UN warns AI boom hides rising environmental cost of data centre expansion: 945 TWh by 2030
businessamlive.comA new United Nations report warns that global data centres could consume 945 terawatt-hours of electricity annually by 2030, nearly three times the combined current consumption of Nigeria, Pakistan, and Bangladesh. The report from UNU-INWEH highlights that AI infrastructure's environmental footprint goes beyond carbon emissions to include massive water and land use, with data centres projected to require 9.3 trillion litres of water and over 14,500 square kilometres of land by decade's end. The report notes that inference, not training, accounts for 80-90% of an AI model's lifetime energy use, with ChatGPT alone processing 2.5 billion daily prompts. It also warns that efficiency gains may backfire due to the Jevons paradox, where cheaper AI drives even more consumption. The UN calls for a six-principle framework including transparency, environmental justice, and lifecycle responsibility to manage these impacts.
