AWS News - 2026-07-24

We are pleased to announce the availability of Amazon EC2 G7e instances in Asia Pacific (Seoul), Europe (London), and Asia Pacific (Tokyo) on Amazon SageMaker AI inference. G7e instances feature up to 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 96 GB of memory per GPU, 5th Generation Intel Xeon processors, and up to 1,600 Gbps of Elastic Fabric Adapter networking bandwidth, delivering up to 2.3x inference performance compared to previous-generation G6e instances.

With this region expansion, you can now deploy inference endpoints on G7e instances closer to your end users in Asia and Europe, reducing latency for generative AI workloads. G7e instances provide up to 768 GB of total GPU memory on a single instance, enabling you to serve medium-to-large language models of up to 70B parameters with FP8 precision without multi-node configurations. These instances are well suited for LLM inference, image and video generation, spatial computing, and scientific computing workloads that require high GPU memory capacity and bandwidth.

G7e instances for SageMaker AI inference are now available in Asia Pacific (Seoul), Europe (London), and Asia Pacific (Tokyo), in addition to previously supported regions. For pricing information on these instances, please visit our pricing page.

AWS GovCloud (US) now offers Claude Sonnet 5 on Amazon Bedrock. Claude Sonnet 5 delivers strong performance across coding, professional work, and agentic tasks while maintaining the balance of capability, cost, and speed. For coding, it navigates large codebases, lands multi-file changes, and carries debugging and refactoring tasks through to completion with fewer rounds of correction. For agents, it calls tools precisely, holds state across many steps, and recovers from errors so more runs finish correctly the first time. For knowledge work, it builds spreadsheets, drafts documents, and turns unstructured material into structured analysis. 

With this launch, Claude Opus 4.8 and Claude Sonnet 5 are available on bedrock-runtime endpoints in AWS GovCloud (US-West and US-East) and bedrock-mantle endpoints in AWS GovCloud (US-West) for performing inference. Bedrock Mantle, Amazon Bedorck's next-generation inference engine, supports the Anthropic Messages API. Amazon Bedrock keeps your data within AWS infrastructure and provides access to Claude Sonnet 5 through a unified service with AWS-managed features like Guardrails, Knowledge Bases, and regional data residency. To learn more, see the Amazon Bedrock documentation and regional availability. 

Today, we're announcing that Amazon Elastic VMware Service (Amazon EVS) is now available in the Asia Pacific (Seoul), Europe (Zurich), and Europe (Stockholm) Regions. This expansion provides more options to leverage the scale and flexibility of AWS for running your VMware workloads in the cloud.

Amazon EVS lets you run VMware Cloud Foundation (VCF) directly within your Amazon Virtual Private Cloud (VPC) on EC2 bare-metal instances, powered by AWS Nitro. You can set up a complete VCF environment in just a few hours, enabling rapid workload migration to AWS to help you eliminate aging infrastructure, reduce operational risks, and meet critical timelines for exiting your data center. This launch supports all existing Amazon EVS features, including VCF 9.0 and 9.1 support to take advantage of the latest VMware features, such as memory tiering.

The added availability in these Regions gives your VMware workloads lower latency through closer proximity to your end users, compliance with data residency or sovereignty requirements, and additional high availability and resiliency options for your enhanced redundancy strategy.

To get started, visit the Amazon EVS product detail page and user guide. 

Amazon Bedrock AgentCore now delivers agent traces and prompts to the same log group as your agent's logs, giving you unified observability for AI agents in a single Amazon CloudWatch log group.

Previously, AgentCore split agent telemetry across multiple destinations trace spans went to the shared `aws/spans` log group while event logs containing prompts, inputs, and outputs went to a separate resource-specific log group. This meant debugging an agent invocation required searching across multiple log groups, and customers could not apply fine-grained access control or customer-managed key (CMK) encryption at the individual agent level. With today's launch, all of an agent's telemetry traces, prompts, structured logs, and standard output is delivered to a single per-agent log group (`/aws/bedrock-agentcore/runtimes/<agent_id>-<endpoint_name>`). You can now correlate traces and logs in one place, scope IAM policies and CMK encryption to individual agents, and export all telemetry by subscribing to a single log group. For multi-agent systems, each agent's complete execution history stays together, making end-to-end debugging straightforward.

