Sam Altman plays down AI water-use claims, says energy is the real issue

Speaking during his India visit, the OpenAI CEO said water claims are "totally untrue" but acknowledged that AI's rising power needs make nuclear and renewables more urgent

OpenAI chief executive Sam Altman has pushed back against claims that artificial intelligence (AI) is driving a surge in water consumption, while acknowledging that the technology’s growing energy footprint is a real concern as AI use scales globally. 

Speaking in an interview with The Indian Express during his visit to India for the AI Impact Summit, Altman said claims about water usage per AI query are “completely untrue” and based on outdated assumptions about how data centres are cooled. 

“Water is totally fake. It used to be true. We used to do evaporative cooling in data centres, but now we don’t do that,” Altman said. Referring to widely shared claims online, he added: “You see these things on the internet where, ‘Don’t use ChatGPT, it’s 17 gallons of water for each query’ or whatever. This is completely untrue, totally insane, no connection to reality.”  Altman’s comments come amid a broader debate over the environmental impact of large-scale AI systems, as data centres expand and electricity demand rises to support training and running increasingly large models. 

 

Energy use, not water, is the bigger issue

While dismissing water-related claims, Altman said concerns about energy consumption are “fair”, particularly when looked at in aggregate rather than on a per-query basis. 

“It’s fair to be concerned about the energy consumption — not per query, but in total, because the world is now using so much AI,” he said, adding that this strengthens the case for moving more quickly towards nuclear, wind and solar power. 

Altman also pushed back against comparisons that equate a single AI query to large amounts of energy use, such as claims that one ChatGPT prompt consumes the equivalent of multiple smartphone battery charges. “There’s no way it’s anything close to that much,” he said.

  He argued that discussions around the energy cost of training AI models are often framed unfairly, especially when not compared with the energy and resources required to “train” a human being. 

“People talk about how much energy it takes to train an AI model. But it also takes a lot of energy to train a human,” Altman said. “It takes like 20 years of life and all of the food you eat during that time before you get smart.” 

In his view, the more relevant comparison is how much energy it takes for an AI system, once trained, to answer a question versus a human doing the same task. Measured that way, he suggested, AI may already be comparable in terms of energy efficiency. 

India’s energy context and AI ambitions

Altman’s remarks come at a time when Indian policymakers are increasingly linking AI growth to questions of power and infrastructure. At the AI Impact Summit, Union IT minister Ashwini Vaishnaw said that AI is both energy- and resource-intensive, and that meeting future compute demand will require a significant expansion of clean power and supporting infrastructure. 

Vaishnaw said that India is in a relatively strong position compared to many countries, noting that more than 51 per cent of the country’s installed power generation capacity now comes from clean energy sources. “About 51 per cent of energy of power generation capacity is from clean sources. And that is one big advantage that India has,” he said. 

The minister also indicated that nuclear power is being considered as part of the mix to support the growing compute requirements of AI systems, alongside continued investments in renewable energy. He said power and infrastructure remain a challenge, but the government is working on expanding clean energy capacity and collaborating with startups on new approaches to address these constraints and support AI-led growth.