Data centers for artificial intelligence could consume by 2030 water equivalent to the annual needs of 1.3 billion people, according to a new report.

A recent report from the United Nations University Institute for Water, Environment and Health highlights that by 2030, data centers supporting artificial intelligence will consume an amount of water equivalent to the annual needs of 1.3 billion people. Researchers found that the environmental impact of AI infrastructure development is often misassessed, focusing on carbon emissions while ignoring water consumption and land use.

Inference costs, which account for 80-90% of total energy consumption, are significant, and the water footprint of AI will require 9.3 trillion liters of water. Paradoxically, the energy efficiency of AI technology could lead to more intensive use, thus increasing the overall environmental footprint. The report warns that despite technological advancements, the environmental impact of artificial intelligence will continue to grow.