Written submission from Dr Anandadeep Mandal (DCU0008)

Written evidence submitted to the House of Commons Environmental Audit Committee

Risks and opportunities to the sustainability of data centres in the UK

27th February 2026

Anandadeep Mandal (PhD)

Associate Professor (Mathematical Finance), Scotcoin Distinguished Chair of Digital Finance

Honorary Research Associate at UCL Centre for Blockchain Technologies

Birmingham Business School, University of Birmingham, UK

Executive summary

• Data-centre demand in Great Britain is being accelerated by artificial intelligence (AI) and cloud computing, translating into unprecedented large-load connection requests. Ofgem and NESO identify around 140 data-centre projects totalling around 50 GW in the demand queue, with around 71 projects (around 20 GW) reporting Final Investment Decision (FID). [R1]
• Grid connection constraints are a first-order sustainability risk: Ofgem reports the connections queue reaching 701 GW (April 2024) with connection offers extending into the late 2030s, and notes that non-viable projects can block strategic investment and distort network signals. [R2, R1]
• Electricity demand growth is material in the 2030 horizon: NESO’s Clean Power 2030 analysis assumes around fivefold growth in data-centre electricity demand by 2030 (driven by high AI development). [R3]
• Water impacts are locally concentrated. The House of Commons Library briefing cites Thames Water evidence that a large data centre could use roughly 4 to 19 million litres per day (comparable to around 50,000 households), and that Thames Water supplied 9.95 billion litres to 140 data centres. [R4]
• Policy coherence is needed across planning, electricity connections, and water regulation. The Environmental Improvement Plan (EIP) sets ambitions such as 50% leakage reduction and 15% reduction in non-household water use by 2050, which sit in tension with rapid expansion of potable-water cooling loads unless strong efficiency and alternative-water requirements are applied. [R5]
• Recommendations include: (i) treating large data centres as ‘strategic demand’ with gating and milestones for connection offers; (ii) mandatory disclosure of energy and water metrics; (iii) siting requirements aligned to water-stress and grid constraints; (iv) incentives/conditions for waste-heat reuse; and (v) robust carbon accounting (including embodied carbon) and 24/7 clean energy matching.

Scope and approach

This submission provides evidence on the sustainability implications of data centres in the UK, focusing on electricity-system impacts (connections, system planning and decarbonisation), water use and local environmental pressures, and policy design across planning, net zero governance and utility regulation. It draws primarily on UK institutional sources (Ofgem, NESO, Government publications, and House of Commons Library briefings), supplemented with international framing where clearly labelled.

Evidence mapped to the Terms of Reference

1. Drivers of demand, and opportunities and challenges

1.1 AI and cloud adoption are driving large-load connection requests. Ofgem’s demand-connections work explicitly links significant growth in data centres to investment in AI and reports an ‘unprecedented’ increase in demand-side connection requests. [R1]

1.2 The scale of connection appetite is large relative to current grid constraints: Ofgem and NESO identify around 140 data-centre projects totalling around 50 GW in the demand queue, with around 20 GW at FID. [R1]

1.3 Opportunities: data centres are essential digital infrastructure underpinning AI deployment, productivity and service innovation, but benefits depend on aligning growth with power-system decarbonisation and local environmental capacity. [R4, R3]

1.4 Challenges: cluster effects can create acute local network constraints and local water-stress impacts even when national shares appear modest. [R4]

2. Environmental impacts now and in the future (energy, water, nature)

2.1 Electricity demand and emissions: operational emissions depend on marginal electricity generation and the pace of grid decarbonisation. System-level impacts also include network reinforcement, queueing and potential displacement of other electrification priorities. [R1, R2, R3]

2.2 Water: impacts are primarily local and depend on cooling design and water source. The Commons Library briefing cites Thames Water’s estimate that a large data centre might use between 4 and 19 million litres per day, comparable to the daily demand of around 50,000 households, and reports 9.95 billion litres supplied to 140 data centres. [R4]

2.3 Nature and local environment: planning impacts include land take, biodiversity net gain delivery, noise and air-quality impacts from standby generation, and cumulative effects in clustered hubs. These are typically assessed site-by-site, raising risks of inconsistent mitigation across regions. [R4]

3. Short, medium and long-term projections

3.1 To 2030: NESO’s Clean Power 2030 analysis assumes around fivefold growth in data-centre electricity demand by 2030, driven by high AI development and offsite computation. [R3]

