UC San Diego starts Qualcomm-backed research into AI-assisted chip design

UC San Diego and Qualcomm Technologies are putting $750,000 behind a research initiative to use AI in designing chips and computing systems. Announced November 2, 2026, AI in Microelectronic Systems, or AIMS, brings university faculty and Qualcomm researchers together to tackle costly design work, with energy efficiency a particular priority.

Earlier research provides the starting point

The collaboration builds on research relationships already in place. UC San Diego professor Yusu Wang has led work on AI-assisted chip design with Qualcomm researchers. Separately, a university team led by Associate Dean Jinendra Ranka has established independent infrastructure for benchmarking AI processors, comparing their performance and energy efficiency with competing technologies.

That benchmarking effort has already produced targeted feedback for Qualcomm’s team, according to the university. AIMS now brings together researchers from the Halıcıoğlu School of Data Science and Computing and its San Diego Supercomputer Center, drawing on machine learning, computer architecture, high-performance computing and semiconductor design.

The problem they are taking on is a balancing act. Modern chips must meet competing demands for processing speed, energy use, memory capacity and communication between components. They also need to handle different kinds of computing work. UC San Diego says chip design still relies primarily on manual creation, even as its expense grows.

Artificial intelligence is creating extraordinary new demands on computing infrastructure, but it also offers powerful new tools for designing that infrastructure,

Rajesh K. Gupta, dean of UC San Diego’s Halıcıoğlu School of Data Science and Computing

Searching for better designs

AIMS will investigate systems-on-chips: integrated circuits that combine multiple computing components in one device. Its scope also reaches into large-scale computing architectures. The university describes the research as relevant to hardware serving mobile devices, autonomous systems, cloud computing and large-scale AI, rather than a project limited to one type of processor.

The proposed role for AI is to explore possible architectures, predict how they will perform and identify promising design choices. Those choices might otherwise require extensive simulation and testing. The researchers aim to use these methods to reduce design expense while paying particular attention to the energy needed to run the resulting systems.

The research agenda spans several connected layers of hardware and its use:

• Processors and hardware accelerators: the computing components that the team will study as part of its design research.
• Memory systems and interconnects: how systems store information and connect their components.
• Workload characterization and performance optimization: understanding the computing work a system handles and improving how it performs.

A gift without required deliverables

Qualcomm is providing a $500,000 gift to the Halıcıoğlu School, which is adding $250,000 in matching funds. The university says Qualcomm’s contribution carries no requirements for research deliverables, services or reporting. That leaves UC San Diego able to direct the money toward its research and educational mission, rather than fulfilling specified obligations to the donor.

Alongside Wang and Ranka, participating researchers include Lily Weng, Mahidhar Tatineni, Martin Kandes, Zhijian Liu and Arya Mazumdar. Gupta also describes the investment as support for preparing future researchers and engineers, giving the initiative an educational purpose alongside its technical agenda.

For now, AIMS is a commitment to investigate new design methods, not a disclosure of a completed chip or a measured efficiency gain. Its stated next step is research across those hardware layers, pursuing lower design costs and better energy efficiency together. Those are the program’s goals, rather than results delivered by the funding announcement.