Increased AI Centres Will Need Massive amounts of Electricity and Water

18th September 2025

The UK's booming AI infrastructure—especially hyperscale data centres will demand staggering amounts of electricity and water. Lots of new jobs but there is a cost.

Here's a breakdown of what experts and government are reporting.

Electricity Demand
UK Data Centre Capacity (2024) - 1.6 gigawatts (GW) and rapidly expanding.

Projected AI Energy Use (Global)
Over 1,000 terawatt-hours (TWh) by 2026 more than double 2022 levels.

UK Impact: AI data centres could consume tens of TWh annually, rivalling the electricity use of millions of homes.

Example: A single 314MW hyperscale centre in Thurrock would draw enough power to supply over 250,000 UK homes.

Water Usage
Cooling Systems
AI servers require intensive cooling, often using evaporative systems that consume vast amounts of water.

Global AI Water Demand
Could reach billions of cubic metres annually, posing risks to water security.

UK Concerns:

Many new AI hubs are in water-stressed regions like South East England.

Labour’s AI Growth Zone in Culham is just 7 miles from the proposed Abingdon reservoir, raising fears of competition between public and industrial water needs.

Sustainability Challenges
Carbon Intensity: UK electricity is cleaner than ever (124 gCO₂/kWh in 2024), but AI workloads must shift to low-carbon hours to stay green.

Policy Gaps
Current water resource plans don’t fully account for AI’s rising demand, risking future shortages.

Recommendations:

Mandatory location-based reporting.

Incentives for water-efficient cooling tech.

Integrated planning of land, energy, and water.

The bottom line
AI’s promise comes with a hefty environmental price tag. If the UK wants to lead in AI without compromising its Net Zero goals or water resilience, it’ll need bold regulation and smarter infrastructure planning.