Announcing instance preference lists for Amazon SageMaker AI training jobs
TL;DR
Amazon SageMaker AI now offers instance preference lists for training and processing jobs. Specify an ordered list of up to five instance types, and SageMaker AI automatically launches on the first type with available capacity, eliminating manual retry loops and capacity-watching scripts.
Nauti's Take
For teams fighting over scarce GPUs, this is a real advantage: fewer homegrown retry scripts, faster starts and reserved capacity used first. The risk sits in the details, since SageMaker does not check whether your container behaves the same on all five instance types.
Anyone running distributed training should test every option upfront and watch the cost per type.
Sources
15.9.26
