AWS Releases
Amazon Web Services releases and Terraform AWS provider. New features, breaking changes, security advisories and deprecations - each summarised in plain English and updated continuously.
Tracking 680 AWS releases · Updated
- AWS What's New aiinfraawsengineeraws-s3aws-sagemakeraws-bedrock ·
SageMaker Studio Workflows adds operators for Bedrock, S3, Glue, MWAA
Amazon SageMaker Unified Studio Workflows now integrates with Amazon Bedrock, S3 Tables, S3 Vectors, AWS Glue Data Catalog, and Amazon MWAA Serverless via 19 new operators. This allows users to orchestrate these AWS services visually within Studio Workflows without writing custom code, simplifying complex data and ML pipelines. This enhancement is available across all SageMaker Studio regions and benefits data workers and builders by reducing context switching.
feature - AWS What's New aisecurityawsaws-s3 ·
Amazon S3 Vectors expands to AWS GovCloud (US) Regions
Amazon S3 Vectors, a vector storage service for AI applications, is now available in AWS GovCloud (US-East) and AWS GovCloud (US-West). This expansion allows government agencies and their partners to leverage S3's elasticity and durability for AI workloads like RAG and semantic search. The service offers dedicated APIs for vector storage and querying without infrastructure provisioning.
feature announcement - AWS What's New aiawsengineeraws-rdsaws-bedrock ·
Amazon RDS for Oracle adds support for Oracle Database 26ai
Amazon RDS for Oracle now supports Oracle Database 26ai, integrating with Amazon Bedrock for foundation models and offering Oracle's Select AI for natural language queries. This enables in-database retrieval augmented generation (RAG) and AI Vector Search without external vector databases. The feature is available in all commercial and AWS GovCloud Regions for Enterprise Edition instances, with upgrade paths from 19c and 21c.
feature - AWS What's New aiinfraawsengineeraws-ec2aws-eksaws-ecs ·
Amazon ECS Managed Instances reduce GPU management fees
Amazon ECS Managed Instances is reducing management fees for GPU instances, with G-series fees dropping by 35% and P-series/Trainium fees by 60%, starting July 1, 2026. This aims to lower costs for customers running accelerated workloads on ECS. The reductions apply automatically to existing users and are available in all ECS Managed Instances regions.
announcement patch - AWS What's New aiinfraawsengineeraws-eksaws-ecs ·
Amazon EKS Auto Mode reduces GPU management fees
Amazon EKS Auto Mode is reducing management fees for GPU and accelerated instance types, with reductions of 35% for G-series and 60% for P-series and AWS Trainium starting July 1, 2026. These savings aim to make EKS Auto Mode, which simplifies GPU infrastructure management for ML and other accelerated workloads, more cost-effective for customers. The fee reductions apply automatically to all existing and new EKS Auto Mode clusters, with identical reductions also available for Amazon ECS.
patch - AWS What's New aiawsgaengineeraws-sagemakeraws-bedrockaws-iam ·
SageMaker Studio integrates with Hugging Face for one-click model deployment
Amazon SageMaker Studio now offers direct integration with Hugging Face, enabling users to deploy or customize models with a single click. This streamlines the process from model discovery to a fully configured Studio environment, reducing prior manual steps like console navigation, permission setup, and quota requests. New customers get a Studio environment with pre-configured permissions, while verified customers receive default GPU access without manual quota increases. The feature is available in all supported AWS Commercial Regions.
feature - AWS What's New aiawsengineeraws-eksaws-sagemaker ·
SageMaker HyperPod enhances LLM inference with Disaggregated Prefill and Decode
Amazon SageMaker HyperPod now supports Disaggregated Prefill and Decode (DPD), an optimization that separates LLM inference prefill and decode phases onto dedicated GPU pools to improve per-token latency and throughput for production LLM workloads. This allows for independent scaling of compute and memory bandwidth resources, benefiting chat assistants, agentic pipelines, and RAG applications. DPD is enabled via the InferenceEndpointConfig custom resource and is available for SageMaker HyperPod clusters using EKS on EFA-capable instance types in all available AWS regions.
