Rethinking Computing: Researchers Tackle AI’s Energy Demands

USC researchers are reimagining how AI systems are trained and powered — through smarter algorithms, innovative hardware, and brain-inspired designs — to dramatically reduce computing’s energy footprint.

2 minute read

May 4, 2025, 11:00 AM PDT

By Clement Lau


A large Google data center building in the Netherlands.

Hilda Weges / Adobe Stock

As artificial intelligence (AI) advances, so too does the need for vast computing power—raising urgent questions about energy consumption. At USC’s School of Advanced Computing and Viterbi School of Engineering, researchers are exploring ways to make computing for AI more energy-efficient. Strategies include trimming and simplifying neural networks, transferring learning from pre-trained models, and reducing the bit-depth of data used in machine learning processes. These methods can dramatically shorten training times while lowering hardware requirements and computational costs, according to Professor Massoud Pedram.

Hardware innovation is another key frontier. While GPUs dominate AI processing, their high cost and energy demands have led researchers like Professor Murali Annavaram to suggest pairing them with CPUs to improve efficiency. Other labs are customizing chips to better match the needs of specific AI tasks, optimizing how data is processed and stored. A more radical approach comes from Professor Peter Beerel, whose lab is pioneering superconducting electronics using niobium to eliminate resistance and enable ultra-low-power computing, supported by new algorithms and design automation tools.

Looking to the future, USC researchers are also pushing the boundaries with neuromorphic computing — systems modeled on the human brain. Professors like Joshua Yang and Barath Raghavan are developing technologies like analog in-memory chips and “information batteries” that store energy as pre-computed data. These innovations promise to dramatically reduce energy use while maintaining or even enhancing performance, signaling a transformative shift in how computing systems are designed to meet the growing demands of AI, edge devices, and real-world applications.

Wednesday, April 30, 2025 in USC News

portrait of professional woman

I love the variety of courses, many practical, and all richly illustrated. They have inspired many ideas that I've applied in practice, and in my own teaching. Mary G., Urban Planner

I love the variety of courses, many practical, and all richly illustrated. They have inspired many ideas that I've applied in practice, and in my own teaching.

Mary G., Urban Planner

Use Code 25for25 at checkout for 25% off an annual plan!

Redlining map of Oakland and Berkeley.

Rethinking Redlining

For decades we have blamed 100-year-old maps for the patterns of spatial racial inequity that persist in American cities today. An esteemed researcher says: we’ve got it all wrong.

May 15, 2025 - Alan Mallach

Logo for Planetizen Federal Action Tracker with black and white image of U.S. Capitol with water ripple overlay.

Planetizen Federal Action Tracker

A weekly monitor of how Trump’s orders and actions are impacting planners and planning in America.

May 14, 2025 - Diana Ionescu

Front of Walmart store with sign.

Walmart Announces Nationwide EV Charging Network

The company plans to install electric car chargers at most of its stores by 2030.

May 7, 2025 - Inc.

Aerial view of Albuquerque, New Mexico at sunset.

New State Study Suggests Homelessness Far Undercounted in New Mexico

An analysis of hospital visit records provided a more accurate count than the annual point-in-time count used by most agencies.

May 16 - Source NM

Close-up on white bike helmet lying on pavement with blurred red bike on its side in background abd black car visible behind it.

Michigan Bills Would Stiffen Penalties for Deadly Crashes

Proposed state legislation would close a ‘legal gap’ that lets drivers who kill get away with few repercussions.

May 16 - Wood TV 8

Muni bus on red painted bus-only lane in downtown San Francisco, California.

Report: Bus Ridership Back to 86 Percent of Pre-Covid Levels

Transit ridership around the country was up by 85 percent in all modes in 2024.

May 16 - Mass Transit