Menu

Taming the Energy Appetite of AI Models

AI computing lacks transparency on carbon emissions, but Lincoln Laboratory’s energy-efficient methods could help check their appetite for energy.

As AI models increasingly consume more energy, especially in data centers, it’s vital to adopt energy-efficient tools to mitigate their impact. The Lincoln Laboratory Supercomputing Center is developing methods to help data centers reduce energy use, from simple measures like power-capping hardware to advanced techniques to halt AI training when it has reached its goal. By demonstrating that these strategies have minimal impact on model performance and could significantly cut energy consumption, it is hoped to spur widespread adoption.

Lincoln Laboratory Supercomputing Center

Research projects

Towards a Whole Brain Cellular Atlas
Microplastic-Free by Design
Supercomputers Reveal Ancient Atmospheric Battle
Evolution of Viral Infectious Disease
Ephemeral Stream Water Contributions to US Drainage Networks
Adaptive Deep Learning Systems Towards Edge Intelligence
Towards a Self-driving Lab?
AI That Speaks Human About Health
All Research Projects

Collaborative projects

ALL Collaborative PROJECTS

OUTREACH & EDUCATION PROJECTS

See ALL Scholarships
100 Bigelow Street, Holyoke, MA 01040