All newly created agents starting July 20, 2026 in supported AWS Regions use unified observability by default starting no configuration needed. For existing agents, set the `UNIFIED_TRACES_DESTINATION_ENABLED=true` environment variable on your agent runtime and upgrade ADOT to version 0.17.1 or later. This feature is available in all AWS commercial regions where AgentCore runtime is supported. Learn more in the AgentCore Developer Guide.

Amazon CloudWatch Logs now supports Application Load Balancer (ALB) logs as vended logs, improving observability and simplifying debugging for network traffic patterns. You can now analyze ALB access, connection and health check logs directly in CloudWatch to gain insights into client connections, traffic distribution, connection status and target health, helping you identify and troubleshoot network issues faster. Additionally, you can set up CloudWatch telemetry enablement rules to automatically configure logging of both existing and newly created ALB resources, for your organization, specific accounts, or specific resources, ensuring consistent monitoring coverage without manual setup.

With this CloudWatch Logs integration, you can track detailed access patterns using CloudWatch Logs Insights queries, create metric filters for monitoring and alarming, and review traffic patterns in real time using Live Tail. ALB logs can be configured through the integrations tab of your application load balancer in AWS Management Console, AWS CLI, or SDKs. You can also configure delivery of ALB logs to Amazon Data Firehose or Amazon S3 with support for Apache Parquet format.

ALB logs delivery to CloudWatch is available in all AWS Commercial and GovCloud regions where Application Load Balancer and CloudWatch are available. ALB logs are charged as vended logs when delivered to CloudWatch Logs and Data Firehose, while delivery to Amazon S3 is free (Parquet conversion is charged at $0.035/GB - N. Virginia).

To learn more about configuring ALB logs in CloudWatch Logs, please visit our documentation. For pricing information, see CloudWatch pricing page.

Amazon Web Services (AWS) announces the availability of Amazon EC2 I8ge instances in AWS Europe (London) and Canada (Central)  regions. I8ge instances are powered by AWS Graviton4 processors and deliver up to 60% better compute performance compared to previous generation Graviton2-based storage optimized Amazon EC2 instances. I8ge instances use the third generation AWS Nitro SSDs, local NVMe storage, and deliver up to 55% better real-time storage performance per TB compared to previous generation Amazon EC2 Im4gn instances . They offer up to 60% lower storage I/O latency and up to 75% lower storage I/O latency variability compared to Im4gn instances.

I8ge instances are storage-optimized instances, and offer up to 120TB of local NVMe storage. They are ideal for workloads that demand rapid local storage with high random read/write performance and consistently low latency for accessing large datasets. These versatile instances are offered in eleven different sizes including two metal sizes, providing flexibility to match customers’ computational needs. They deliver up to 180 Gbps of network performance bandwidth and 60 Gbps of dedicated bandwidth for Amazon Elastic Block Store (EBS), ensuring fast and efficient data transfer for the most demanding applications.

To begin your Graviton journey, visit the Level up your compute with AWS Graviton page. To get started, see AWS Management Console, AWS Command Line Interface (AWS CLI), and AWS SDKs. To learn more, visit the I8ge instances page.

Today, AWS announces the general availability of the AWS Lambda Durable Execution SDK for .NET, empowering C# developers to build resilient, long-running workflows using Lambda durable functions. With this SDK, developers can create multi-step applications like payment processing pipelines, AI agent orchestration, and human-in-the-loop approvals directly in their applications without implementing custom progress tracking or integrating external orchestration services. 

Lambda durable functions extend Lambda's event-driven programming model with operations that checkpoint progress automatically and pause execution for up to a year when waiting on external events. The AWS Lambda Durable Execution SDK for .NET provides an idiomatic C# experience for building with Lambda durable functions. It includes steps for progress tracking, callback integration for human and agent-in-the-loop workflows, durable invocation for reliable function chaining, and waits for efficient suspension. The SDK installs from NuGet into the .NET toolchain you use today. The local testing emulator in the SDK enables developers to build and debug locally before deploying to production.