feature - AWS What's New aiawsgaengineeraws-eksaws-sagemaker ·
SageMaker HyperPod adds AMI versioning and auto-patching
Amazon SageMaker HyperPod now provides visibility into AMI versions and automatically applies security patches to clusters without disrupting workloads. This enhances security and consistency by enabling administrators to detect drift, roll back to previous versions, and apply patches non-disruptively. These features are available for HyperPod clusters orchestrated by Amazon EKS in all supported AWS Regions.
feature patch - AWS What's New aiinfraawsengineeraws-sagemakeraws-iam ·
SageMaker Unified Studio now supports Terraform provisioning
Amazon SageMaker Unified Studio now supports Terraform for provisioning, allowing platform teams to integrate SageMaker domains into existing infrastructure-as-code pipelines. This enhancement ensures consistency across environments and simplifies the deployment of managed workspaces with built-in governance and access control. The feature is available in all AWS Regions where SageMaker Unified Studio is offered and requires the Terraform AWS Cloud Control Provider.
feature - AWS What's New aiawsengineeraws-bedrock ·
Amazon Bedrock AgentCore increases default runtime quotas
Amazon Bedrock AgentCore has increased default runtime quotas, providing customers with greater capacity for scaling AI agent workloads. New limits support up to 5,000 concurrent sessions in key US regions and 200 agent interactions per second globally. This enhancement allows for more simultaneous AI agents and higher throughput capabilities without immediate configuration changes, benefiting developers building AI-powered applications.
feature - AWS What's New aiinfraawsengineeraws-eksaws-ecs ·
AWS Network Firewall Adds Container Attribute Inspection for EKS/ECS
AWS Network Firewall now supports container attribute-based rules for securing containerized workloads on EKS and ECS. This simplifies network security by allowing rules based on native constructs like namespaces and cluster names, rather than complex IP addresses. This enhancement is particularly beneficial for dynamic environments running generative AI applications, offering adaptive security and compliance.
feature - AWS What's New aiawsengineeraws-ec2aws-lambdaaws-eksaws-ecs ·
AWS AppConfig launches managed A/B testing and experimentation tools
AWS AppConfig now offers managed tools for running A/B tests and feature experiments, removing the need for separate infrastructure. These tools incorporate Amazon's best practices and AI-driven guidance to help users build robust experiments with controlled exposure and traffic allocation. Available for applications running on various AWS services and on-premises, this feature enables data-driven decision-making for feature rollouts. Users set up experiments in AppConfig and analyze results via CloudWatch or other tools, promoting winning variations through safe rollouts.
feature announcement - AWS What's New aiawsengineeraws-bedrock ·
Amazon Bedrock AgentCore expands to four new AWS Regions
Amazon Bedrock AgentCore, a platform for building and optimizing agents, is now available in four additional AWS Regions: Asia Pacific (Bangkok), Asia Pacific (Malaysia), Europe (Milan), and Europe (Spain). This expansion allows customers in these regions to deploy agents with lower latency and closer proximity to their end users. All core AgentCore capabilities, including runtime, security, and tool connectivity, are available at launch in these new regions.
feature announcement - AWS What's New aiawsengineeraws-bedrock ·
Claude Sonnet 5 now available on AWS via Bedrock and Claude Platform
Anthropic's Claude Sonnet 5, their most capable Sonnet model yet, is now accessible on AWS through Amazon Bedrock and Claude Platform. This model offers enhanced performance for coding, agentic tasks, and professional knowledge work at Sonnet pricing. Customers can choose Bedrock for AWS-managed features and data residency or Claude Platform for direct access to Anthropic's native experience with AWS billing and authentication.
feature announcement - AWS What's New aisecurityawsengineeraws-sagemakeraws-bedrock ·
AWS Security Hub CSPM adds AI Security Best Practices standard
AWS Security Hub CSPM has launched a new AI Security Best Practices standard, featuring 31 automated controls to assess AI resource security. This standard helps continuously evaluate Amazon Bedrock and SageMaker workloads against security best practices, detecting misconfigurations without manual effort. It covers critical domains like network isolation and encryption, with controls spanning various AI infrastructure components and is available in all Security Hub CSPM regions.
