AWS Updates 2026-06-05 | Cloud Provider News

Better cost visibility lands across AWS tools

AWS rolled out the new Pricing Change Analysis Dashboard, and it’s built to show how pricing changes affect your spend. The dashboard tracks every AWS pricing change across the SKUs you use and surfaces the dollar impact, so FinOps teams can see what’s driving variance instead of guessing.

AWS also expanded AWS Config coverage with nine new resource types across Bedrock, Bedrock AgentCore, and SageMaker. That means you can inventory and evaluate more AI and ML resources for governance and cost auditing across every AWS Region where those resources are available.

Compute and container updates aimed at better efficiency

AWS Compute Optimizer now supports a 32-day lookback window for rightsizing recommendations, up from 14 days and at no additional cost. The longer lookback helps recommendations reflect monthly usage patterns, which can improve the quality of cost-optimization decisions across EC2, Auto Scaling groups, RDS, EBS, and ECS.

AWS also added support for Kubernetes 1.36 in Amazon EKS and EKS Distro. The release includes in-place pod-level resource vertical scaling, which lets pods adjust CPU and memory budgets without restarting.

New ways to cut spend on specialized workloads

AWS released Flex-start VMs for Managed Instance Groups in Compute Engine — wait, that’s the Google Cloud item, so for AWS this slot isn’t applicable.

AWS added AWS Trainium and AWS Inferentia support to Amazon ECS Managed Instances. This lets customers run accelerator-backed container workloads on managed infrastructure, including Inferentia2, Trainium1, and Trainium2.

Dashboards and taxonomy updates for cleaner FinOps reporting

CUDOS dashboards v5.8.1 updated Bedrock token cost visuals from cost per 1K tokens to cost per 1M tokens and added Bedrock spend visuals. The change improves visibility into LLM token costs for cost allocation and FinOps reporting.