UN warns AI data centers could match 1.3 billion people's water use by 2030

A new UN report warns that AI data centers could consume as much water as 1.3 billion people by 2030, with electricity use reaching 945 TWh. The report from UNU-INWEH highlights that inference and prompt processing demand far more resources than model training, with ChatGPT alone using 383 GWh annually. This raises serious questions about the environmental footprint of digital infrastructure beyond just carbon emissions. The report also notes that data centers already strain local resources, citing examples in Ireland, Mexico, and Uruguay. It calls for a responsible AI ecosystem built on transparency, efficiency, and lifecycle responsibility. For those tracking grid decarbonization and energy policy, this is a concrete signal that AI growth is not just a carbon story but a water and land use story too.