‘Humans use lot of energy too’: Sam Altman on resources consumed by AI, data centres

OpenAI CEO Sam Altman has entered the fray on one of AI's most contentious issues: its massive energy appetite. Speaking at a recent event, Altman directly countered critics who point to the enormous electricity demands of AI data centers powering models like GPT-4 and Sora. His central argument pivots the conversation, stating that human civilization already consumes a staggering amount of energy for various activities, and that AI's consumption should be viewed in this broader context.

Altman didn't dismiss the scale of the problem. He acknowledged that future AI systems will require even more computational power, necessitating a major energy breakthrough. He specifically pointed to advanced nuclear fusion as a critical solution, a technology he has personally invested in through companies like Helion Energy. This aligns with his previous warnings that an AI-driven future is fundamentally an energy-limited one, and that scaling current models requires rethinking our power infrastructure.

The implications are significant for both policy and public perception. As governments and environmental groups scrutinize tech's carbon footprint, Altman's comments reframe AI not just as a consumer, but as a potential tool for optimizing energy use across other industries. The debate now centers on whether the benefits of advanced AI—in scientific discovery, efficiency gains, and problem-solving—justify its substantial energy investment, and who bears the responsibility for creating a sustainable path forward.