AI models are booming, and they are becoming more and more demanding in terms of resources with each passing day. And the concern related to AI data centres consuming natural resources is something that has been in discussion for a very long time now. According to a United Nations report suggests that by 2030. AI energy use could double to consume 3 per cent of the world’s electricity, produce emissions equal to the UK, and consume more water for cooling than the annual drinking water required for the global population.
Anticipations are that AI will also follow the economic principle known as the ‘Jevons Paradox,’ which predicts that when technological improvement increases the efficiency of a resource, it leads to a rise, instead of a fall, in the total consumption of that particular resource. For those unaware, the paradox is named after economist William Stanley Jevons, who observed this effect with the use of coal in 19th-century England.
Perplexity CEO Aravind Srinivas Says This Metric Will Define The AI Race Winner
Keeping in mind that AI models are becoming cheaper and more attractive, the report expects this to encourage new uses and higher volumes of use, eroding and possibly erasing any savings from efficiency advances. If we look at the previous stats, the data centres already consume as much electricity as Saudi Arabia, which ranks as the world’s 11th largest electricity consumer. The only thing that can put a halt on this is AI adopted rates dropping instead of increasing, or efficiency improvements in chips and models.
What’s more concerning is that if electricity use doubled as projected by 2030, then the associated carbon footprint would require 6.7 billion trees grown over ten years to offset this demand. That’s not it, as data centres would also need 9.3 trillion litres of water and land, which is almost ten times the size of Mexico City.