SageMaker Training Plans introduced the ability to extend reserved GPU capacity in increments from one day up to 182 days, allowing users to continue running AI training jobs without reconfiguring workloads when they exceed initial time estimates.
These extensions can be managed via API or console, facilitating cost-efficient planning and infrastructure provisioning. SageMaker automates resource allocation based on plan parameters, supporting scalable, continuous AI model training across AWS regions with minimal user intervention.
Source & further reading
SageMaker Training Plans now enables extending of existing capacity commitments without workload reconfiguration ↗Source published 2026-03-17. 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.