aws AWS What's New ·

Amazon SageMaker Adds Instance Preference Lists for Training and Processing Jobs

aiawsgaengineeraws-sagemaker
feature

Amazon SageMaker now supports instance preference lists for AI training and processing jobs, allowing users to specify a prioritized list of acceptable instance types. This enhancement helps users find compute capacity more quickly, especially for high-demand GPUs during peak periods. The feature eliminates the need for complex retry logic or submitting multiple jobs concurrently to secure an available instance. It is generally available today across all AWS Regions where SageMaker operates.

Features (1)
  • Specify Prioritized Instance Preferences for SageMaker Jobs

    SageMaker users can now provide a prioritized list of desired instance types and counts when submitting training or processing jobs. This allows SageMaker to automatically launch the job on the first available configuration from the user's preferences, reducing wait times and the need for manual retries or custom logic.

Read the original announcement →

https://aws.amazon.com/about-aws/whats-new/2026/09/amazon-sagemaker-training-processing-instance-pref-lists/

Related releases