Lords demand 18-month AI energy efficiency strategy as data centres strain UK grid and water resources

UK peers push for a national AI energy efficiency strategy within 18 months

Parliamentarians have called for a national strategy to manage the energy and water footprint of AI infrastructure as data centres expand across the UK. A proposed amendment to the Planning and Infrastructure Bill would make the energy use and efficiency of AI-related infrastructure a statutory planning consideration, supported by guidance and reporting duties.

The aim: ensure AI delivers public benefit without overloading local grids or water systems. The proposal also seeks to capture surplus heat and power from data centres and feed it back into the system during periods of high demand.

What the amendment would do

• Require the Secretary of State to publish a national AI energy efficiency strategy within 18 months.
• Treat projections of energy use and efficiency measures as material planning considerations for AI-related developments.
• Mandate developers to account for both supply and efficiency in planning applications, including options for heat reuse and demand-side flexibility.
• Enable surplus energy from data centres to be used to support the grid at peak times.
• Assess and manage water impacts, set targets for alternative cooling technologies, and require tests of local water resilience before approval.

Why this matters for government

Electricity demand is expected to rise sharply by mid-century, and AI data centres are a significant part of that growth. National Grid ESO's Future Energy Scenarios point to higher system demand by 2050, underscoring the need for coordinated planning and efficiency measures. Source

Small modular reactors may help, but they are unlikely to be online until the mid-2030s. Meanwhile, several major tech companies have softened clean energy targets, making clear policy signals and enforceable planning tests more important.

Water is a parallel constraint. Defra has estimated a potential daily shortfall close to 5 billion litres by 2050, and data centre cooling can add to local stress. The national framework for water resources highlights the need for regional coordination and demand reduction. Source

Government response to date

Ministers noted that the Department for Science, Innovation and Technology (DSIT) is developing a National Policy Statement to guide planning for data centres and AI infrastructure. It is intended to be a material consideration for local planning decisions, with publication targeted for 2026.

The government has also created an AI Energy Council, with a sustainability working group to identify options for low-carbon energy supply to AI and tools to reduce AI-related emissions. Supporters of the amendment welcomed these steps but warned that AI's energy and water demands are growing faster than current policy development.

Practical steps for central and local government now

• Set interim expectations in plan-making and development management for AI/data centre proposals until the national strategy is in place.
• Require whole-life energy and water impact assessments, including projections, efficiency measures, and resilience under peak conditions.
• Prioritise designs that enable heat recovery, grid flexibility (e.g., demand response), and on-site or contracted low-carbon power.
• Enforce water efficiency and alternative cooling targets; where appropriate, apply water neutrality or offset requirements in stressed catchments.
• Secure metering, data-sharing, and annual reporting conditions on energy and water use to support monitoring and enforcement.
• Coordinate early with network operators and water companies on capacity, connection timelines, and upgrade funding.
• Use planning obligations and conditions to lock in efficiency upgrades, heat networks interconnection, and standby generation standards.

Risks of delay

• Grid constraints that stall economic projects or trigger costly reinforcements.
• Local water stress, especially during heatwaves and droughts.
• Stranded or underperforming assets if efficiency and cooling standards tighten later.
• Erosion of public trust if AI growth drives visible environmental impacts without clear benefits.

Timeline and next steps

The amendment seeks a national AI energy efficiency strategy within 18 months. The DSIT policy statement is slated for 2026, creating a near-term gap. Departments and planning authorities can close that gap by applying material consideration tests now, standardising evidence requirements, and setting clear conditions for efficiency, heat reuse, and water stewardship.

The goal is simple: enable AI infrastructure that strengthens grids and water systems rather than straining them.

Further reading and skills
For teams building AI literacy to support policy and planning, see relevant training options for public-sector roles: Courses by job.