Story perspectives

Revolutionizing AI: DARPA and Google Tackle Energy Efficiency

8/22/2025

41 8

Summary

Story summary

DARPA's ML2P program seeks to create energy-efficient machine learning for battlefield applications, focusing on reducing electricity demands. Briefings are scheduled for August 26, 2025, in Arlington, VA, and online.
Google has developed a method to assess the energy and carbon footprint of AI models, resulting in notable reductions in consumption and emissions.
Research on optimizing matrix multiplication improves speed and energy efficiency for AI, utilizing mixed-precision computing across different hardware.

1 of 1

Story summary

DARPA's ML2P program seeks to create energy-efficient machine learning for battlefield applications, focusing on reducing electricity demands. Briefings are scheduled for August 26, 2025, in Arlington, VA, and online.
Google has developed a method to assess the energy and carbon footprint of AI models, resulting in notable reductions in consumption and emissions.
Research on optimizing matrix multiplication improves speed and energy efficiency for AI, utilizing mixed-precision computing across different hardware.