Data Drought: AI Is Using Vast Amounts Of Water

Artificial intelligence (AI) is having an unintended impact on a vital resource – water. As AI systems expand their reach, they demand colossal computing power housed in sprawling data centres.

These data centres, in turn, require vast quantities of water for cooling, adding to concerns about the environmental sustainability of AI growth.


Microsoft’s Soaring Water Consumption


In a shocking revelation, Microsoft, one of the tech industry’s biggest companies, consumed more than 2,500 Olympic-sized swimming pools worth of water in its data centres in the previous year alone.

This figure, as disclosed in Microsoft’s latest environmental report, has sent ripples of concern throughout the tech world. The numbers reflect a 34% increase in water consumption between 2021 and 2022, with the company guzzling nearly 1.7 billion gallons.


AI: The Culprit Behind the Surge


While not all of Microsoft’s water consumption can be attributed to AI, experts suggest that the majority of this surge is directly linked to the growth of artificial intelligence. Shaolei Ren, a researcher at the University of California, Riverside, has been dedicated to understanding AI’s environmental impact. According to Ren, the staggering increase in water usage is primarily due to the appetite for water-cooling by AI-powered servers.


Google’s Rising Thirst


Microsoft is not alone in this water crisis. Google, another tech titan, reported a 20% increase in its water usage during the same period. It’s worth noting that Google’s data centres, scattered across different parts of the United States, displayed varying rates of water consumption. This indicates that water demand is not uniform across data centres and can have localised environmental effects.



Measuring the Water Impact of AI


A forthcoming paper by Professor Shaolei Ren and his team sheds light on the precise water consumption associated with AI models. According to their research, each set of 5 to 50 prompts, or questions, posed to AI models like ChatGPT consumes a staggering 500 millilitres of water. This statistic shows the hidden environmental cost behind seemingly innocuous interactions with AI systems.

Several technology companies are grappling with their water usage and environmental responsibilities. Google, for instance, declared its commitment to responsible water usage, vowing to scrutinise and optimise water consumption across its operations. This includes conducting thorough assessments of where water is used and assessing the potential impact on surrounding ecosystems.


Navigating the AI-Water Crisis


As the AI industry continues its explosive growth, the water crisis in data centres emerges as an alarming challenge that demands immediate attention. The revelations from Microsoft and Google show the need to address the environmental consequences of AI’s thirst for water. Tech giants, researchers, and policymakers must collaborate to develop sustainable solutions that ensure AI’s development does not come at the expense of our planet’s most precious resource.

Balancing technological progress with environmental protection will be an important part of the road ahead.