OpenAI CEO Sam Altman has urged the global tech industry to accelerate the transition to solar and nuclear energy as the country is rapidly establishing large-scale data centres to power the next generation of AI models. Speaking at the Express Adda interview hosted by Anant Goenka of The Indian Express Group, Altman addressed growing concerns over the environmental footprint of large-scale AI operations, particularly water and electricity usage in data centres.
When asked about the environmental impact of running large AI data centres, Altman dismissed exaggerated claims about water consumption, stating, “We used to do evaporative cooling in data centres, but ChatGPT queries taking up gallons of water is fake.” He clarified that while individual query water usage has been overstated in public discourse, the broader issue of total energy consumption remains critical. “What is fair though, is the energy consumption not per query but in total and we need to move to solar, nuclear energy quickly,” he added.
Altman’s comments come at a time when AI companies face increasing scrutiny over their carbon and resource intensity. Training and running frontier models like those powering ChatGPT requires enormous amounts of electricity, which is often equivalent to the consumption of small cities.
Altman addresses misinformation on water usage for coolant
Altman acknowledged that early media reports had sensationalised per-query water figures, sometimes citing outdated or misapplied evaporative cooling statistics. Modern data centres, including those used by OpenAI’s partners, increasingly employ advanced cooling techniques such as liquid cooling, free-air cooling in suitable climates, and closed-loop systems that drastically reduce water loss.
However, he stressed that the real sustainability challenge lies in the overall electricity demand, which is projected to grow exponentially as AI adoption scales. Altman advocated for a swift move to low-carbon, high-output energy sources, specifically solar for scalable renewable capacity and nuclear for reliable, high-density baseload power, to decouple AI growth from fossil fuel dependency.
Altman recommends new energy opportunities for India’s AI ambitions
During the discussion, Altman highlighted India’s potential to lead in sustainable AI infrastructure, given its ambitious renewable targets and interest in small modular reactors (SMRs). He suggested that pairing clean energy expansion with AI development could position countries like India as global leaders in responsible technological progress, which is currently the broader industry effort. Several major tech firms have already signed long-term power purchase agreements for solar, wind, and emerging nuclear projects to offset their data centre footprints.