Sam Altman debunks myth that a single ChatGPT query wastes six hours of shower water

When a meme started circulating that every time you ask ChatGPT a question you’re essentially running a six‑hour shower, the claim quickly snowballed into headlines and angry comments. Sam Altman, the chief executive of OpenAI, finally weighed in on the debate during an episode of the Sources podcast.

Altman laughed at the notion before getting to the numbers. “I don’t have the exact calculation in front of me, but from memory it’s something like… for every 38,000 ChatGPT queries, you use about the same amount of water it takes to produce a single almond in California,” he said, chuckling.

In other words, a lone query consumes a minuscule slice of water – nowhere near the volume of a prolonged shower. To put it in perspective, the water used to grow a handful of almonds is something most people never think about, yet it is comparable to the energy needed for a single AI response.

Altman also reminded listeners that the myth rests on an outdated view of how data centres cool their hardware. “Older facilities relied on evaporative cooling, which can waste a lot of water,” he explained. “Modern large‑scale centres are built with sophisticated closed‑loop cooling systems, and their total water use is roughly on par with a regular office building – sinks, toilets and all.”

He went on to point out the double standard that often appears in sustainability discussions. “People snack on twelve almonds without a second thought about the water behind them, yet the same scrutiny is dumped on AI because it’s new and visible,” Altman observed.

While the conversation around AI’s carbon and water footprints is certainly important, Altman’s remarks suggest that the specific claim of a “six‑hour shower per query” is more myth than metric. As the industry continues to adopt greener hardware and smarter cooling, the actual environmental impact per request is expected to shrink even further.