A groundbreaking new study by KAIST has quantified the hidden electricity costs of advanced AI agents for the first time. These autonomous systems consume up to 136.5 times more energy per query than conventional generative AI tools.
To complete a complex task, the agent has to talk to itself and rerun its core AI programming over and over — meaning answers can take 153.7 times longer to generate.
Expensive computer graphics processors sit idle more than half the time, burning electricity while simply waiting for external websites and apps to respond.
The KAIST research team, led by Distinguished Professor Minsoo Yoo, found that a single complex request to an AI agent burns through an average of 348.41 watt-hours of electricity.
If these agents become mainstream and handle an estimated 13.7 billion requests globally per day, total data center electricity demand would skyrocket to roughly half of the average energy consumption of the entire United States.
The study was presented at IEEE HPCA 2026. Benchmark code was released open source.
Professor Yoo emphasized that competition in the AI era is shifting from “smarter AI” to “more efficient AI.”