​The Agentic AI Infrastructure Playbook

​Agentic AI is putting pressure on infrastructure in ways most teams did not anticipate — and the early adopters are all reporting the same emerging patterns:

• ​Reasoning chains triggering 5–20× workload volatility

• ​Orchestration frameworks failing under chained tool executions

• ​Unpredictable token cascades driving sudden cost overruns

• ​Governance gaps emerging once systems begin acting autonomously

​This session brings together leaders solving these problems right now — including members of Google’s internal agent team — to share what has actually worked inside real production environments.

​What You’ll Learn

• ​Architectural patterns that stabilize agentic workloads when reasoning never stops

• ​Where observability must evolve to capture reasoning steps, not just prompts

• ​How leading teams are containing cost spikes created by multi-agent loops

• ​Governance & safety controls required before agents scale across teams

​If you’re responsible for planning or defending your 2026 AI infrastructure roadmap, this is the most concentrated, real-world intelligence you will get in one room this quarter.

​Featured Speakers:

​Event Agenda:

​5:15 – 6:00 PM: Registration & Networking

​6:05 – 6:25 PM: Beyond Agentic Slop: Where Enterprise Agents Actually Work — and Why Yours Don’t

​Speaker: Amjad Masad, CEO & Founder, Replit

​6:25 – 6:45 PM: Session with Google Cloud

​Speaker: Mike Clark, Director of Product Management, Google Cloud

​6:45 – 7:00 PM: Breaking the Memory Wall: The Infrastructure Required for Statefulness

​Speaker: Simon Ben-David, Chief Technology Officer, Weka

​7:10 – 8:00 PM: Closing Networking

​Who’s in the Room

​Invite-only audience of senior AI, data, and engineering leaders:
C-suite executives, VP/Director-level AI strategists, and hands-on practitioners from companies ranging from 200–5,000 employees.

​Why Attend

​Peer-Level Depth: Off-the-record conversations with leaders rebuilding AI infrastructure in real time.

​Actionable Frameworks: Concrete architectures for orchestration, inference, cost containment, and governance.

​Future-Facing Insight: A clear view into how agentic workloads will reshape the stack — from memory and I/O to economics and operational strategy.