UN's WARNING ABOUT AI - Sep 2026
AI may feel invisible, but the infrastructure behind it is very physical—and increasingly thirsty.
Every AI query is processed inside data centers packed with powerful chips that generate enormous amounts of heat. Many facilities use water-based cooling systems to keep those machines operating, while additional water can be consumed indirectly by the power plants supplying their electricity.
The scale could become enormous. UN-linked estimates cited in recent reporting suggest that by 2030, annual AI-related water consumption could equal the basic domestic water needs of around 1.3 billion people. As AI adoption and data-center construction accelerate, that demand could become especially important in regions already struggling with water scarcity.
But this future isn’t inevitable. Newer data centers are increasingly experimenting with closed-loop systems, air cooling, reclaimed or non-potable water and technologies designed to use little or even zero water for cooling. The environmental footprint of AI will depend heavily on where facilities are built, how their electricity is generated and which cooling technologies they use.
AI could help humanity solve enormous problems, but its growth comes with a physical cost we rarely see on our screens. The challenge isn’t simply building more powerful AI—it’s building it without placing unnecessary pressure on the resources humans already depend on.
Source: United Nations estimates / Financial Times reporting.
