Shocking ChatGPT water usage cost 1 query uses 1/15th teaspoon

Key Points

• ChatGPT water usage revealed 1 query uses 1/15th teaspoon
• Sam Altman shared the claim in a new blog post
• Energy use equals 0.34 watt-hours per query
• AI industry faces growing scrutiny on resource use

ChatGPT water usage is the latest surprise fact shared by OpenAI CEO Sam Altman. In a recent blog post, Altman claimed that each ChatGPT query consumes about 0.000085 gallons of water—roughly one fifteenth of a teaspoon.

This small detail was part of a larger post where Altman discussed his vision for AI’s future. He also explained that an average ChatGPT query uses about 0.34 watt-hours of electricity.

For comparison, that’s similar to what an oven uses in just over a second or what a high-efficiency lightbulb uses in a couple of minutes.

But it’s the water number that caught the eye of many in the tech community. Altman didn’t provide a direct source for the figure, and OpenAI hasn’t commented yet on how it was calculated. Still, the statistic highlights an important and growing discussion around AI’s environmental impact.

Sam Altman claims an “average” ChatGPT query uses about 0.34 watt-hours and about 0.000085 gallons of water, or “roughly one fifteenth of a teaspoon” (@jaypeters / The Verge)https://t.co/nBsrYf0IXBhttps://t.co/06QCtVhGEjhttps://t.co/ZOzeer1FAj

— Techmeme (@Techmeme) June 10, 2025

AI power use sparks comparisons with Bitcoin mining

As AI models like ChatGPT grow more powerful and popular, their resource usage is drawing attention. Water is used to cool the data centers that run these large models. While one query may use just a fraction of a teaspoon, the scale of global usage adds up fast.

Researchers have already warned that AI could soon outpace Bitcoin mining in energy consumption. According to a 2024 forecast, AI power demand may exceed that of crypto mining before year’s end.

Last year, The Washington Post reported that generating a simple 100-word AI email could use more than a bottle of water.

“A ChatGPT query uses; the average query uses about 0.34 watt-hours, about what an oven would use in a little over one second, or a high-efficiency lightbulb would use in a couple of minutes. It also uses about 0.000085 gallons of water; roughly one fifteenth of a teaspoon.”… pic.twitter.com/BF3Qrfntx8

— EN (@EarningsNugget) June 11, 2025

Water usage can also vary by data center location. In areas where water is scarce or cooling systems are less efficient, the environmental cost of AI could be higher than users realize.

Adding to this scrutiny, Nvidia recently warned the UK that current AI infrastructure is not prepared to handle the soaring demand for power and cooling. Meanwhile, tools like OpenAI’s O3 Pro model and emerging deep research agents such as OpenAI’s new deep research AI agent are pushing the boundaries of what AI can do—while increasing resource consumption.

Altman’s post suggests he expects the cost of intelligence to drop over time, ideally nearing the cost of electricity. But critics argue that this ignores the broader environmental footprint, which includes water and other resources.

Sam Altman @sama claims an average ChatGPT query uses ‘roughly one fifteenth of a teaspoon’ of water

“People are often curious about how much energy a @ChatGPTapp query uses; the average query uses about 0.34 watt-hours, about what an oven would use in a little over one second,… pic.twitter.com/dwgO5XttkW

— Water Mark 🚰 (@OtayMark) June 10, 2025

As the AI industry races forward, transparency about these impacts will be key. Figures like ChatGPT water usage may sound small in isolation, but multiplied by millions of daily queries, they add up.

For now, Altman’s blog post has sparked another round of debate on what it truly costs to power the future of intelligence.

Why transparency matters in AI resource consumption

The tech world is increasingly calling for more transparency around how much energy and water large AI models consume. While Altman’s blog post offered a glimpse into ChatGPT water usage, many are asking for more concrete data and independent verification.

Environmental advocates argue that AI companies should publish regular reports detailing the resource usage of their data centers. Without clear information, it’s hard to weigh the true environmental cost of using AI tools that have become part of daily life for millions.

Transparency is also becoming more important as legal and ethical questions grow. Just this week, Reddit sued Anthropic over concerns about how AI models are trained on publicly available content—another sign that the industry is moving into more regulated and scrutinized territory.

There’s also a growing push to develop more sustainable AI systems. Innovations in data center cooling, more efficient hardware, and energy sourcing from renewables could help reduce the footprint.

Until then, even small figures like a teaspoon of water per query serve as a reminder that every digital interaction has a physical cost.