Tech Desk 
Nov 22, 2025
An undated image. — Unsplash
In a remarkable feat at the intersection of AI and eco-friendly technology, researchers at the University of California, Riverside (UCR) have developed a new artificial intelligence (AI) tool that reduces emissions by 45% during data centre operations, while extending server lifespan by approximately 1.6 years.
The new AI-driven system optimises the flow of compute-heavy tasks through data centre infrastructure by monitoring hardware conditions, energy profiles, and cooling performance in real time.
Instead of just evenly distributing workloads, the AI system routes tasks through the most efficient pathways in order to minimise power consumption and heat build-up while ensuring high performance.
This approach is believed to address cost and sustainability challenges faced by large-scale computing.
“Our results show that sustainability in AI cannot rely solely on clean energy,” stated Mihrimah Ozkan. To effectively transform idle capacity into an asset, the UCR tool utilises periods of lower-carbon electricity and cooler conditions to reduce wear on critical components.
The UCR system has emerged as a need of the hour as data centres are a rapidly growing contributor to global energy consumption, and the International Energy Agency warns that emissions could surge without efficiency improvements.
UCR researchers stated the system is ready for pilot deployment in commercial data centres this year.
