Led by Verma, a professor of electrical and computer engineering and director of the Keller Center for Innovation and Entrepreneurship, the team reimagined computer chips specifically to run powerful AI systems using much less energy than today’s most advanced semiconductors. This allows systems made with the Princeton technology to run AI programs directly on mobile hardware, rather than communicating over the cloud, enabling applications such as piloting software for drones or AI assistants that keep information securely on local devices. The technology is being commercialized by EnCharge AI, a company based in Santa Clara, Calif. that was co-founded by Verma and two tech-industry veterans.

Jia and Valavi were graduate students studying with Verma when they began working on the chip designs, which enable memory cells to perform complex computations without shuttling data back and forth to a central processor. Jia is now an assistant professor of electronic engineering at Tsinghua University in Beijing. Valavi is a lecturer in electrical and computer engineering at Princeton, with a continued research focus on the intersection of machine learning and efficient hardware design.

Earlier this year, Verma and his team partnered with the U.S. Defense Department’s largest research organization to investigate how the new hardware could be used in a range of mobile and space-based applications. EnCharge AI plans to release its first commercial chips, manufactured by TSMC, next year.

Named after New Jersey’s most famous inventor, Thomas Edison, the Edison Patent Award has been given by the R&D Council for more than 40 years. Winners are selected by a team of R&D Council researchers who evaluate patents based on the significance of the problem addressed, its socioeconomic value, novelty and commercial impact. This year, 14 patents created by 66 inventors and five individual award winners will be honored during the 45th annual Edison Patent Award Ceremony and Reception on November 21, 2024 at Bell Works in Holmdel, NJ.