AI data centers may consume water equivalent to 1.3 billion people by 2030

A United Nations University report warns that the water and land footprint of artificial‑intelligence infrastructure could rival the basic water needs of 1.3 billion people in sub‑Saharan Africa by 2030. To power AI models, global data‑center electricity use is projected to reach 945 terawatt‑hours, enough to rank the sector as the world’s 11th largest electricity consumer, while the land occupied would exceed 14,500 km². The report notes that most of the energy is used during inference – the processing of user queries – which now accounts for 80‑90 % of AI’s total power demand. For example, ChatGPT handles roughly 2.5 billion queries a day, consuming about 383 GWh of electricity annually and a water volume comparable to the annual needs of half a million people in Africa.

The rapid expansion of AI‑focused data centres is also sparking community opposition in several countries. Projects in the United States, Chile and elsewhere have drawn criticism for their massive water use for cooling and for the broader environmental strain they place on local resources. High‑profile figures such as Erin Brockovich have voiced concerns, highlighting the growing conflict between technological growth and water scarcity.