AI Industry Repeats Old Energy Sector Mistakes on Water Use, Expert Warns
By Ikenna Omeje
A growing debate over the environmental footprint of artificial intelligence has drawn parallels with the oil and gas industry’s long-standing struggle over resource stewardship, as experts warn that the sector may be underestimating its true water impact.
Steve Coffee, President of the Produced Water Society, has cautioned that the AI industry is replicating historical mistakes made by the energy sector—particularly in how environmental costs are measured and reported.
In a recent post on LinkedIn, Coffee referenced new research from Virginia Tech which found that up to 90 per cent of a data centre’s total water footprint is not generated on-site through cooling systems, but indirectly through the electricity supply chain powering the facilities.
This finding suggests that corporate disclosures focusing on on-site water efficiency may significantly understate the broader environmental impact of AI operations.
“When an Al company says it’s reducing on-site water use, it’s describing a fraction of the actual impact,” Coffee noted.
“The real cost is upstream in the fossil fuel plants, the water-cooled nuclear facilities, and the natural gas infrastructure generating the electricity.”
The observation highlights what analysts describe as an “accounting gap” in the tech industry’s environmental reporting—one that closely mirrors earlier challenges in the oil and gas sector, where full lifecycle emissions and resource use were not immediately incorporated into regulatory or corporate frameworks.
According to Coffee, the energy industry only adopted comprehensive “full-cycle” thinking after years of regulatory pressure and public scrutiny. He warned that AI could face a similar reckoning if sustainability considerations are not embedded early in its rapid expansion.
“What the energy sector learned, mostly through regulatory pressure and community backlash, is that stewardship has to be embedded before the infrastructure scales, not retrofitted after the problems show up,” he said.
The oil and gas industry’s experience with produced water management was cited as a key example. While the volumes of wastewater generated during extraction were always significant, structured governance and environmental oversight lagged behind industrial growth, creating long-term environmental and regulatory challenges.
