Episode
Guess How Much Water It Takes To Ask ChatGPT A Question | Really? no, Really?
- Published
- Feb 17, 2026
- Duration seconds
- 1697
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/really-no-really-with-jason-alexander-peter-tilden-6004825/episodes/guess-how-much-water-it-takes-to-ask-chatgpt-a-question-really-no-really/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/really-no-really-with-jason-alexander-peter-tilden-6004825/guess-how-much-water-it-takes-to-ask-chatgpt-a-question-really-no-really.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Some AI data centers known as "hyperscalers" need to consume as much water as the whole city of Philadelphia. On this this episode, Jason Alexander and Peter Tilden ask Eric Olson, the assistant professor of finance and the director of the Center of Energy Studies at the University of Tulsa, to not only reveal how much much water it takes for ChatGPT to answer one question but to also address why building an AI data center in some place like Phoenix, Arizona isn't a good option. They also address the concerns of rising energy costs from these AI data centers being passed on to the consumer and if recycled water, or greywater, could be a potential solution. Would it make sense to move the data centers to areas with colder climates like Alaska or Antartica to handle the evaporative cooling? Eric also clarifies what we've all heard about job creation during the developmental phase. The conversation also delves into why businesses like Netflix spend so much money on servers to keep up with the attention society and the demand society. Learn more about your ad choices. Visit megaphone.fm/adchoices