Speaking on the\u00a0Sources Podcast with Alex Heath, Altman said modern data centers have largely moved away from evaporative cooling systems.<\/p>\n
He claimed that producing a single almond in California uses roughly as much water as 38,000 ChatGPT<\/a> queries.<\/p>\n \u201cFor every 38,000 ChatGPT queries, that is the same amount of water that is used in the production of a single almond in California,\u201d Altman said, adding that the estimate includes \u201cfull-on\u201d water accounting rather than water use at a single data center.<\/p>\n Altman also rejected widely circulated comparisons suggesting that one ChatGPT query consumes as much water as running a shower for hours. He said such claims have become a persistent narrative that is difficult to disprove but do not reflect the cooling technologies used in modern data centers.<\/p>\n Sam Altman on OpenAI\u2019s next model and the AI backlash<\/p>\n I recently sat down with Sam for a two-part conversation. We covered a lot of ground: the need to slow down frontier research, alignment concerns, OpenAI’s next big model, Astra, competition with Anthropic, his thoughts on an\u2026 pic.twitter.com\/kfz4UrAKBm<\/a><\/p>\n \u2014 Alex Heath (@alexeheath) September 1, 2026<\/a><\/p>\n According to Altman, a modern large data center uses roughly as much water as an office building, mainly for everyday use.<\/p>\n The podcast discussion also covered OpenAI\u2019s next major model, the company\u2019s Astra project, competition with Anthropic, artificial general intelligence, AI safety and alignment, consumer hardware, frontier research and Altman\u2019s views on a potential initial public offering.<\/p>\n Earlier, Qazinform News Agency<\/a> reported that users worldwide had reported<\/a> disruptions affecting ChatGPT and several other artificial intelligence services.<\/p>\n