What is AI’s environmental footprint?
AI uses energy, water, materials, infrastructure and human labour.
What is it?
AI uses electricity, water for cooling, materials for chips and infrastructure. AI data centres are growing fast, and consumption depends on the size of the model, the type of task and the energy source. On the other hand, AI can also help save resources.
An everyday example
Generating a video uses far more energy than a short text answer, while smart energy management can cut the consumption of a whole building.
Why does it matter?
Rising consumption affects climate targets, electricity prices and local water supplies. That is why public, comparable data are needed.
A common misconception
That the footprint can be judged from a provider’s name, or that every question costs the same. The footprint has to be measured.
In more detail
Why do data centres use electricity?
For computing, cooling, data services and related infrastructure.
Is every AI question equally demanding?
No. The cost depends on the model, the length of the answer, the hardware and the part of the system involved.
Can AI also reduce resource use?
It can help optimise some processes, but the overall effect has to be measured — including the extra use it encourages.
What are the important questions?
Where the electricity comes from, water use, where data centres are located, chip production, how long equipment lasts and what the real benefit is.
Remember: The environmental footprint has to be measured, not guessed from the size of the brand.
Reliable sources
Last reviewed 1 November 2026. An explainer, not professional, legal or medical advice. For rules and figures that change, check the official version.