NVIDIA AI Updates: April 10, 2026

1. Slinky Enables Large-Scale Slurm Workloads on Kubernetes at 8,000+ GPU Scale

NVIDIA. NVIDIA published a technical deep dive on Slinky, the open-source project (from SchedMD, now part of NVIDIA) that natively integrates Slurm cluster management with Kubernetes. The slurm-operator represents all Slurm daemons as Kubernetes CRDs, supporting full lifecycle orchestration, high availability, and autoscaling. Production deployments at NVIDIA have scaled to over 8,000 GPUs with no measurable performance impact from the Kubernetes layer. The v1.1.0 release adds dynamic topology support and DaemonSet-style worker pod scaling. Source