Exploring Princeton’s Cutting-Edge AI Chip Technology

AI Workload Chip Prototype

Princeton researchers have completely rethought the physics of computing to construct a chip for modern AI workloads, and with new U.S. government support they will observe how fast, compact, and power-efficient this chip can become. An early prototype is shown above. Credit: Hongyang Jia/Princeton University

Princeton’s cutting-edge AI chip project, supported by Naveen Verma

Professor Naveen Verma will lead a U.S.-backed project to supercharge AI hardware based on a suite of key innovations from his Princeton laboratory. Credit: Sameer A. Khan/Fotobuddy

The disclosure unfolded as part of a broader campaign by DARPA to fund “innovative advances in science, devices, and systems” for the next era of AI computing. The initiative, dubbed OPTIMA, incorporates projects across various universities and companies. The solicit for proposals for the initiative approximated total funding at $78 million, although DARPA has not disclosed the complete list of institutions or the overall funding quantity the initiative has granted to date.

The Emergence of EnCharge AI

In the Princeton-led undertaking, researchers will cooperate with Verma’s startup, EnCharge AI. Based in Santa Clara, Calif., EnCharge AI is commercializing technologies based on discoveries from Verma’s lab, including several crucial papers he collaborated on with electrical engineering graduate students dating back to 2016.

Encharge AI “brings leadership in the development and execution of reliable and scalable mixed-signal computing architectures,” as per the project proposition. Verma co-established the firm in 2022 with Kailash Gopalakrishnan, a former IBM Fellow, and Echere Iroaga, a leader in semiconductor systems design.

Gopalakrishnan mentioned that advancement within existing computing architectures, along with enhancements in silicon technology, began decelerating at precisely the moment when AI triggered substantial new requirements for computational power and efficiency. Not even the finest graphics processing unit (GPU), utilized to operate present-day AI systems, can remedy the obstacles in memory and computational energy facing the industry.

“While GPUs are the best accessible tool today,” he stated, “we concluded that a fresh kind of chip will be necessary to unlock the potential of AI.”

Transforming AI Computing Landscape

Between 2012 and 2022, the computation power required by AI models multiplied by about 1 million percent, according to Verma, who is also director of the Keller Center for Innovation in Engineering Education at

Leave a Reply

Your email address will not be published. Required fields are marked *