AI news ·
UC Merced researchers use AI to find energy-efficient chip materials for computing
UC Merced received $810,000 of a $6 million grant to study materials that could cut AI's power use. A single ChatGPT query consumes roughly 10 times more electricity than a Google search.
UC Merced joins $6 million effort to make AI computing more efficient
Researchers at UC Merced are developing materials that could cut the power consumption of artificial intelligence systems. The university received $810,000 of a $6 million systemwide grant to investigate topological materials - substances where electrons move in ways that enable faster, lower-energy switching in computer chips.
The problem is concrete. A single ChatGPT query uses about 0.34 watt-hours of electricity, roughly 10 times more than a Google search. U.S. data centers consumed 183 terawatt-hours in 2024, accounting for more than 4% of the nation's total electricity consumption - equivalent to Pakistan's annual electricity demand.
The project involves principal investigators from five UC campuses (Santa Barbara, Merced, San Diego, Irvine, and Berkeley) alongside scientists from Lawrence Livermore and Los Alamos national laboratories.
Two research tracks
Elizabeth Nowadnick, a chemical and materials engineering professor at UC Merced, is leading the campus team. The group is pursuing two goals: using AI to accelerate discovery of topological materials and related quantum materials, and developing new electronic switching methods that reduce computational costs.
"We are investigating topological materials whose electronic structures can be rapidly switched with minor disturbances, meaning that the cost of each switching operation is minimized," Nowadnick said. Because these changes are electronic rather than structural, they can happen extremely fast.
By using AI to identify the most promising candidates before laboratory testing, the team could dramatically speed up the discovery of chip materials that switch faster while consuming less energy.
Automating materials simulation
Nowadnick's team includes one postdoctoral researcher and two physics Ph.D. students. They are using density functional theory (DFT) - a computational tool that simulates material properties - to model how topological materials might behave in computer systems.
The researchers are also building an AI system with computer scientists at UC San Diego called TritonDFT. The system automates DFT calculations and allows users to interact with the code using natural language, similar to ChatGPT. This autonomous approach can orchestrate the entire DFT workflow across physics, computing, and high-performance computing expertise.
TritonDFT has the potential to speed up calculations and make them accessible to researchers without specialized DFT training.
Learn more about AI Research Courses and Generative AI and LLM Courses to understand how these technologies are applied in scientific discovery.