What AI actually costs the planet
A grounded look at AI's energy and water footprint — neither dismissive nor doom-laden.
Updated 31 Jul 2026 fast-moving — check a current source
Growth figures from IEA reporting published early 2026. Per-prompt figures vary enormously by model, provider and grid — treat any single number with suspicion, including the ones here.
AI has a real environmental cost, and it’s worth understanding plainly — without the hand-waving that says it’s nothing, or the catastrophising that says it’s everything. The cost comes in two phases, and they’re very different in size.
Training vs use
- Training a large model is energy-intensive and happens occasionally — a big, one-off cost amortised across everyone who later uses the model.
- Using it (each prompt and response, called inference) is far smaller individually, but it happens billions of times, so at scale it adds up.
There’s also water: many data centres use water for cooling, which matters most in places already under water stress.
Keeping it in proportion
A single chat is a small cost — roughly in the ballpark of other routine digital activity, not a flight. The honest concern is aggregate and growth: a lot of people using AI a lot, and infrastructure expanding to meet it. The footprint also depends heavily on how clean the local electricity grid is.
The right frame isn’t “is one prompt bad?” — it’s “is all of this, growing this fast, on this grid, sustainable?” That’s a question for providers and policy as much as users.
The growth is the story
The IEA’s reporting puts numbers on the aggregate concern. Global data-centre electricity demand grew around 17% in 2025, with AI-focused facilities growing considerably faster, and the agency expects data-centre consumption to roughly double by 2030. In the United States, data-centre expansion accounts for something like half of projected electricity demand growth.
Two things are true at once, and the argument usually goes wrong by picking one. Per-prompt efficiency is genuinely improving — providers now compete on getting the same work done for fewer tokens, which is a real reduction per unit of work. And total consumption is rising fast anyway, because usage is growing faster than efficiency. Efficiency gains that get spent on more usage are not reductions.
What an individual can do is use it deliberately — see using AI without the waste.
Sources
Everything above was checked against these on 31 Jul 2026. Providers change things without notice — if a detail matters to a decision, follow the link.