Energy-Efficient Computing and Communication: Perspectives on Chiplet Interconnects and Photonic Networks

We are in the midst of a transformative computing revolution, driven by the explosive growth in AI and machine learning (ML) demands. Energy efficiency has emerged as a cornerstone in the design of high-performance supercomputers to meet the ever-growing requirements of daily training and inference workloads.

In this talk, we will delve into the evolution of chiplet interconnects and photonic interconnects, alongside advancements in process and packaging technologies. We will highlight how energy efficiency is fundamentally tied to innovations at every level—device, component, package, and system—enabling the next generation of energy-efficient, high-performance computing systems.

Shenggao Li is a Senior Member of IEEE. He earned his B.S., M.S., and Ph.D. degrees from Northwestern Polytechnical University, Tsinghua University, and The Ohio-State University, respectively.

He is currently with TSMC, California, serving as a Director focused on high-bandwidth chiplet and photonic interconnects. Previously, he was with Intel as a lead designer for PCIe physical layer implementations in multiple generations of Xeon and Core microprocessors, with hundreds of millions of units deployed in global data centers.

Dr. Li co-chairs the UCIe Consortium’s Formfactor and Compliance Working Group, initiated the UCIe 4.0 proposal, and is a TPC Co-Chair for IEEE CICC 2026. He holds 100+ patents and publications.