Featured Talk & Panel Discussion

上午 10:50 - 下午 12:15

Dr. Da Yang, Senior Director of Product, Semiconductor and EDA, NVIDIA

Topic: AI Supercomputing Meets EDA: Accelerating the Future of Chip and System Design

Abstract:

As semiconductor designs advance toward trillion-transistor packages and quadrillion-transistor systems, engineering workflows cannot scale by simply adding more people or running existing tools faster. This session explores how agentic AI, AI physics and accelerated computing are reshaping chip and system design. AI physics enables near-real-time simulation and broader design-space exploration, while long-running engineering agents connect CAD, EDA, verification, and optimization tools through governed workflows that can plan, execute, evaluate, and improve. Da Yang will show how NVIDIA’s full-stack approach and collaboration with the EDA and semiconductor ecosystem can shorten design cycles, improve engineering productivity, and help bring increasingly complex chips and systems to market.

Mr. Aveek Sarkar, Head of Ecosystem and Alliance Management Division, TSMC

Topic: The Road to Trillion-Transistor Multi-Die Systems and Beyond

Abstract:

The proliferation of AI is placing unprecedented demands on silicon infrastructure as intelligence spreads beyond the data center to the edge. This presentation will explore the critical requirements for compute, memory, connectivity, and power, and how advanced-node silicon, multi-die advanced packaging, and associated technologies enable the performance and energy efficiency metrics needed for the next generation of AI. We conclude with a vision for how the collaboration across the stack—from the system companies and ODMs to the chip design companies, and TSMC along with its Open Innovation Platform® (OIP) partners, can address the requirements for co-design and multi-physics convergence, paving the road to trillion-transistor multi-die systems and beyond.

Key Technologies Covered

Advanced-Node Silicon
Multi-Die Advanced Packaging
DTCO / STCO

Mr. James Jau, Vice President, Quanta Computer Inc.

Topic: AI Supercomputer Scale-Up: Evolution for Today and Beyond

Abstract:

AI supercomputing is rapidly evolving as advances in accelerator performance reshape requirements across system architecture, interconnect bandwidth, power delivery, and cooling. As AI infrastructure moves toward larger and more tightly integrated scale-up systems, holistic system-level co-design is becoming essential to meet growing demands for compute density and performance.

This presentation explores the evolution of AI supercomputer scale-up architectures and the key technologies driving this transformation—from advanced scale-up fabrics and optical interconnects to next-generation power delivery and cooling solutions. Looking beyond today’s rack-scale systems, it will also examine emerging System-on-Wafer (SoW) technologies and their potential to shape the next frontier of large-scale AI computing.

Key Technologies Covered

AI Scale-Up Architectures
High-Speed Scale-Up Fabrics
Electrical and Optical Interconnects
End-to-End Power Delivery
Advanced Cooling Solutions
System-on-Wafer (SoW)

Featured Speakers

李祥宇

S&A Country Manager, Taiwan, Synopsys

Roger Lee is the S&A Country Manager for Taiwan at Synopsys, leading the company’s business strategy and engagement with Taiwan’s semiconductor ecosystem. He works closely with customers and industry partners to accelerate innovation across the semiconductor value chain through AI-powered simulation, Multiphysics simulation, and digital engineering. His focus is on enabling silicon-to-system co-design for next-generation IC design, advanced packaging, AI infrastructure, and high-performance computing (HPC).

With more than 30 years of leadership experience in the technology industry, Roger has held executive roles spanning enterprise software, consulting services, computing platforms, semiconductor technologies, and manufacturing solutions across Greater China and Southeast Asia. Before joining Ansys in 2018, he served as Greater China General Manager at JDA Software, VP at Hewlett-Packard (HP) Taiwan, and Sales Director at Oracle Taiwan, where he built extensive expertise in business leadership, strategic planning, and market development. Mr. Lee holds a Bachelor’s degree in Industrial Engineering from Chung Yuan Christian University.

Dr. Da Yang

Senior Director of Product, Semiconductor and EDA, NVIDIA

Da Yang is Senior Director of Product at NVIDIA, where he leads the AI product development for Semiconductor and EDA applications. Before joining NVIDIA, Da was Vice President of Technology at Tokyo Electron America, leading new product development for fabless design customers. Prior to Tokyo Electron, Da had worked at Qualcomm and IBM covering various design optimization and process development roles.

Mr. Aveek Sarkar

Head of Ecosystem and Alliance Management Division, 台積公司

Aveek Sarkar is Head of Ecosystem and Alliance Management Division at TSMC. Along with his team, he is responsible for enabling TSMC’s Open Innovation Platform and its partner ecosystem covering EDA, IP, Cloud, design services and 3DFabric. Prior to TSMC, Aveek held leadership roles in Synopsys, Ansys, Apache Design Solutions and Sun Microsystems. He has a MBA (Santa Clara University), MSEE (Oregon State University) and B.Tech. (Indian Institute of Technology, Kanpur, India).

Mr. James Jau

Vice President, Quanta Computer Inc.

James Jau is Vice President of Quanta’s Feynman Research Center (FRC), leading R&D in advanced AI supercomputing and quantum computing, with a focus on system architecture, platform development, and ecosystem collaboration. He joined Quanta in 1999 and established its cloud engineering team, overseeing servers, storage, networking, and rack systems.

With over 30 years of server design experience, James has led major industry innovations, including the development of Taiwan’s first locally designed blade server with integrated networking in 2003 and NVIDIA DGX-1, the world’s first NVLink-enabled deep learning system, in 2016.

As a former member of the Open Compute Project (OCP) Incubation Committee, he contributed to projects including Facebook’s Tioga Pass, Big Sur, and Big Basin, as well as Microsoft’s Open Cloud Server infrastructure. He also continues to guide Quanta’s quantum computing development and initiatives.

James previously held leadership roles at Wang Laboratories and Digital Equipment Corporation. He holds a bachelor’s degree in Nuclear Engineering and a master’s degree in Electrical and Computer Engineering and has been granted 36 cloud computing-related patents.