Researchers Identify the 'Hidden Energy Cost' of AI Agents for the First Time
KAIST researchers say agentic AI can raise response latency 153.7 times and leave GPUs idle for 54.5% of execution time.
7 Articles
7 Articles
TechXplore
NanowerkResearchers identify the 'hidden energy cost' of AI agents for the first time
As the era of AI agents—systems that can reason and act autonomously—begins, the power consumption of data centers is emerging as a critical challenge. A KAIST research team has, for the first time, analyzed the computational cost and energy consumption of AI agents, finding that they can consume up to 136.5 times more energy per query than conventional generative AI.
Full-scale deployment of autonomous AI systems will require hundreds of gigawatts of electricity, far beyond the capacity of existing data centers. This is the conclusion reached by experts at the Korea Advanced Institute of Science and Technology (KAIST), who conducted the first comprehensive analysis of the energy footprint of agents capable of autonomously planning and executing multi-step tasks. Unlike classic chatbots, which respond with a …
While technology companies are rushing to embed advanced “AI agents” - systems capable of performing complex tasks independently - new research from South Korea raises significant warning signs about the energetic price of this revolution. The study, conducted at KAIST University, found that AI agents consume up to 136.5 times more electricity per query compared to regular chatbots. The reason for this lies in the nature of their operation: unli…
Coverage Details
Bias Distribution
- 75% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
