
Amazon SageMaker HyperPod now supports Karpenter autoscaling, enabling automatic and intelligent GPU cluster scaling for inference and training workloads with reduced operational complexity.


<div>
<p>
Amazon SageMaker HyperPod now supports managed node autoscaling using Karpenter, offering several key benefits for machine learning workloads:
</p>
<ul>
<li>Automatic cluster scaling for inference and training demands</li>
<li>Just-in-time GPU compute provisioning</li>
<li>Scale to zero nodes during low-demand periods</li>
<li>Workload-aware node selection to optimize instance types and costs</li>
<li>Seamless integration with event-driven pod autoscalers</li>
</ul>
<p>
Customers can enable autoscaling on HyperPod using the UpdateCluster API with AutoScaling mode set to "Enable" and AutoScalerType set to "Karpenter". The feature is available in all AWS Regions supporting SageMaker HyperPod EKS clusters.
</p>
</div>


Amazon SageMaker HyperPod now supports autoscaling using Karpenter

Related articles

Related articles

Aug 29
2025
Introducing auto scaling on Amazon SageMaker HyperPod

Aug 11
2025
Amazon SageMaker HyperPod now provides a new cluster setup experience

Jul 2
2026
Amazon SageMaker HyperPod now supports AMI versioning and auto-patching

Apr 23
2026
Amazon SageMaker HyperPod now supports automatic Slurm topology management