Amazon SageMaker HyperPod, infrastructure designed for large-scale foundation model training, added visibility into AMI versions running on clusters using semantic versioning. This permits detection of version drift and enables rollbacks, including associated components like NVIDIA drivers and CUDA.
An opt-in auto-patching feature applies backward-compatible security patches as nodes become idle to avoid interrupting demanding workloads. These enhancements improve cluster security posture and operational consistency at scale. The features are supported on HyperPod clusters orchestrated by Amazon EKS and available in all supported AWS Regions.
Source & further reading
Amazon SageMaker HyperPod now supports AMI versioning and auto-patching ↗Source published 2026-07-02. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.