Agent Infrastructure Roundtable 2026
About the Event
Five thousand people will be at the Berkeley Agentic AI Summit 2026 talking about what agents can do. This is the room next door for the people who have to make those agents run at scale.
The next wave of agentic AI needs product builders and infrastructure builders in the same room. This evening brings together leaders across silicon, systems, cloud, frameworks, and enterprise agent teams to discuss the stack agents will depend on.
The conversation will explore what changes when agents move from prototypes to persistent, production-scale systems:
How should infrastructure evolve for long-running, stateful agents?
Where should security, governance, memory, and orchestration live?
How do we align silicon, systems, platforms, and enterprise demand?
What partnerships are needed to move agent infrastructure into production?
This is not a product promotion event, and nothing is being sold. ARAMAS and BGA are hosting because the conversation needs a host, not a sponsor. We sit at the intersection of these builds and are bringing that center of gravity together.
Expect a candid panel discussion followed by an informal mixer with dinner and drinks.
Panelists
Dr. Wei Li
Adjunct Professor at Stanford University and former Vice President and General Manager of AI & Analytics at Intel. Dr. Li brings deep experience in AI strategy, governance, full-stack platforms, and the commercial scaling of AI systems.
Anurag Juneja
Senior Director of Agentic AI & Data Governance at Salesforce. Anurag leads governance for Data 360 and Agentforce, shaping how autonomous agents operate across compliant enterprise data architectures.
Ramakrishnan Sivakumar
Principal Machine Learning Software Engineer at AMD and creator of the open-source Lemonade SDK. His work focuses on ML frameworks, local AI, and optimized deployment across GPUs and NPUs.
Micah Villmow
Senior Research Engineer at the MBZUAI Institute for Foundation Models Silicon Valley Lab. Micah works on frontier models, AI coding agents, and multi-agent developer workflows, building on nearly two decades of experience in GPU computing and AI inference.
Panel Moderator
Freddy Chiu
AI Software Architect at Intel and a technical leader with nearly two decades of experience in AI infrastructure and software architecture. Freddy will lead the discussion, drawing on his work across hardware, software, and cross-industry partner engineering.
Who Should Attend
Builders and leaders working across:
AI infrastructure and accelerated computing
Enterprise agents and agent platforms
Cloud, data center, and systems architecture
AI governance, identity, and security
Developer frameworks and agent orchestration
Silicon, hardware, and infrastructure partnerships
No pitches. No media. Just a focused conversation among the people building the infrastructure behind the agent economy.