3.2 Connection pipeline as an upper bound: the demand queue contains around 50 GW of data-centre projects, indicating the potential scale of growth if projects progress and grid capacity is delivered. [R1]

3.3 To 2050: Government planning assumes economy-wide electrification drives at least a doubling of annual electricity demand by 2050; data-centre growth therefore competes with other strategic electrification loads for grid capacity. [R6]

3.4 Water-resource constraints tighten over time: Government analysis highlights a projected England public supply deficit by 2050 (order of 5,000 Ml/day), implying that new industrial potable-water demands should be evaluated against regional water-resource plans and drought resilience. [R7]

4. Impacts on net zero and Carbon Budgets

4.1 Data-centre growth affects net zero via (i) additional electricity demand, (ii) potential delays to other decarbonisation investments if grid access is constrained, and (iii) local environmental constraints (water) that can affect feasibility of wider transition plans. [R1, R2, R7]

4.2 Ofgem notes that non-viable projects can block important projects and create inaccurate signals for network investment, which is directly relevant to delivering carbon budgets on time. [R1]

5. AI acceleration and adequacy of governance (Government, CCC, OEP)

5.1 Ofgem identifies AI investment as a key driver of data-centre demand growth. NESO’s Clean Power 2030 analysis incorporates an assumption of high AI development driving large growth in data-centre demand. [R1, R3]

5.2 CCC has noted data-centre demand is uncertain and that its advice uses Government projections in the baseline, implying the need for transparent sensitivity analysis and explicit treatment of local constraints (water stress and clustering). [R8]

6. Adequacy of existing policies (EIP, planning and infrastructure policy)

6.1 The Environmental Improvement Plan sets ambitions including 50% leakage reduction and 15% reduction in non-household water use by 2050, which suggests a policy need to constrain potable-water cooling where alternatives exist and to mandate high water-efficiency designs. [R5]

6.2 Planning policy is evolving toward treating data centres as strategic infrastructure; the Commons Library briefing indicates a move toward more national direction (including a forthcoming National Policy Statement in 2026). Sustainability outcomes will depend on embedding clear standards for water, energy, carbon, and heat reuse within national planning guidance. [R4]

7. How impacts should be factored into future policy (Land Use Framework, regional planning, housing, water reform)

7.1 Electricity: connections reform should include demand-side gating for large loads, with milestones (e.g., land, permits, financing) to prevent speculative projects from occupying queue capacity. Ofgem’s evidence on queue size and demand-offer growth supports this. [R1, R2]

7.2 Water: site selection and permitting should align with regional water-resource management plans, require non-potable or recycled water where feasible, and require transparent public reporting of water consumption and discharge. [R4, R7]

7.3 Place-based planning: local cumulative impacts in clustered regions justify regional capacity caps or sequential tests (grid capacity, water stress, heat off-take) before consent for additional facilities. [R4]

8. Location and optimum siting

8.1 Location is central to sustainability outcomes. Key factors include (i) deliverable grid connection and reinforcement needs, (ii) regional water stress and drought risk, (iii) proximity to credible waste-heat off-takers and heat networks, and (iv) cumulative local environmental impacts. [R1, R4, R7]

8.2 Practical implication: encourage siting in regions with surplus grid capacity and lower water stress, coupled with commitments to heat reuse and flexible connections where appropriate. [R1, R4]

9. CCC treatment of data centres in Seventh Carbon Budget advice

9.1 CCC correspondence indicates its baseline uses Government projections and that data-centre demand is inherently uncertain, implying the Committee should request published sensitivity ranges and explicit accounting of high-growth/cluster scenarios. [R8]

9.2 Given Ofgem’s evidence of large connection appetite (around 50 GW), CB7 sensitivities should test materially higher demand and local constraints, not only national averages. [R1]

10. Technologies to minimise environmental impacts and readiness

10.1 Electricity and carbon: (i) 24/7 clean energy matching (hour-by-hour) rather than annual MWh matching; (ii) flexible or phased connections; (iii) higher-efficiency compute and cooling designs; and (iv) on-site storage and demand response where technically and contractually feasible. [R4, R1]

10.2 Water and cooling: closed-loop cooling, non-potable water use, and water-circularity approaches can reduce withdrawals, but benefits depend on site conditions; therefore standardised reporting and enforceable design requirements are needed. [R4, R7]

