World Environment Day: AI could chug water enough to sustain all 8.1 billion people on Earth for over a year, UN report warns
&imwidth=600&imheight=450&format=webp&quality=medium "World Environment Day: AI could chug water enough to sustain all 8.1 billion people on Earth for over a year, UN report warns")
AI could leave humans high and dry with the amount of water and electricity it uses which will skyrocket in the coming years. Photograph: (AFP)
World Environment Day: A report by the United Nations has painted a scary picture of the future in the age of Artificial Intelligence. The water it uses can quench the thirst of all 8.1 billion people on Earth for over a year. Electricity, carbon emissions and electronic waste will skyrocket.
Humans seem to have dug their own graves with Artificial Intelligence, as a new study has found that the technology would take a massive toll on the climate in the next few years. According to a new report, by 2030, AI will need nearly 3 per cent of the world's electricity to function. The carbon emissions will be the same as that of United Kingdom in 2025, and electronic waste will be equivalent to trashing 250 Eiffel Towers every year. The most stark observation is about water, as the amount it would need will be enough to quench the thirst of every person on Earth for more than a year and a half. That's a pretty grim picture of the future painted by the United Nations University Institute for Water, Environment and Health.
Researchers say carbon emissions are only a part of the toll AI takes on the climate, and reducing the harm AI would cause to the environment would depend on several more factors than just that. "Low-carbon is not automatically low-water or low-land,” the report states. Data centres consumed an estimated 448 terawatt-hours of electricity in 2025, with AI processes accounting for around 20 per cent of it. By 2034, AI-related electricity use could reach 40 per cent, or 374 terawatt-hours.
AI used 9.3 trillion litres of water in 2025
These data centres run 24*7 and need massive amounts of water to cool the systems. The data centre totals could double to 945 terawatt-hours, which can power 1.3 billion people in sub-Saharan Africa for over five years. The consumption of water, a major problem linked with data centres, will also considerably rise. In 2025, they used an estimated 9.3 trillion litres of water. This water can meet the needs of all 8.1 billion people on Earth for more than a year and a half.
Also Read: Company accidentally blows away Rs 4,800 crore on Claude AI in a month because of a tiny oversight
Trending Stories
)
&im=FitAndFill=(700,400))
&im=FitAndFill=(700,400))
)
)
)
)
)
)
)
)
&im=FitAndFill=(700,400))
)
&im=FitAndFill=(700,400))
)
)
)
&im=FitAndFill=(700,400))
&im=FitAndFill=(700,400))
)
&im=FitAndFill=(700,400))
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
How much electricity ChatGPT uses?
The report further states that the cost of daily use of ChatGPT has exceeded the environmental cost of training. Training a model like ChatGPT-5 requires around 100 gigawatt-hours of electricity, one billion litres of water and a piece of land as large as 215 football fields. However, with an estimated 2.5 billion prompts per day, the consumption has gone way beyond the line. Researchers found that a concise response mode can reduce ChatGPT's output by 30 per cent, saving as much electricity as would be needed by nearly 760,000 people in sub-Saharan Africa in a year. Avoiding the use of words like please or thank you further helps.
Switching to green energy creates another problem
However, the biggest problem is with AI-generated videos. The report states that a single high-resolution AI video clip requires more than 415 watt-hours of electricity, way more than what hundreds of AI images would need. The report further found that using renewable energy to power the data centres does not solve the problem. Switching from coal to bioenergy drops carbon emissions by 72 per cent, but increases the water footprint 30-fold and the land footprint 100-fold due to the agriculture required.
