Is your VMware migration stalling? Your bare metal layer might be why.

In this episode of the VMblog Expert Interview Series, we sit down with Rob Hirschfeld, Founder and CEO of RackN — the company behind the Digital Rebar platform — for a candid, deeply technical conversation about what’s really breaking down in enterprise infrastructure today.

Rob and I are genuine VMware OGs, having worked with ESX since Alpha code around 2000, which gives this conversation a rare depth of perspective. From the Broadcom licensing shockwave to the rise of AI infrastructure demands and the accelerating trend of cloud repatriation, we cover the full landscape of where enterprise IT is struggling — and where it has the most opportunity.

The Broadcom Wake-Up Call Nobody Is Solving

Rob makes a striking observation: he doesn’t know a single customer who has committed to staying with Broadcom — but he also doesn’t know a single customer who has successfully migrated off of it. The reason? Most teams focus on VM migration tools while ignoring the deeper operational stack — storage, networking, orchestration, onboarding, and lifecycle management — that VMware quietly handled for years.

Why DIY Tools Like MaaS and Terraform Fall Short at Scale

Tools like Canonical’s MaaS and Terraform represent only about 10% of the bare metal management spectrum. Without lifecycle control, inventory discovery, BIOS configuration, patching, and decommissioning built into a single cohesive process, teams are left managing fragile, homegrown automation that typically runs at only around 80% reliability. RackN’s Digital Rebar, by contrast, delivers over 99.9% success rates on reprovisioning cycles.

AI Infrastructure Is Just Bare Metal — But Harder and More Urgent

Rob pushes back on the idea that AI requires a completely different infrastructure philosophy. GPUs, SmartNICs, DPUs, InfiniBand, ARM-based servers from NVIDIA — it’s all still bare metal at the core. What’s changed is the cost, the complexity, and the executive pressure to keep quarter-million-dollar servers running like a pit crew keeps a race car on the track.

Day Zero to Day Omega: What Full Lifecycle Management Actually Looks Like

Rob walks through RackN’s complete lifecycle framework — from pre-boot manifest validation (Day Zero) through provisioning (Day One), ongoing patching and reprovisioning (Day Two), all the way to secure decommissioning (Day Omega). For AI workloads running immutable, cloud-native OS patterns, that Day Two cycle of tearing down, patching, and rejoining clusters isn’t optional — it’s the entire operational model.

Fix the Foundation First — Then Pick Your Platform

The most counterintuitive advice in the conversation: stop making the virtualization platform decision before you’ve stabilized your bare metal layer. Rob shares a story about a customer fighting an OpenShift virtualization migration for four straight quarters — until RackN was brought in and had OpenShift running on bare metal in under an hour. Getting the foundation right makes every platform decision above it easier, faster, and less risky.

Cloud Repatriation Is Not a Retreat — It’s a Step Forward

As organizations bring workloads back on-premises — driven heavily by AI cost and continuity concerns — Rob cautions against rebuilding the fragile, slow on-prem environments that drove people to the cloud in the first place. The goal is a modern, cloud-native data center with hyperscale-grade processes baked in, not a return to legacy IT habits.