While dismissing extreme claims, Altman admitted that total AI energy consumption is a legitimate issue as global usage accelerates. The OpenAI CEO said it is “fair” to question aggregate power demand, adding that the solution lies in rapidly expanding nuclear, wind, and solar energy capacity to support AI’s growth sustainably.

Altman also criticized what he called “unfair” comparisons between AI model training costs and individual human tasks. He argued that evaluating efficiency should compare a trained AI answering a query versus a trained human doing the same task — not the decades of human development versus a one-time model training cycle. As AI adoption expands worldwide, he suggested the conversation should shift from viral energy myths to infrastructure modernization and long-term clean energy investment.