Sam Altman would like to remind you that humans use a lot of energy, too

Source: TechCrunch

Event and Statements

OpenAI CEO Sam Altman addressed concerns about AI’s environmental impact while speaking at an event hosted by The Indian Express (video).

Altman, who was in India for a major AI summit (TechCrunch coverage), said that claims about AI’s water usage are “totally fake.” He noted that water consumption was a real issue when data centers relied on evaporative cooling, but that practice has largely been abandoned.

“Now that we don’t do that, you see these things on the internet where, ‘Don’t use ChatGPT, it’s 17 gallons of water for each query’ … This is completely untrue, totally insane, no connection to reality.”

Altman acknowledged that concerns about total energy consumption are “fair,” emphasizing that the world’s overall AI usage is growing rapidly. He argued that the solution is to transition quickly to low‑carbon power sources such as nuclear, wind, and solar.

Legal and Industry Context

• There is no legal requirement for tech companies to disclose their energy or water usage, so independent researchers are attempting to study the impact (NPR).
• Data centers have also been linked to rising electricity prices (TechCrunch).

When asked whether a single ChatGPT query uses the equivalent of 1.5 iPhone battery charges—a figure cited in a previous conversation with Bill Gates—Altman replied, “There’s no way it’s anything close to that much.”

Comparison with Human Energy Use

Altman criticized discussions that focus solely on the energy required to train AI models, calling them “unfair.” He highlighted that human cognition also consumes substantial resources over a lifetime:

“It takes like 20 years of life and all of the food you eat during that time before you get smart. And not only that, it took the very widespread evolution of the 100 billion people that have ever lived … to produce you.”

He suggested the more relevant comparison is the energy needed for a single inference (a ChatGPT query) versus the energy a human expends to answer the same question. In his view, AI may already be comparable or more efficient on that basis.

Full Interview

The conversation about water and energy usage begins at around 26:35 in the interview video linked above.