Panel

Title: LLM Token Economy: How is networking going to play in the age of agents?

Abstract:

New business models in LLM workloads are reshaping the interconnect landscape. While ongoing challenges such as low latency, high bandwidth, and reliability persist, the emergence of scale-up networking creates new opportunities for performance with orders of magnitude more bandwidth. On the other end of the spectrum, applications are scaling across multiple data centers, trying to harness a million endpoints. The ability of software to deliver performance across the system holistically has become more challenging than ever, pushing the need for co-design. This HotI panel assembles experts from across the hardware and software stack to foster a discussion of these complex, multi-dimensional problems. The panel will explore the trade-offs between closed, and open AI models. Simultaneously, the panel will feature bleeding edge application developers, who will share their experiences leveraging current hardware, articulate their most pressing challenges, and envision the game-changing features that could revolutionize interconnect performance for their applications.


Moderator:

Peter Olcott
Peter Olcott
Principal at First Spark Ventures
Peter Olcott is a Principal at First Spark Ventures, where he focuses on early-stage investments in AI and networking technologies. With a background in both engineering and business, Peter brings a unique perspective to the intersection of hardware and software innovation.

Panelists:

Phill Brown
Phill Brown
AI Systems Engineer at Meta
Phil leads the design of AI Systems for Meta, integrating high performance accelerators, networking and facilities to create the industry-leading training & inference systems we use to deliver our AI services to our customers. Previously Phil worked at Graphcore (a UK-based startup building AI accelerators) and at Cray developing and delivering large scale high performance computing systems. Phil holds a PhD in computational chemistry from the University of Bristol.
Davor Capalija
Davor Capalija
Senior Fellow, AI Software and Architecture at Tenstorrent Inc.
TBD
Ethan Lockshin
Ethan Lockshin
AE, Commercial at Groq
TBD
Sid Sheth
Sid Sheth
Founder and CEO at d-Matrix
Sid Sheth co-founded d-Matrix in 2019 with a vision to build a new class of compute platform to deploy AI inference efficiently at-scale. In his role as CEO, Sid is leading a global team in developing a first-of-a-kind Generative AI inference solution using in-memory compute chiplet-based architectures, that promises to change the trajectory of commercially viable generative AI. Sid has spent more than two decades as a business and technical leader transforming startups into industry leaders and steering publicly traded corporations toward groundbreaking innovations. Prior to d-Matrix, as Sr. Vice President & GM for the Networking Business Unit at Inphi Corporation, he incubated and grew the group into a $1B+ business by focusing on the cloud and enterprise data center segment. Prior to Inphi, Sid was at NetLogic Microsystems (now Broadcom) and Intel where he ran marketing and worked in R&D for cutting-edge networking chips and processors. Sid earned his Master of Science in Electrical Engineering from Purdue University, where he serves on the Electrical and Computer Engineering Advisory Board. He is a regularly featured speaker at industry tradeshows and conferences, is a published author at ISSCC and holds several patents.