"Full-stack AI" sounds appealing, but the IT reality is more complex

At SUSECON 2026, ITQ’s Francisco Perez van der Oord said “full-stack AI” usually means a validated mix of proprietary and custom components, not one complete vendor stack.

He argued enterprises are already building AI-driven operations across mixed VM and Kubernetes environments instead of waiting for a fully turnkey AI platform.

The trend suggests AI infrastructure is becoming an enterprise planning and operations problem, not just a software feature.

This could reshape how vendors like VMware, Nutanix, Red Hat, SUSE, and hyperscalers fit into production environments.