Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124

“People are often Curiosity about how much energy a chatzipt query uses, ” Sam AltmanCEO OpenWrote on one side Long blog post Last week. Altman writes that the average query uses 0.34 watt-hour energy: “A stove will use the key for more than a second, or a high-star Lightblb in a few minutes.”
For a company with 800 million weekly active user (and Growing), The question of how much energy is using all these searchs is becoming increasingly under pressure. However, experts say that Ultman’s image does not mean much without much public context on how it came in in the opening – with the definition of “average” what it is, whether it incorporated the generation of the image, and the use of the AI models to train and cool the opener’s server.
As a result, the hug of the AI company Sasha Luccioni does not keep too much stock in the number of Ultman. “He could have pulled it out of his ass,” he says. (Opina did not respond to any request for more information about how it came in this number))
As AI accepts our lives, we try to fight the climate change as well as the promise of transforming our energy system, supercharging carbon emissions. Now, a new and growing company of research is trying to keep a harsh number on how much carbon we are actually emitting with all our AI use.
This attempt is complicated by the fact that the main players like Openai publish little environmental information. An analysis submitted for Peer Review this week focuses on the need for more environmental transparency in AI models. In the new analysis of Lukioni, he and his colleagues use data from OpenRUtorGreater language model (LLM) traffic is a leadersboard, 84 percent of LLM use in May 2025 was for zero environmental manifestations. This means that customers are choosing models with completely unknown environmental impacts.
“It blows my mind that you can buy a car and find out how many miles per gallon, we use all these AI tools every day and we have any skills metric, emissions, nothing,” said Luccioni. “It’s not compulsory, it’s not regulator. Given where we are with the climate crisis, it should be at the top of the agenda for the regulators everywhere.”
As a result of this lack of transparency, Lucyoni says people are facing assumptions that do not feel any meaning but what is taken as a gospel. For example, you may have heard that the average chatzipt request takes 10 times more energy than the average Google search. Luxioni and his colleagues have identified the claim in a public comment that Google’s main body, John Hainsi, was created in 2021.
He has nothing to do with the product of a company (Google) with which he has nothing to do with (OpenAI) is best – however, Lucy’s analysis shows that this image has repeatedly repeated in press and policy reports. (I got a pitch with these accurate figures when I was writing this piece))