including cross-recognition of reactor assessment work — as a tool for reducing the duplication that has historically made UK GDA so time-consuming for US-designed reactors.
National Policy Statement EN-7 – Nuclear as Nationally Significant Infrastructure
The updated National Policy Statement for Nuclear Energy Generation (EN-7) (published 12 November 2025) confirms that new nuclear generating stations are nationally significant infrastructure projects within the meaning of the Planning Act 2008. The policy statement requires decision-makers to give substantial weight to nuclear’s role in energy security, decarbonisation and economic resilience.
While EN-7 is primarily drafted with grid-connected nuclear in mind, its significance for co-located SMR deployment is indirect but material. First, it embeds nuclear firmly within the national infrastructure priority category, strengthening the case for expedited Development Consent Order (DCO) treatment. Second, it reinforces the presumption that nuclear development aligned with energy security and decarbonisation objectives should be facilitated rather than obstructed.
In the context of AI Growth Zones, the interaction between EN-7 and designated AI infrastructure sites creates the possibility of dual-priority treatment: strategic AI infrastructure coupled with nationally significant nuclear generation.
Regulatory Proportionality and the ALARP/BAT Reform Agenda
In November 2025, the Department for Energy Security & Net Zero (DESNZ), the Ministry of Defence, industry and the statutory regulators published “Ways of Working – principles to guide the application of ALARP and BAT in the nuclear industry”. The document does not amend the legal duties of operators. It does something more important: it articulates a shared regulatory and industry position that evidencing compliance must be proportionate to actual risk, that conservative modelling must not accumulate beyond what is justified by real-world exposure, and that there may be more than one compliant solution where risk is properly optimised.
For advanced nuclear projects, this is not cosmetic guidance. Historically, FOAK deployment has been slowed not only by engineering challenge but by iterative and sometimes over-conservative evidencing cycles. The 2025 framework explicitly addresses the question of “when enough evidence is enough” and promotes early engagement, graded governance proportional to risk, and the recognition of relevant good practice and replication arguments.
For a data centre-anchored SMR programme, this materially reduces one of the structural uncertainties in UK deployment: regulatory sequencing risk driven by excessive evidencing burdens. The legal standard remains unchanged. The regulatory culture in applying it is being deliberately recalibrated.
Conclusions
Taken together, these regulatory developments do not resolve all of the delivery challenges described in this article. Construction risk remains and regulatory approvals can still take years. However, they do confirm that the UK Government is an active participant in proactively facilitating advanced nuclear technologies, including in the context of digital infrastructure — rather than operating as a passive regulator. That presents a more supportive environment and is the precondition for the sovereign capital engagement that the construction-phase risk structures depend on.
Baker McKenzie’s nuclear practice is positioned to advise across the full architecture of this deployment wave: from regulatory structuring and capital stack design, through sovereign capital engagement and grid integration, to the commercial and compliance architecture of the energy campuses that will define the next generation of digital infrastructure.