10.3 Heat reuse: waste-heat recovery can create local benefits where heat networks and off-takers exist, but requires coordinated planning and financeable heat offtake arrangements. [R9, R10]

11. Role of renewables

11.1 Renewables reduce carbon footprint most effectively when additional, locationally aligned, and (ideally) temporally matched to consumption. The Commons Library briefing highlights the limitations of annual matching for claims of decarbonisation. [R4]

12. Community benefits: powering and heating local communities

12.1 Waste-heat reuse can contribute to heat networks and local heating, but delivery depends on (i) proximity to heat demand, (ii) infrastructure investment, and (iii) planning conditions and commercial structures that secure long-term offtake. London policy work provides a practical pathway for heat reuse. [R9]

12.2 UK case evidence: Queen Mary University of London has announced reuse of data-centre waste heat to provide hot water and heating for its campus, demonstrating feasibility in institutional settings. [R10]

13. Lessons from outside the UK

13.1 Ireland illustrates precautionary lessons on concentration of data-centre demand and system exposure, as summarised in the Commons Library briefing. [R4]

13.2 Nordic examples highlight the value of district-heating integration for heat reuse, but depend on local heat-network maturity; this underlines the need for UK heat-network development if similar benefits are expected. [R9]

14. Competition for resources and impacts on other sectors

14.1 Electricity: large-load queue volumes can crowd out or delay other strategic electrification and low-carbon projects where connections are constrained. Ofgem highlights the risk of non-viable projects blocking important investments. [R1, R2]

14.2 Water: projected water-supply deficits and EIP demand-reduction ambitions mean that additional industrial potable-water demand can displace water availability for households and other industrial decarbonisation priorities unless alternative water sources and high-efficiency designs are mandated. [R5, R7]

Recommendations

1. Introduce ‘strategic demand’ rules for large data centres in connections reform: require milestones (e.g., land rights, planning progress, financing, equipment orders) and enforceable drop-out provisions to prevent speculative queue blocking.
2. Mandate public reporting for large data centres: electricity consumption, peak demand, Power Usage Effectiveness (PUE), water usage (withdrawal, source type, discharge), cooling technology, and backup generation testing/emissions.
3. Adopt siting standards that integrate: (i) grid deliverability, (ii) water-stress screening and drought resilience, (iii) biodiversity and local environmental capacity, and (iv) credible waste-heat reuse plans where feasible.
4. Require robust carbon accounting: include embodied carbon for new builds and require credible decarbonisation claims (preferably 24/7 clean energy matching rather than annual matching alone).
5. Create policy incentives and planning conditions for heat reuse into heat networks or anchor loads (hospitals, campuses, housing), including support for heat-network infrastructure where it unlocks community benefit.

References

R1: Ofgem (2026) Demand connections - Call for Input (12 February 2026). Available at: https://www.ofgem.gov.uk (accessed 27 February 2026).

R2: Ofgem (2024) Connections reform - Open letter (April 2024). Available at: https://www.ofgem.gov.uk (accessed 27 February 2026).

R3: National Energy System Operator (NESO) (2024) Clean Power 2030 analysis: assumptions and pathways including data-centre demand growth. Available at: https://www.neso.energy (accessed 27 February 2026).

R4: House of Commons Library (2024) Data centres - environmental impacts, electricity and water use (briefing paper CBP-10315, 15 November 2024). Available at: https://researchbriefings.files.parliament.uk (accessed 27 February 2026).

R5: Defra (2023) Environmental Improvement Plan 2023. Available at: https://www.gov.uk (accessed 27 February 2026).

R6: Department for Energy Security and Net Zero (DESNZ) (2024) Clean Power 2030 Action Plan (main report) - electricity demand outlook to 2050. Available at: https://www.gov.uk (accessed 27 February 2026).

R7: UK Government (2025) Water use in AI and data centres (report). Available at: https://assets.publishing.service.gov.uk (accessed 27 February 2026).

R8: House of Commons Environmental Audit Committee (2025) Correspondence with the Climate Change Committee on Seventh Carbon Budget and data-centre demand uncertainty. Available at: https://committees.parliament.uk (accessed 27 February 2026).

R9: Greater London Authority (2025) Optimising Data Centres in London - Heat Reuse. Available at: https://www.london.gov.uk (accessed 27 February 2026).

R10: Queen Mary University of London (2024) Data centre waste heat to provide hot water and heating for campus (news release). Available at: https://www.qmul.ac.uk (accessed 27 February 2026).