Red Hat OpenShift 4.20 Boosts AI, Security, Hybrid Cloud (thenewstack.io)

Red Hat announced general availability of OpenShift 4.20 at KubeCon + CloudNativeCon NA 2025, a release aimed at speeding enterprise AI adoption while tightening security and hybrid-cloud/virtualization capabilities. Key AI-focused additions include the LeaderWorkerSet (LWS) API for orchestrating and scaling distributed AI/ML workloads, Runtime Image Volume Source for AI Models (which lets teams add or swap models in minutes without rebuilding containers), and support for the Model Context Protocol (MCP) to connect agents with external data and services in multitenant environments. OpenShift Lightspeed gains multicluster support and surfaces as a generative-AI assistant in the console to help automate tasks and troubleshooting with natural-language guidance. Security, sovereignty and virtualization updates broaden OpenShift’s enterprise appeal: Red Hat Advanced Cluster Security 4.9 is GA, and the Trusted Artifact Signer and Profile Analyzer have been improved for safer supply chains. New controls for digital sovereignty include bring-your-own OpenID Connect integration and initial post‑quantum cryptography support for platform traffic. OpenShift Virtualization adds CPU load‑aware rebalancing, ARM support, and migration toolkits, plus bare‑metal support on Oracle Cloud Infrastructure to simplify moving legacy VMs and storage into containerized, hybrid deployments. Together these changes accelerate model lifecycle management, enable scalable multitenant AI platforms, and address compliance and modernization needs for regulated and hybrid-cloud organizations.