Each time somebody makes use of ChatGPT to jot down a 100-word electronic mail, roughly 519 millilitres of water is consumed, virtually equal to a normal water bottle.
That estimate comes from a peer-reviewed 2025 paper revealed in Communications of the ACM by Pengfei Li, Shaolei Ren and their colleagues on the College of California, Riverside.
The research accounts for each the direct water used to chill knowledge centre servers and the oblique water required to generate the electrical energy that powers them.
When that is scaled to thousands and thousands of customers making a number of queries daily, the numbers turn into staggering.
By 2027, international AI infrastructure is projected to eat between 4.2 billion and 6.6 billion cubic metres of water yearly. That’s almost equal to half of the UK’s complete annual water withdrawal.
Extra worryingly, a lot of this water is being drawn from areas which are already going through water stress.
Why AI Information Centres Eat So A lot Water
Information centres generate monumental warmth. The highly effective chips that run trendy AI techniques, particularly high-end graphics processing items, can every eat and launch between 300 and 700 watts of warmth whereas working beneath heavy load.
To regulate this warmth, many knowledge centres depend on evaporative cooling. On this system, water is pumped via the power to soak up warmth from servers. A portion of that water then evaporates into the ambiance.
Round 80 p.c of the water drawn into an evaporative cooling system is completely misplaced via evaporation.
The remaining water might return to the system, however typically at increased temperatures and with chemical residues.
The brand new technology of AI-focused hyperscale knowledge centres is bigger, denser and much more heat-intensive than the overall cloud infrastructure constructed within the 2010s.
A single massive knowledge centre campus can now eat extra water in a day than a city of 10,000 folks makes use of for ingesting, sanitation, cooking and agriculture mixed.
In line with the 2024 US Information Centre Vitality Utilization Report by Lawrence Berkeley Nationwide Laboratory, ready for the US Division of Vitality, knowledge centres consumed round 17.4 billion gallons of water instantly for cooling in 2023.
One other 211 billion gallons had been consumed not directly via electrical energy technology.
The report additionally famous that knowledge centre load progress has tripled over the previous decade and is projected to double or triple once more by 2028.
Google, Microsoft And Meta’s Water Use
Main know-how firms have began disclosing their water consumption in annual sustainability studies. The pattern is evident: water use is rising.
Google’s 2024 Environmental Report mentioned the corporate consumed round 8.1 billion gallons of water through the 12 months, with about 95 p.c used at knowledge centres. This marked an 8 p.c improve from 2023.
The earlier years had additionally seen sharp will increase, with Google’s water consumption almost doubling in three years.
Microsoft has reported smaller numbers, however the sample is comparable. The corporate consumed round 1.7 billion gallons of water in 2022, a 34 p.c improve from the earlier 12 months.
Impartial reporting on Microsoft’s knowledge centre cluster in West Des Moines, Iowa, the place GPT-4 coaching runs had been performed in 2022, discovered {that a} single coaching run consumed 11.5 million gallons of water in July 2022 and 13.4 million gallons in August.
The identical cluster has since expanded to 5 amenities, drawing 68.5 million gallons yearly from the native municipal water system.
Meta consumed round 813 million gallons of water globally in 2023. Amazon, which operates the world’s largest cloud infrastructure, doesn’t publish mixture water consumption figures.
AI Is Increasing In Water-Careworn Areas
The Li and Ren paper tasks that by 2027, international AI demand might result in water withdrawal equal to greater than 4 instances Denmark’s annual utilization, or almost half of the UK’s complete yearly withdrawal.
The issue just isn’t solely the quantity of water used, but additionally the place that water is coming from.
Microsoft acknowledged in its 2023 sustainability report that round 42 p.c of its water consumption got here from areas categorized as water-stressed beneath the World Sources Institute’s ranking system.
Google reported that 15 p.c of its freshwater withdrawals in 2023 got here from areas going through excessive water shortage.
The implications are already seen. In Chile, Google paused a deliberate $200 million knowledge centre close to Santiago after an environmental court docket dominated that the corporate had not adequately assessed the affect on the Central Santiago Aquifer.
The nation had been going through drought for 15 years and had begun rationing residential water in 2022.
In Querétaro, Mexico, the place 32 new knowledge centres are deliberate, the state skilled its worst drought in a century in 2024.
Microsoft has secured rights to round 25 million litres of water yearly from an area aquifer that’s already operating a 60-million-litre annual deficit.
In Arizona, a $14 billion knowledge centre challenge was withdrawn in 2024 after native residents opposed the rezoning.
What Corporations Are Not Disclosing
The obtainable figures are based mostly solely on what firms have chosen to reveal. The precise water footprint of the AI trade is probably going a lot bigger.
There are three main disclosure gaps.
The primary is the distinction between water withdrawal and water consumption. Withdrawal refers back to the complete quantity of water taken from a supply, whereas consumption refers back to the water completely misplaced, largely via evaporation.
Corporations typically report solely one among these figures, which may make their footprint seem a lot smaller.
The second hole is the distinction between direct cooling water and oblique water used for electrical energy technology.
The Li and Ren analysis estimates that oblique water use could possibly be round 12 instances increased than direct cooling water use. Only a few company studies embrace this determine.
The third hole is the absence of facility-level knowledge. An organization-wide annual complete doesn’t inform native communities whether or not their very own aquifer or water system is beneath strain.
The UC Riverside paper is essential as a result of it makes use of publicly obtainable data to estimate these hidden gaps.
As unbiased researchers produce credible estimates, know-how firms might face rising strain to reveal extra detailed water knowledge.
The Huge Query For AI
The worldwide infrastructure for AI is being constructed at a rare tempo. Behind the joy over synthetic intelligence are large industrial amenities that rely closely on cooling techniques and electrical energy.
Every particular person AI question could seem small. However when multiplied by billions of interactions, the environmental price turns into important.
Supporters argue that AI might assist remedy local weather and water issues via higher local weather modelling, improved irrigation techniques and extra correct drought prediction.
However the true query is whether or not AI can ship these advantages sooner than its personal water consumption grows.
For now, that query stays unanswered.