The government has been called upon to create a national artificial intelligence (AI) energy efficiency strategy in the next 18 months as the House of Lords questioned the potential demands the influx of data centres could have on the UK’s grid.
The Lords Environment and Climate Change Committee spokesperson Earl John Russell tabled the motion as an amendment to the Planning and Infrastructure Bill. It would see the energy use and efficiency of AI-related infrastructure as a statutory planning consideration. This would be supported by a national strategy, guidance and reporting duties.
He said: “More widely, AI can be used to heat and cool our buildings, improve our transport sector and to improve any number of industrial processes requiring large amounts of energy.
“This amendment would require the Secretary of State to publish a national AI energy efficiency strategy within 18 months.”
Russell wants the amendment to establish a framework for the “projections of energy use as material planning considerations and mandate developers to account for both supply and efficiency measures in their applications”. The measure also aims to ensure that surplus energy resources from data centres can be fed back into the national grid at times of energy need.
“Our overall electricity demand is set to double by 2050. SMRs [small modular reactors] will help, but we should note that they will not come online until the mid-2030s,” Russell said.
“Meanwhile, many big tech companies are rowing back from their clean energy targets; Google is one example.
“[The Planning and Infrastructure Bill] is notably silent on AI, meaning that planning frameworks lack explicit provisions to assess or moderate the substantial energy and water demands of AI.
“I believe that, with the right legislation, we can make provision to require that AI is used for public good by ensuring that its power is also applied to finding ways we can drive national energy savings and efficiencies.”
The amendment also looks at the potential water usage of data centres and how that could affect the UK’s water system.
The shortfall between water supplies and expected demand has been estimated by the Department for Environment, Food and Rural Affairs (Defra) to be close to 5bn litres per day by 2050.
A significant contributor to the increase in demand is expected to come from an influx of data centres being built in the UK to account for the surge in AI use. Data centres are necessary to house the Language Learning Models (LLMs) that constitute AI applications as they need an enormous amount of computing power and storage to function. Data centres use cooling techniques that utilise large amounts of water to avoid overheating.
Russell said: “Facing a warmer world, it is essential that national policy demands a clear water efficiency strategy, enforces targets for alternative cooling technologies and ensures that planning authorities rigorously assess water availability and resilience before consenting to new developments.”
NCE contacted the Department for Science, Innovation and Technology (DSIT) for a comment on the call for a strategy addressing these concerns but did not receive a reply.
Nevertheless, during the debate, Ministry of Housing, Communities and Local Government (MHCLG) under-secretary of state Baroness Taylor spoke about some of the work DSIT is doing to allay the concerns about AI data centres’ water and energy consumption.
“Officials from DSIT are in the process of developing a national policy statement to guide this process, which will also be treated, importantly, as a materially significant consideration in the local authority-led planning process,” she said.
“This statement will include an assessment of the sustainability of the sector, and we are aiming for publication in 2026.
“The government has also established the AI Energy Council to provide expert insight into the energy needs of AI and the role of AI in an efficient and sustainable energy system.”
In order for this council to do this it has established a sustainability working group to explore options to accelerate the development of low-carbon energy solutions to power AI and tools to reduce carbon emissions from AI
While Russell heralded the creation of the council, he also stated that the technology’s energy and water demands are accelerating at a pace quicker than the government can act.
“The simple truth is that AI’s energy and water demands are outpacing our development of policies to regulate AI’s energy and water usage,” he said.
Like what you’ve read? To receive New Civil Engineer’s daily and weekly newsletters click here.
