Virtual
In-person
Enterprise AI infrastructure in 2026: A reality check for platform engineers
After two years of hype, many enterprises feel they’ve overspent on AI infrastructure without realizing the promised value. In this session, Dan Ciruli, Kelsey Hightower, and Luca Galante cut through the buzzwords to show what enterprise AI infrastructure actually looks like in production.
In most cases, it is often less about training models from scratch and more about how platform engineering teams can fold AI inference and fine-tuning directly into their existing infrastructure.
Join the discussion to learn:
- Why you don’t need infrastructure for training models from scratch, and how to start treating inference and fine-tuning just like serving standard web APIs.
- Strategies for folding AI workloads and autonomous agents into your existing heterogeneous environments (VMs + containers + legacy databases) without rebuilding everything.
- How to build an Agentic Development Platform (ADP) based on actionable reference architectures to deploy real-world AI applications.
