Datacentre power and water use to double by 2030

AI’s electricity demand will triple the total combined consumption of Pakistan, Bangladesh and Nigeria

AI is driving a surge in global energy and water demand, with new research from the United Nations suggesting that both will double in the next four years.

A report by the UN University Institute for Water, Environment and Health (UNU-INWEH) has found that that datacentres and other infrastructure carry substantial “carbon, water and land footprints” that are often overlooked in public debate.

The environmental costs of AI and datacentres cannot be understood through carbon emissions alone, the researchers argue.

"If we keep judging AI sustainability by carbon alone, we might think that renewables make AI infrastructure clean but that is solving one problem while creating other problems, often in places that didn't ask for it," said lead author Dr Miriam Aczel.

Datacentres consumed around 448TWh of electricity in 2025, according to the report, but the researchers predict that by 2030, datacentres could consume 945TWh annually - nearly triple the combined consumption of Pakistan, Bangladesh and Nigeria. Meanwhile their water demand, projected as 9.3 trillion litres annually, could match the basic needs of 1.3 billion people.

While renewable power can help reduce carbon emissions, improved efficiency is also likely to increase usage in an unhelpful example of Jevons Paradox, the report notes.

"A lot of people think that the environmental footprint of AI reduces as technology improves and processes become more efficient," said co-author Professor Kaveh Madani But that is only a partial picture of the overall problem.

"More efficient and affordable AI and energy mean more consumption of AI, making the overall footprint far bigger than what we save through efficiency gains."

Moreover, datacentres are often sited close to communities "where water stress is worst," the researchers say. "The communities living near these sites are not necessarily the ones using the AI being run there."

The report urges consideration of energy, water and land use in all new AI infrastructure planning, along with transparency and the adoption of standardised reporting metrics. Investors should treat electricity, carbon, water, and land footprints as material risks in AI infrastructure portfolios, and communities should be involved early in the decision-making process, the researchers recommend.