Sam Altman Highlights Evolutionary Energy Costs Behind Human Intelligence in AI Debate

• February 23, 2026

The ongoing debate about the energy demands of artificial intelligence compared to the human brain has taken a new turn with insights from Sam Altman, CEO of OpenAI. Altman stressed that any fair comparison must consider not only the energy consumption of the brain itself but also the extensive evolutionary and developmental energy inputs required to achieve human-level intelligence.

Energy Investment Beyond Brain Activity

Typically, discussions around AI’s energy footprint focus on the immediate power consumption of training and operating large neural networks versus the biological energy used by the brain. However, Altman argues that this view is incomplete. The human brain’s complexity and capabilities are the product of millions of years of evolution coupled with the substantial caloric intake humans require to reach cognitive maturity—an extended period spanning roughly two decades.

According to Altman, measuring only the brain’s direct energy use overlooks the vast energetic expenditure embodied in the evolutionary process, which includes not just brain development but the entire biological and environmental framework enabling intelligence. He pointed out that humans consume tons of food throughout their growth before reaching full intellectual capacity, which represents a significant, albeit indirect, energy investment.

This perspective reframes the conversation about AI energy efficiency by highlighting that human intelligence is not merely a function of brain activity at a single moment but a cumulative outcome of long-term energy use extending across developmental and evolutionary timescales. It suggests that when AI systems are compared to human cognition, the comparison should account for the entire background of biological and evolutionary energy consumption that supports the brain’s functioning.

The statement from Altman arrives amid growing public and scientific interest in the environmental implications of large-scale AI deployment, where the scale of computational power needed for advanced models draws intense scrutiny. His comments invite a broader understanding of what constitutes “intelligence” costs, encouraging the AI community to consider energy use more holistically.

In doing so, Altman’s remarks underscore a distinction between the immediate operational energy that AI systems demand and the extended, often less visible, energetic investment underpinning natural intelligence. This context may influence future discussions on AI development, environmental impact, and the design of more sustainable computational models.

While specific numerical comparisons or energy metrics were not disclosed, the emphasis on evolutionary and developmental factors adds complexity to the dialogue on AI efficiency. It encourages deeper reflection on how intelligence, whether artificial or biological, relates to energy consumption over time.

Sam Altman compares the energy consumption of AI to humans, emphasizing evolutionary and developmental energy investments needed for intelligence.