Earlier this year, when my colleague Casey Crownhart and I spent six months researching the climate and energy burden of AI, we came to see one number in particular as our white whale: how much energy the leading AI models, like ChatGPT or Gemini, use up when generating a single response.
This fundamental number remained elusive even as the scramble to power AI escalated to the White House and the Pentagon, and as projections showed that in three years AI could use as much electricity as 22% of all US households.
The problem with finding that number, as we explain in our piece published in May, was that AI companies are the only ones who have it. We pestered Google, OpenAI, and Microsoft, but each company refused to provide its figure. Researchers we spoke to who study AI’s impact on energy grids compared it to trying to measure the fuel efficiency of a car without ever being able to drive it, making guesses based on rumors of its engine size and what it sounds like going down the highway.
But then this summer, after we published, a strange thing started to happen. In June, OpenAI’s Sam Altman wrote that an average ChatGPT query uses 0.34 watt-hours of energy. In July, the French AI startup Mistral didn’t publish a number directly but released an estimate of the emissions generated. In August, Google revealed that answering a question to Gemini uses about 0.24 watt-hours of energy. The figures from Google and OpenAI were similar to what Casey and I estimated for medium-size AI models.
So with this newfound transparency, is our job complete? Did we finally harpoon our white whale, and if so, what happens next for people studying the climate impact of AI? I reached out to some of our old sources, and some new ones, to find out.
The numbers are vague and chat-only
The first thing they told me is that there’s a lot missing from the figures tech companies published this summer.
OpenAI’s number, for example, did not appear in a detailed technical paper but rather in a blog post by Altman that leaves lots of unanswered questions, such as which model he was referring to, how the energy use was measured, and how much it varies. Google’s figure, as Crownhart points out, refers to the median amount of energy per query, which doesn’t give us a sense of the more energy-demanding Gemini responses, like when it uses a reasoning model to “think” through a hard problem or generates a really long response.
The numbers also refer only to interactions with chatbots, not the other ways that people are becoming increasingly reliant on generative AI.
“As video and image becomes more prominent and used by more and more people, we need the numbers from different modalities and how they measure up,” says Sasha Luccioni, AI and climate lead at the AI platform Hugging Face.
This is also important because the figures for asking a question to a chatbot are, as expected, undoubtedly small—the same amount of electricity used by a microwave in just seconds. That’s part of the reason AI and climate researchers don’t suggest that any one individual’s AI use creates a significant climate burden.
A full accounting of AI’s energy demands—one that goes beyond what’s used to answer an individual query to help us understand its full net impact on the climate—would require application-specific information on how all this AI is being used. Ketan Joshi, an analyst for climate and energy groups, acknowledges that researchers don’t usually get such specific information from other industries but says it might be justified in this case.
“The rate of data center growth is inarguably unusual,” Joshi says. “Companies should be subject to significantly more scrutiny.”
We have questions about energy efficiency
Companies making billion-dollar investments into AI have struggled to square this growth in energy demand with their sustainability goals. In May, Microsoft said that its emissions have soared by over 23% since 2020, owing largely to AI, while the company has promised to be carbon negative by 2030. “It has become clear that our journey towards being carbon negative is a marathon, not a sprint,” Microsoft wrote.
Tech companies often justify this emissions burden by arguing that soon enough, AI itself will unlock efficiencies that will make it a net positive for the climate. Perhaps the right AI system, the thinking goes, could design more efficient heating and cooling systems for a building, or help discover the minerals required for electric-vehicle batteries.
But there are no signs that AI has been usefully used to do these things yet. Companies have shared anecdotes about using AI to find methane emission hot spots, for example, but they haven’t been transparent enough to help us know if these successes outweigh the surges in electricity demand and emissions that Big Tech has produced in the AI boom. In the meantime, more data centers are planned, and AI’s energy demand continues to rise and rise.
The ‘bubble’ question
One of the big unknowns in the AI energy equation is whether society will ever adopt AI at the levels that figure into tech companies’ plans. OpenAI has said that ChatGPT receives 2.5 billion prompts per day. It’s possible that this number, and the equivalent numbers for other AI companies, will continue to soar in the coming years. Projections released last year by the Lawrence Berkeley National Laboratory suggest that if they do, AI alone could consume as much electricity annually as 22% of all US households by 2028.
But this summer also saw signs of a slowdown that undercut the industry’s optimism. OpenAI’s launch of GPT-5 was largely considered a flop, even by the company itself, and that flop led critics to wonder if AI may be hitting a wall. When a group at MIT found that 95% of businesses are seeing no return on their massive AI investments, stocks floundered. The expansion of AI-specific data centers might be an investment that’s hard to recoup, especially as revenues for AI companies remain elusive.
One of the biggest unknowns about AI’s future energy burden isn’t how much a single query consumes, or any other figure that can be disclosed. It’s whether demand will ever reach the scale companies are building for or whether the technology will collapse under its own hype. The answer will determine whether today’s buildout becomes a lasting shift in our energy system or a short-lived spike.