Sam Altman has pushed back against mounting criticism over the enormous resource demands of artificial intelligence, particularly concerns surrounding the water consumed by data centres powering AI systems.
Sam Altman has pushed back against mounting criticism over the enormous resource demands of artificial intelligence, particularly concerns surrounding the water consumed by data centres powering AI systems. The OpenAI CEO has dismissed viral claims about ChatGPT’s water footprint, arguing that public alarm over the fresh water required to cool large-scale data centres is overstated and fails to account for advances in modern cooling technology.
Speaking on the Sources Podcast, Altman addressed growing resistance to the rapid expansion of AI infrastructure and challenged the viral claim that a single ChatGPT prompt consumes as much water as leaving a household shower running for six hours.
Calling the narrative an online myth, Altman suggested that the actual water consumption associated with individual ChatGPT queries is dramatically lower. According to his estimate, around 38,000 ChatGPT queries consume an amount of water comparable to what is required to produce a single almond in California.
“I saw this thing going around about the water usage of ChatGPT and it was like, ‘every time you run a single ChatGPT query, it’s like you run your shower for like 6 hours and the water never comes back and it’s just it’s done’. I don’t have the exact calculation in front of me, but I think the real number is something like doing this from memory, it might be wrong. For every 38,000 ChatGPT queries, that is the same amount of water that is used in the production of a single almond in California,” he said on the podcast .
Altman also questioned the scrutiny directed at AI infrastructure, pointing to the everyday consumption of water-intensive products such as almonds. He argued that people can consume several almonds without considering their water footprint, while AI systems face intense criticism over comparatively small per-query consumption.
“There’s a question, ‘where this came from’ because the people that are scarfing down 12 almonds at a time don’t feel like they’re doing something horrible from a water perspective. It is true that data centers at one point used evaporative cooling but they have not done that in a long time,” he noted.
He further argued that perceptions surrounding data-centre water consumption are rooted in outdated assumptions about how these facilities operate. According to Altman, older data centres relied more heavily on evaporative cooling, which could result in significant water loss, whereas newer facilities have adopted more efficient cooling infrastructure.
Altman claimed that a modern large-scale data centre’s water consumption is broadly comparable to that of a conventional office building, including water used by employees for sinks and toilets.
“If you look at a modern very large data center, it uses the equivalent amount of water as an office building in terms of people running the sinks and the toilets and whatever. So that has been a robust meme and difficult to disprove, but I don’t think it holds up to any scrutiny,” Altman concluded .