Labour’s first artificial intelligence growth zone will be sited close to the UK’s first new reservoir in 30 years, sparking fears that the AI push will add to the “severe pressure” on water supplies in the area.

Keir Starmer announced on Monday that he would hugely increase artificial intelligence capacity and reduce planning restrictions on companies that wanted to build datacentres by setting up “growth zones” with fewer constraints.

The first of these will be in Culham, Oxfordshire, only seven miles from a reservoir planned by Thames Water in Abingdon, which was supposed to provide water to people in the severely water-stressed south-east of England. This is the area of the country most at risk of running out of water, according to the Environment Agency. Oxfordshire has faced particular issues, with areas reliant on bottled water during heatwaves.

AI datacentres use a large amount of water, as their servers generate heat. To prevent computer systems overheating and shutting down, the centres use cooling towers and outside air systems, both of which need clean, fresh water. AI consumes between 1.8 and 12 litres of water for each kilowatt hour of energy usage across Microsoft’s global datacentres. One study estimates that global AI could account for up to 6.6bn cubic metres of water use by 2027 – the equivalent of nearly two-thirds of England’s annual consumption.

Even without a big increase in AI datacentres, by 2050, England faces a shortfall of nearly 5bn litres of water a day between the sustainable supplies available and the expected demand. This is more than a third of the 14bn litres of water currently put into public supply. The south-east faces a potential deficit of more than 2.5bn litres a day in the same period.

AI could wipe out gains made by businesses in reducing their water consumption; the government is seeking a 9% reduction in non-household (business) consumption by 2037-38 from 2019-20 levels, and currently businesses are on course to achieve a reduction of 6.1%.

Adrian Ramsay MP, Green Party co-leader, said: “While communities will face heatwaves, droughts and water shortages over the coming decades, this strategy locks us into pumping huge amounts of water into AI datacentres. One estimate said AI-related infrastructure may soon consume six times more water than Denmark, a country of 6 million people. What will this mean for residents in water-stressed communities like Culham in Oxfordshire?”

Prof Hannah Cloke from the University of Reading added: “The south of England already has severe pressure on water resources, which is getting more acute as we build more homes and look to grow hi-tech industries, all of which need more power and water. We know that as the climate changes in the UK there will be a more variable supply of water from the sky with hotter drier summers which will exacerbate the demands on cooling systems that datacentres need.”

AI datacentres also use a lot of energy. The International Energy Agency estimates that datacentres’ total electricity consumption could double from 2022 levels to 1,000 terawatt hours in 2026, approximately Japan’s level of electricity demand. AI will result in datacentres using 4.5% of global energy generation by 2030, according to calculations by the research firm SemiAnalysis. In January, Amazon, the world’s largest corporate purchaser of renewable energy, announced it had bought more than half the output of an offshore windfarm in Scotland.

A Thames Water spokesperson said it was not a statutory consultee for datacentres, meaning the company did not need to be asked about the availability of water. Thames intends to ask the government whether sustainable sources, such as surface water, can be used to cool the datacentres, rather than water directly from the reservoir.

The government is setting up a dedicated AI Energy Council chaired by the science and energy secretaries. The government said it intended to work with “energy companies to understand the energy demands and challenges” that would fuel AI’s development. It hopes small modular nuclear reactors could power datacentres.

Prof Gopal Ramchurn, professor of artificial intelligence at the University of Southampton, said: “AI expansion has been a concern for the National Grid but the speed at which AI compute demand is growing has taken everyone by surprise and, unless we balance the above tradeoffs right, with appropriate policies, all the cheap and green energy we have will be used by big tech companies, pricing out families suffering energy poverty already.”

Some scientists have said there are environmental opportunities from AI. Dr Shaun Fitzgerald, teaching fellow at the University of Cambridge, said: “The increased energy use of AI should not be considered in isolation, but rather coupled with the potential energy reduction that intelligent control of our energy systems could enable. There are incredible opportunities to make more of what we already have in our energy system, and integrated AI control has the potential to unlock them.”

A government spokesperson said: “We recognise that datacentres face sustainability challenges such as energy demands and water use. Many newer datacentres are already addressing these issues, using advanced cooling systems that significantly reduce water consumption.

“Through the AI Energy Council, we’ll also build on this progress by exploring bold, clean energy solutions – from next-generation renewables to small modular reactors – to ensure our AI ambitions align with the UK’s net zero goals. We’re also unlocking £104bn in water infrastructure over the next five years, which includes supporting water supply resilience in and around datacentres.”