# AWS Batch adds GPU quota sharing for SageMaker training jobs

_Published Thursday, October 29, 2026 at 1:05 PM EDT · AI, Infrastructure · Latest · Tier 2 — Notable_

![AWS Batch adds GPU quota sharing for SageMaker training jobs — Primary](https://d2908q01vomqb2.cloudfront.net/f1f836cb4ea6efb2a0b1b99f41ad8b103eff4b59/2026/10/27/ML-21246-featured-image.png)

AWS has introduced quota management in AWS Batch for SageMaker Training jobs, allowing administrators to allocate compute capacity by team or project and automatically share idle resources. Teams can reserve their allocation, lend unused capacity, or both lend and borrow.

The scheduler reclaims lent resources when the owning team submits work. Administrators can also enable interruption of lower-priority jobs within a team's allocation so higher-priority jobs start sooner. The SageMaker Python SDK supports submitting training jobs to queues configured for quota management.

Borrowed capacity can be reclaimed, interrupting jobs; teams seeking fewer interruptions can reserve capacity and disable priority-based preemption.

## Sources

- [Artificial Intelligence](https://aws.amazon.com/blogs/machine-learning/lend-and-borrow-resources-across-teams-with-aws-batchs-quota-management-for-sagemaker-training-jobs/)

---
Canonical: https://techandbusiness.org/newswire/vSb55Sk2OsotA2p5F1Jq2T
Published: 2026-10-29T17:05:56.283Z
Story chronology: 2026-10-29T16:58:57.000Z
Retrieved: 2026-10-29T20:40:06.234Z
Publisher: Tech & Business (techandbusiness.org)