AI chief addresses growing concerns about the environmental impact of artificial intelligence
OpenAI CEO Sam Altman has disclosed new details about the energy consumption of ChatGPT queries, offering rare insight into the environmental footprint of artificial intelligence systems. His comments come amid rising concerns about the sustainability of rapidly expanding AI technologies.
In a blog post titled “The Gentle Singularity,” published on June 11, Altman broke down the resource demands of a typical ChatGPT interaction. According to his calculations, an average query consumes approximately 0.34 watt-hours of electricity, equivalent to running a high-efficiency lightbulb for a few minutes or an oven for just over one second.
Water and energy usage under scrutiny
Altman also revealed that each ChatGPT query uses about 0.000085 gallons of water, roughly one-fifteenth of a teaspoon. While these figures may seem negligible for individual interactions, they accumulate significantly given ChatGPT’s hundreds of millions of daily users.
“People are often curious about how much energy 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,” Altman wrote.
His disclosure follows warnings from experts that AI’s energy demands could soon rival—or even exceed—those of cryptocurrency mining, raising sustainability concerns as tech giants race to develop more powerful models.
A future of “wildly abundant” intelligence and energy
Looking ahead, Altman predicted a dramatic shift in the 2030s, where intelligence and energy become “wildly abundant.” He suggested this could unlock unprecedented human progress, from breakthroughs in physics to space colonization and advanced brain-computer interfaces.
“With abundant intelligence and energy (and good governance), we can theoretically have anything else,” he wrote.
However, he also acknowledged the risks of AI’s growing influence, noting that even minor flaws in such a widely used system could have massive consequences.
“A small new capability can create a hugely positive impact; a small misalignment multiplied by hundreds of millions of people can cause a great deal of negative impact,” he cautioned.
AI costs plummeting, capabilities expanding
Altman reiterated his previous prediction that AI costs will continue dropping exponentially, potentially by 10 times every year, making the technology increasingly accessible.
“You can see this in the token cost from GPT-4 in early 2023 to GPT-4o in mid-2024, where the price per token dropped about 150x in that period,” he noted, comparing the trend to Moore’s Law but with far greater momentum.
As AI systems become increasingly powerful and widespread, Altman’s remarks underscore both their transformative potential and the urgent need to manage their environmental and societal impact responsibly.