12th January
Data centres – which includes those that provide AI – use large amounts of water, principally in providing cooling., but also (depending on the source of their electricity) from the water needed at fossil fuelled power plants where water is again needed for cooling. It is estimated that the water footprint of AI in 2025 was somewhere between 313 and 765 billion litres which apparently is roughly the same as the amount of bottled water consumed globally. (1)
This consumption of water is going to increase as our use of and development of AI continues to expand. By 2027 experts predict global water use by could reach 4.2–6.6 trillion litres annually – equivalent to about half the UK’s annual water consumption. (2)
In a world where freshwater is already in short supply, this is a serious issue as highlighted in this comment from a UK Government blog:
“AI has potential to address challenges in healthcare, inequality and climate monitoring. However, these benefits must be weighed against the full environmental impact of building and running AI systems. This requires development of a comprehensive and transparent series of metrics that include, carbon emissions, energy consumption, water usage, biodiversity and social impact. Only by looking at AI through these combined lenses can we ensure it delivers net benefit rather than widening inequality or disproportionately impacting communities and regions.” (3)
A UK government report ‘Water use in AI and Data Centre’ highlights the issue of the 2050 projected daily 5 billion litre water deficit versus the expansion of date centres in the UK. It notes with concern that in coming to its projection of a 5 billion litre water shortage, the Environment Agency has not included any figures for the anticipated increased water demands that will be made as data centres expand. This is aggravated by a lack of information as data centres are not required to be transparent about the amount of water and electricity that are and will be using. Mandating transparency would encourage data centre to find ways of minimising the resources they need and would enable the relevant authorities to make better decisions as to where new data centres could or should not be located. (They should not for example be located in areas already prone to water shortages). (4)
Equally if information about the likely consumption of scarce resources were more readily available, it would enable a better discussion as to whether we should be encouraged the unrestricted growth of AI and data centre, and have the ability to weigh up the benefits of AI against its detriments.
- https://www.theguardian.com/technology/2025/dec/18/2025-ai-boom-huge-co2-emissions-use-water-research-finds
- https://www.aidrinkswater.com/report.html
- https://sustainableict.blog.gov.uk/2025/09/17/ais-thirst-for-water/
- https://assets.publishing.service.gov.uk/media/688cb407dc6688ed50878367/Water_use_in_data_centre_and_AI_report.pdf
The Green Tau 