Ashesh Badani, Red Hat | Red Hat Summit 2026
summary
Red Hat says the real AI bottleneck is the platform not the model and gives concrete fixes: Hummingbird runs autonomous pipelines that build thousands of container images weekly with humans in roughly 10% of merge requests, and token budgets assign inference spend like cloud spend. OpenShift becomes the convergence point for VM migration and AI workloads, a proposed agent ops framework enforces identity, permissions and auditability, and pre-configured landing zones embed DORA/GDPR compliance and smart API gateways route queries to smaller distilled models.
AI-written summary · about
chapters (11)
- 0:00Welcome and introduction
- 1:13AI model shift to inferencing
- 2:39GPU efficiency and convergence
- 4:26Token economics and experimentation
- 6:31Secure AI at scale
- 8:39Networking and agent identity
- 10:04Sovereignty and hybrid cloud
- 11:50Model as a service strategies
- 13:36Customer success factors
- 15:44Budgeting for AI workloads
- 17:31Future vision and predictions
