OpenAI CEO Sam Altman has a novel argument in relation to the controversy of the quantity of power assets it takes whereas coaching AI fashions.
Talking at an occasion on Friday, Altman dismissed the useful resource calls for of synthetic intelligence, arguing that claims in opposition to OpenAI utilizing gallons of water to coach its AI mannequin ChatGPT had been faux.
“We used to do evaporative cooling in knowledge centres, however now that we do not try this. You see these like issues on the web the place ‘do not use ChatGPT, it is 17 gallons of water for every question’ or no matter… That is fully unfaithful. Completely insane. No connection to actuality,” Altman mentioned on the Specific Adda occasion in the course of the India AI Affect Summit 2026.
Information centres historically use enormous quantities of water to forestall overheating and funky down electrical elements. Water know-how firm Xylem and World Water Intelligence mentioned in a report final month that the water used for cooling will greater than triple within the subsequent 25 years as AI compute demand rises.
Whereas Sam Altman dismissed the water utilization claims, he warned that the exponential use of AI calls for that we transfer shortly in the direction of renewable sources of power.
“What’s truthful although is the power consumption, not per question, however in complete as a result of the world is now utilizing a lot AI is actual and we have to transfer in the direction of nuclear or wind and photo voltaic in a short time,” he mentioned.
Sam Altman compares AI and human power makes use of
Talking additional on the occasion, Sam Altman additional mentioned that evaluating AI power utilization to that of people was unfair, arguing that it takes 20 years of life and power to coach a human being.
“One of many issues that’s all the time unfair on this comparability is individuals discuss how a lot power it takes to coach an AI mannequin relative to how a lot it prices a human to do one inference question. Nevertheless it additionally takes lots of power to coach a human. It takes like 20 years of life and the entire meals you eat throughout that point earlier than you get sensible,” Altman mentioned.
He argued that AI has most likely caught up on power effectivity if we evaluate it to how a lot power people have consumed when it comes to answering a query.
“And never solely that, it took just like the very widespread evolution of the hundred billion those who have ever lived and discovered to not get eaten by predators and discovered tips on how to like work out science and no matter to provide you and then you definitely took no matter you you recognize you took. So the truthful comparability is in case you ask ChatGPT a query, how a lot power does it take as soon as its mannequin is skilled to reply that query versus a human? And possibly AI has already caught up on an power effectivity foundation measured that method,” he mentioned.
Altman’s feedback had been obtained with criticism on the web, with netizens bashing him for his opinion.
Zoho founder Sridhar Vembu additionally lashed out at Altman, saying, “I don’t wish to see a world the place we equate a chunk of know-how to a human being.”