AI is growing fast, and keeping up means building more data centers, manufacturing advanced chips and powering the tech behind it. All of that comes with a carbon cost. AI advocates claim that in the long run, AI will save energy and cut carbon emissions across global industries.
However, a new study published in the journal Communications Earth & Environment points out a flaw in this argument. By the time those long-term savings come, we would already have accumulated a substantial cumulative carbon debt, according to its calculations.
Yassine Charabi, a geographer at Kuwait University, built a mathematical simulation to map out how AI expansion affects the planet over time. He plugged in data such as global energy forecasts, data center growth rates and hardware replacement schedules. The model then factored in everyday electricity use, emissions from grid electricity and the carbon generated during chip manufacturing.
Trapped in the Carbon Valley
Charabi ran the simulation 10,000 times, and the results revealed a long stretch of negative impact he calls the Carbon Valley . This is when building and powering AI produces more carbon pollution than the technology can save.
Under the fastest AI growth scenario, the model shows that AI will stay trapped in this valley for nearly a decade, with AI-enabled savings failing to catch up until late 2031. By that time, the world will have already accumulated a peak cumulative carbon debt of around 2.85 gigatons of carbon dioxide.
But it doesn't have to be this way.
Charabi's model shows that to shrink the Carbon Valley, we will need to accelerate AI deployment in green industrial processes , because each year of delay adds 0.45 gigatons of CO 2 .
The study finds that the climate benefits depend on how much AI is actually tied to emissions-reducing uses. Also, location matters, since cooler climates reduce the amount of cooling-related energy used compared with hotter regions.
The cost of hesitating
Relying on the technology simply to become more efficient will not solve the problem, as Charabi notes in his paper.
"Efficiency improvements alone do not ensure absolute decoupling between AI expansion and electricity demand under rapid deployment pathways."
If these approaches do not come soon, clean energy fixes may arrive too late to reverse the initial damage, he warns. "Under the accelerated AI data-center deployment scenario analyzed here, subsequent reductions in annual emissions alter future trajectories but do not erase earlier additions to cumulative CO₂ within the cumulative-emissions accounting framework adopted in this study."
Written for you by our author Paul Arnold , edited by Gaby Clark , and fact-checked and reviewed by Robert Egan —this article is the result of careful human work. We rely on readers like you to keep independent science journalism alive.
If this reporting matters to you, please consider a donation (especially monthly). You'll get an ad-free account as a thank-you.
Publication details
Yassine Charabi, Rapid artificial intelligence deployment increases near-term pressure on global carbon budgets, Communications Earth & Environment (2026). DOI: 10.1038/s43247-026-03746-y
Who's behind this story?
Paul Arnold
BSc Biology from University of London. BBC documentary producer with world travel experience. Freelances from southern Spain.
Full profile →
Gaby Clark
MA in English, copy editor since 2021 with experience in higher education and health content. Dedicated to trustworthy science news.
Full profile →
Robert Egan
Bachelor's in mathematical biology, Master's in creative writing. Well-traveled with unique perspectives on science and language.
Full profile →
© 2026 Science X Network
Citation :
AI could be trapped in a 'Carbon Valley' unless action is taken soon (2026, June 17)
retrieved 17 June 2026
from https://phys.org/news/2026-06-ai-carbon-valley-action.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
part may be reproduced without the written permission. The content is provided for information purposes only.
Read the full article at Phys.org →