feature - AWS What's New aiawsgaengineeraws-bedrock ·
Claude Sonnet 5 available on AWS via Bedrock and Claude Platform
Anthropic's Claude Sonnet 5, the latest generation Sonnet model, is now accessible on AWS. This release offers top-tier intelligence for coding, agents, and professional tasks at Sonnet pricing, balancing capability, cost, and speed. It's available through Amazon Bedrock with AWS managed features and directly via the Claude Platform on AWS, offering unified AWS billing and authentication. Customers can leverage its enhanced performance for complex coding, precise tool calling in agents, and structured analysis of knowledge work.
feature announcement - AWS What's New aiawsgaengineeraws-sagemaker ·
SageMaker Inference speeds generative AI scale-out with container caching
Amazon SageMaker Inference now supports container image caching, reducing generative AI scale-out time by up to half. This feature pre-caches container images for new instances, eliminating the latency associated with pulling large images from ECR during scale-out events. The enhancement is automatically applied to existing SageMaker Inference configurations and is available in all AWS commercial regions.
feature patch - AWS What's New aiawsgapreviewengineerhealthcare ·
Amazon WorkSpaces for AI agents now generally available
Amazon WorkSpaces for agents is now generally available, allowing AI agents to securely operate desktop applications without modernization. This enables automation of critical business workflows like claims processing and trade settlement by providing agents with managed cloud workspaces that offer the same security and governance as human users. The service supports any agent framework using Model Context Protocol (MCP) and offers features like real-time session control and domain-joined fleet support.
feature announcement - AWS What's New aiawsgaengineeraws-sagemaker ·
SageMaker AI supports serverless customization for Gemma 4 models
Amazon SageMaker AI now offers serverless model customization for Google DeepMind's Gemma 4 E4B and 31B models, including supervised fine-tuning, direct preference optimization, and reinforcement fine-tuning. This allows users to tailor foundation models with proprietary data for domain-specific tasks without managing infrastructure. Serverless customization is now available for Gemma and expands options for other families like Nova, Nemotron 3, Qwen, Llama, gpt-oss, and DeepSeek.
feature - AWS What's New aiawsgaengineeraws-bedrock ·
Claude Opus 4.8 available on AWS GovCloud (US)
Anthropic's Claude Opus 4.8 is now generally available on AWS GovCloud (US), offering enhanced capabilities for coding, professional knowledge work, and autonomous tasks. This model provides deeper reasoning, better coding context, and improved agentic task handling for developers and enterprises. Access is provided through Amazon Bedrock, ensuring data stays within AWS infrastructure with managed features.
feature announcement
About AWS release tracking on ReleaseBytes
AWS ships hundreds of service updates a month across EC2, Lambda, RDS, S3, EKS and the rest of the catalogue — far more than anyone can follow by reading the official What's New feed. ReleaseBytes ingests announcements from AWS's official release channels and the Terraform AWS provider changelog, summarises each one in plain English, and tags anything that is a breaking change, security advisory or deprecation so you can see at a glance whether it affects your workloads.
Frequently asked questions
How often are AWS release notes updated on ReleaseBytes? ›
Continuously. ReleaseBytes monitors the official AWS release channels around the clock and publishes a plain-English summary of each announcement shortly after it lands.
What kinds of AWS changes does ReleaseBytes track? ›
New features, enhancements, bug fixes, security advisories, breaking changes, deprecations and end-of-life announcements. Every item is tagged by type so you can filter to just the changes that need action.
How can I get alerts for new AWS releases? ›
Set up a free email or Slack alert filtered to AWS, subscribe to the weekly digest, or follow the RSS feed. Teams can also install the ReleaseBytes GitHub App or connect via MCP.
Where does the AWS release data come from? ›
From the official sources: Amazon Web Services releases and Terraform AWS provider. Every item links back to the original vendor announcement.