Artificial intelligence lives on power and water, fed to it in vast quantities by data centres around the world. And those centres are increasingly located in the global south.
One estimate from the University of California, Riverside says AI’s total water demand by 2027 could be more than half the total annual water withdrawal of the UK. But all we really have are estimates. Big tech firms have been secretive about the amount of public water used by individual data centres, and up to half of all data centres don’t even measure how much water they use, according to one survey.
Colon, a municipality north of Mexico City/in central Mexico, is home to Microsoft’s first hyperscale data centre campus in the country. The town of 67,000 is suffering extreme drought. Its two dams have nearly dried up, farmers are struggling with dead crops, and families are relying on trucked and bottled water to fulfil their daily needs. Mexico, leveraging its proximity to the US, is hoping to convince Big Tech to ‘nearshore’ their facilities here. The State of Queretaro is offering favourable land loans, cheap electricity and a pool of local talent.
Similar stories are playing out around the world. In Uruguay, Google admitted that a planned data centre in Montevideo would require 7.6 million liters of drinking water per day, while the country was suffering a historic three-year drought. In the US, a bill has been introduced in the Senate to compel Big Tech to reveal the environmental impacts of AI after reports of conflict over water between farmers and Big Tech in the desert of Arizona.
In this episode of Context, we explore why Big Tech is choosing water-stressed regions for their water-hungry operations. And why are governments giving data centres access to water when so many of their citizens are going thirsty?
Subscribe to our YouTube channel | Watch more videos like this
