Oracle has begun offering access to Nvidia’s GB200 NVL72 systems via its cloud platform, integrating thousands of Blackwell GPUs into its infrastructure. With plans to scale beyond 100,000 GPUs, Oracle’s latest move pushes toward a 2.4 zettaflop AI supercluster. Available across public, sovereign, and on-premise clouds, the deployment marks a major step in reshaping large-scale AI training capabilities.
Kubernetes 1.33, code-named “Octarine,” has landed with 64 feature enhancements, marking a major leap in container orchestration. Highlights include stable support for sidecar containers, native user namespaces for tighter security, nftables integration for faster networking, and expanded dynamic resource allocation for AI workloads. The release signals Kubernetes’ growing focus on resilience, scalability, and high-performance computing in cloud-native environments.
