Episode 101: Hybrid AI Infrastructure and the Rise of Non-Human Identity

Agents don't log in. They connect. And that changes everything about access control. In this episode of Built for Trust, Nick Lippis sits down with Tom Gillis of Cisco and Robert Kim of Presidio to discuss hybrid AI infrastructure and the operational challenges no one has fully solved yet. They explore why on-prem AI is becoming practical for enterprises that can't easily retrofit their data centers for liquid cooling, how Presidio built an agent framework that lets users switch between cloud and private models, and why Ethernet is emerging as the preferred fabric for enterprise AI infrastructure. Tom draws a direct line from infrastructure to security: agents have the access footprint of a human but none of the judgment, and today's access control frameworks weren't built for that. The network, he argues, is where non-human identity has to live and the only enforcement point organizations can truly rely on when agents begin operating at scale. Want to go deeper? Watch the related on-demand AI Networking Summit session, Hybrid AI Infrastructure and Agentic Ops Security in Practice: https://hubs.li/Q04jpjxF0 New episodes of the Built for Trust Podcast come out every Thursday. Subscribe to get notified.