How AI tools like ChatGPT are causing water usage to skyrocket


The global popularity of AI tools like ChatGPT is causing the water usage of Big Tech firms to skyrocket.

Data centres siphon off a large volume of liquids from the local water supply to cool down their equipment, which generates excess heat when performing the computations needed to make apps like ChatGPT run.

Microsoft’s annual environmental report claims its global water usage shot up over a third between 2021 to 2022, coming in at a total of almost 1.7 billion gallons.

This is roughly enough to fill 2,500 Olympic-sized swimming pools, as per estimates by AP News.

This compares to the far more modest 14 per cent increase in water usage Microsoft reported between 2020 and 2021.

Google is also dealing with spiralling water usage.

The search giant used about 20 per cent more water than it did last year as per its most recent environmental report. This equates to roughly 5.6 billion gallons, or the equivalent of what it takes to irrigate 37 golf courses annually, as per Google’s own estimates.

Shaolei Ren, a researcher at the University of California, Riverside, who is authoring a third-party research paper on the project, claims that the majority of these eye-popping jumps in water usage are due to the fantastic rise of AI products over the past year, including Open AI’s ChatGPT.

The scientist told AP News he estimates that GPT uses around 500 millilitres of water whenever you ask it between five to 50 questions.

These eye-popping levels of water usage are only set to grow in the coming years.

Business consultancy McKinsey expects the demand for data centres to roughly double over the next decade, at least in the US, which represents around half of the overall market.

According to estimates from 2021, the average Google data centre uses roughly 450,000 gallons of water every day.

Google estimates this is roughly equal to the volume of water needed to irrigate 17 acres of turf lawn grass or to grow cotton for and produce 160 pairs of jeans.

How much power does AI use?

Unfortunately, water isn’t the only problem that the data centre industry is dealing with alongside the rise of AI.

Data centres are said to account for up to three per cent of global electricity consumption, which is projected to rise to four per cent by 2030.

What are tech firms doing to save water?

Acknowledging water usage as an issue, Microsoft told AP News that it is “working on ways to make large systems more efficient, in both training and application”.

In 2021, Google blogged about some of the strategies it was taking to reduce water consumption.

These included using alternatives to freshwater whenever possible, which is considered the most valuable, and instead using wastewater, industrial water, or even seawater.

For example, Douglas County, in the US state of Georgia, cools one of its Google data centres by recycling local municipal wastewater that would otherwise flow into the Chattahoochee River.

There are also other methods of cooling data centres that are potentially more environmentally friendly.

These include simply building data centres in areas that have significantly cooler climates, or building more energy-efficient infrastructure that requires less water to cool.

We’ve also seen signifiantly more creative approaches to the issue appear over the past few years; in 2018, Microsoft built an underwater data centre just off the coast of Orkney, Scotland, which pumped in local cold seawater for a more efficient approach to cooling.