Could soon use 3% of world's electricity and more water than we need to drink: UN report

A United Nations report warns that artificial intelligence could dramatically increase its environmental footprint, potentially using 3% of global electricity by 2030 and drawing more cooling water than the world can drink. The study highlights the Jevons paradox, whereby efficiency gains spur overall consumption, not reductions, as cheaper AI expands usage. To avoid this trap, the report calls for governance spanning the entire value chain—resource extraction to end-of-life recycling—guided by transparency, design efficiency, equity, lifecycle responsibility, and global cooperation. It notes that data centers already consume electricity equivalent to Saudi Arabia’s total consumption and would require 9.3 trillion litres of water; offsetting this demand would require about 6.7 billion trees, underscoring the sector’s vast footprint and potential inequities in access to AI infrastructure.