The Water
Question.
84% of water-intensive sites are stressed — and water is no longer a question data centres can outsource.
This is part of the paid GridteamAI Quarterly Sector Report.
£295 single volume · £995/yr annual · Team £2,495 · Enterprise £4,995 · Corporate £9,995. Secure checkout coming soon; email to purchase today.
84% of water-intensive sites are stressed — and water is no longer a question data centres can outsource.
The Environment Agency’s National Framework for Water Resources 2025 is the principal English water-resources strategy, and it singles out data centres as a sector requiring active attention — alongside agriculture, hydrogen production and the electrification of heat. Approximately 84% of water-intensive UK data centre developments are proposed for areas already water-stressed or projected to be so by 2040.
Slough — Europe’s densest data centre cluster — sits in an area the Environment Agency formally classifies as "seriously water stressed": the single most acute UK case. Thames Water estimates that under a high-growth scenario data centres would drive about 30% of new regional water demand — around 270 million litres per day, equivalent to the household consumption of ~1.85 million people.
The sector-wide UK water-consumption total is unknown — the Environment Agency calls reliable data "vital for long-term planning", and the most-cited industry survey covered only 73 of ~500 operational UK facilities. That data gap is itself the policy problem; mandatory reporting is the likely first-order intervention.
The air-to-liquid cooling transition (Gap 19) changes — but does not eliminate — water exposure. Air-cooled evaporative towers consume substantial make-up water, peaking at summer ambient extremes when catchment stress is also highest; closed-loop direct liquid cooling consumes less in volume but introduces procurement-sensitive make-up requirements. It is a profile change, net-positive, but not a water "solution" at hyperscale density.
AI Growth Zone applications must include a statement from the local water supplier confirming sufficient access to water — the first explicit UK water-availability planning condition at zone level, making water companies formal gatekeepers.
UK data centre water is no longer a routine procurement matter. It is a strategic infrastructure consideration with policy, planning and reputational dimensions — and civil-society legal challenge is now part of the planning environment. The forward outlook is about the pace and detail of the framework response, not whether the response happens.
The shape of this volume. Section 1 sets out the headline figures. Section 2 develops twelve findings, each confidence-annotated. A chapter-weight pull at p. 14 anchors the argument. Section 3 sets out the stress overlap and the National Framework. Section 4 develops the Thames quantification and the cooling link. Section 5 isolates the policy stack and the operational responses. Section 6 makes recommendations, with a one-page tear-out at p. 25.
The overlap is structural, not incidental: fibre proximity, power availability and skilled-labour access all concentrate DCs in the same population-dense, water-stressed geographies. Slough is the acute case — "seriously water stressed".
The 84% overlap (ink) is structural, not coincidental. Slough is the single most acute UK water-stress concentration.
HIGH — verified against an Environment Agency, government or water-company primary source. MEDIUM — named industry, civil-society or analytics source cross-referenced to a primary source. LOW — GridteamAI scenario projection or interpretation, marked as such.
Findings 01 – 12 · summary
Findings 01–06 in detail · pp. 11–13 · 07–12 summarised above
The Environment Agency’s National Framework for Water Resources 2025 — water for growth, nature and a resilient future — is the principal English water-resources strategy published in 2025. Section 9, Taking action on other significant water-using sectors and emerging demands, explicitly singles out data centres, positioning them alongside agriculture, hydrogen production and the electrification of heat as sectors whose demand pathways require coordinated planning rather than ad hoc accommodation. The Framework is non-statutory but materially influential: water companies’ Water Resources Management Plans, EA licensing decisions and DEFRA policy positions all reference it as the operative baseline. The implication is that water has been formally elevated from a routine procurement matter to a strategic infrastructure consideration — the most material UK water-policy development affecting the sector this decade.
The Environment Agency has stated that reliable data on UK data centre water use is "vital for long-term planning" and is not currently available; the sector-wide total is unknown. The most-cited data point — the techUK industry survey — covered only 73 of approximately 500 operational UK facilities, and the 73 were self-selecting, introducing bias that almost certainly understates sector-average consumption. This gap is itself the policy problem: the EA cannot model future demand, planning authorities cannot assess catchment-level cumulative impact, and water companies cannot plan supply augmentation against a precise forecast. Closing it is the first-order intervention — mandatory water-consumption reporting for facilities above a threshold IT load, comparable to the Energy Performance of Buildings treatment of large commercial buildings, is now a likely policy direction.