How Much Water Does ChatGPT Use? Sam Altman Reveals

Sam Altman Reveals How Much Water and Power ChatGPT Uses Per Query

AI tools like ChatGPT are now a daily part of our digital lives—but behind their smart replies lies a massive infrastructure that consumes more resources than most people realize. Recently, OpenAI CEO Sam Altman gave the world a rare glimpse into just how much energy and water it takes to generate a single AI response.

In a new blog post, Altman shared that a single query on ChatGPT uses about 0.34 watt-hours of electricity. To put it in perspective, that’s roughly equivalent to turning on a small LED bulb for a few minutes. It might seem insignificant on its own, but multiply that by billions of queries per day, and the energy use becomes staggering.

Even more surprising is the water usage. According to Altman, answering just one question requires the equivalent of a teaspoon of water—used in cooling the powerful servers that run these AI models. Again, it sounds small, but scaled up, it represents thousands of gallons consumed each day just to keep AI systems operational.

These numbers haven’t been independently verified, but they have sparked fresh concerns about the environmental impact of large-scale AI deployment. Data centers, already known for their heavy electricity demands, are now being scrutinized for their hidden water footprints as well.

Altman believes that as technology evolves, resource consumption will decrease—just like it did in past tech revolutions. But many experts argue that action needs to be taken now, not later. With water shortages becoming a global issue, even a teaspoon per interaction adds up.

This revelation is a wake-up call for both users and developers. While AI continues to push boundaries in productivity and innovation, the sustainability question cannot be ignored. If we’re to truly embrace a tech-driven future, it must be one that balances progress with environmental responsibility.

As OpenAI and other tech giants move forward, the hope is that they’ll invest just as much in green infrastructure as they do in model development. After all, the intelligence of our future machines should also reflect in the choices we make for our planet.