UN report projects sharp rise in AI data centres’ water and electricity use by 2030
The Facts
- The United Nations University Institute for Water, Environment and Health published a report examining the environmental costs of AI, including carbon, water and land impacts tied to data-centre energy use.
- The report projects that by 2030, AI-related data-centre water use could equal the basic domestic needs of about 1.3 billion people.
- The report says global data centres supporting AI could consume about 945 terawatt-hours of electricity in 2030, roughly triple current levels cited in the coverage.
- Several reports citing the UN study say data centres used 448 terawatt-hours of electricity in 2025.
- The report says AI’s environmental footprint is not limited to electricity demand; it also includes pressure on water resources, land use and electronic waste.
- The UN report calls on AI companies to publish information about their environmental footprint and urges governments to require standardized environmental reporting from AI developers.
- The report presents the growth of AI infrastructure as a policy issue because its resource demands could affect electricity grids, water supplies and land availability in places where data centres are built.
- A key unresolved issue identified in the coverage is limited transparency and a lack of comparable public data on AI companies’ environmental impacts, which the report says makes the full scale of the footprint hard to measure.
How left and right are reading this
- Both agree
- AI’s buildout carries material environmental costs beyond electricity alone, with water, land and waste pressures serious enough to affect the places where data centres are built and to warrant better public accounting of those impacts.
- They split on
- Less a disagreement than a question of emphasis: the sheer scale of AI’s projected resource consumption, versus the policy and transparency tools needed to measure and manage it.
Context
What is the main finding of the UN report?
The report projects that by 2030, AI-related data centres could use 945 terawatt-hours of electricity a year and 9.3 trillion litres of water, with water demand comparable to the basic domestic needs of 1.3 billion people NDTV,RTVE.es,europa press.
Why does the report say this matters beyond carbon emissions?
The report says AI infrastructure affects multiple resources at once: electricity generation, water for cooling and operations, land for facilities and supply chains, and electronic waste from hardware turnover infobae,El Tiempo,La Jornada.
What actions does the UN report recommend?
It urges AI companies to disclose their environmental footprint and calls on governments to require standardized reporting, arguing that better measurement and transparency are needed to assess and manage the sector’s impacts infobae,El Universal,France 24.
View all 90 sources
Wire services (2)
Independent coverage (50)
About these frames
See this differently than someone you know would? Two ways to keep it going.
The dial works on any URL — paste an article you read elsewhere this week.