To get started, see the Lambda durable functions developer guide and the AWS Lambda Durable Execution SDK for .NET on NuGet. For Regional availability and pricing details, see the AWS Regional Services List and AWS Lambda Pricing.

 

Amazon RDS for MySQL now supports version community MySQL 9.7 in the Amazon RDS Database Preview Environment, allowing you to evaluate the latest Release on Amazon RDS for MySQL. This preview environment provides a sandbox where you can test applications and explore new MySQL 9.7 capabilities before they become generally available.

MySQL 9.7 is the latest Long-Term Support release for community MySQL. MySQL Long-Term Support releases include bug fixes, security patches, as well as new features. Please refer to the MySQL 9.7 release notes for more details about this release.

Amazon RDS Database Preview Environment database instances are retained for a maximum of 60 days and are automatically deleted after the retention period. Amazon RDS database snapshots created in the preview environment can only be used to create or restore database instances within the preview environment. Amazon RDS Database Preview Environment database instances are priced the same as production RDS instances created in the US East (Ohio) Region. For further information, see Working with the Database Preview Environment.

AWS now enables customers who pay through electronic funds transfer to configure preferences for how credit memos are automatically applied to outstanding invoices. Customers can choose from different application preferences directly on the Billing and Cost Management console to match their internal payment processes. These options include combinations of applying credit memos to the original invoice, next eligible invoice, and oldest unpaid invoice. By default, credit memos are applied to the original invoice first, then to future invoices, unless a different preference is selected.

Automatic credit memo application preferences are available in all commercial AWS Regions. To get started, visit the Payment Preferences page in the AWS Billing and Cost Management console. To learn more, see Managing balance application preferences. 

Amazon Elastic Container Service (Amazon ECS) introduces zone-aware routing for ECS Service Connect, enabling customers to reduce cross Availability Zone (AZ) data transfer costs and latency by automatically prioritizing service-to-service traffic within the same AZ.

With this launch, ECS Service Connect preferentially routes requests to endpoints in the same AZ as the originating task while dynamically adjusting traffic weights as endpoints scale to maintain balanced load across target services. Previously, as customers distributed their applications across AZs for resiliency, service-to-service traffic led to significant cross-zone data transfer, requiring trade-offs between cost and resilience. Zone-aware routing eliminates this trade-off, and when local endpoints become unhealthy or fall below capacity thresholds, traffic automatically redistributes across healthy AZs to maintain availability without overloading any single zones.

Zone-aware routing is enabled by default for all new and existing services and requires no additional infrastructure or application code changes. Existing services require a one-time redeployment to enable the new routing behavior. You can use Amazon VPC Flow Logs with AZ metadata to monitor cross-AZ traffic patterns and validate routing effectiveness. This feature is available in all AWS commercial and AWS GovCloud (US) Regions, where ECS Service Connect is supported at no additional cost. For more details, refer to our documentation and launch blog post.

AWS Clean Rooms now supports larger worker types of 32 vCPUs with 244 GB of memory for SQL analyses, allowing you to run more compute and memory-intensive workloads. Previously, the largest worker type available was 16 vCPUs with 120 GB of memory. Larger workers can help improve performance for complex queries on large datasets, optimizing costs for customers. For example, an advertiser collaborating with a publisher can use larger workers to run complex feature engineering across billions of records to train their multi-touch attribution model, accelerating time-to-value and optimizing costs.

AWS Clean Rooms helps companies and their partners easily analyze and collaborate on their collective datasets without revealing or copying one another’s underlying data. For more information about the AWS Regions where AWS Clean Rooms is available, see the AWS Regions table. To learn more about collaborating with AWS Clean Rooms, visit AWS Clean Rooms.

Today, AWS announces a research preview of aws-bench, an open-source benchmark that measures how accurately and efficiently AI agents complete real-world AWS tasks. Model providers and AI researchers building agents that operate on AWS infrastructure need an objective, reproducible way to measure performance and diagnose failures. aws-bench provides a public suite of test cases derived from analysis of real AWS usage, including investigation, troubleshooting, and infrastructure creation tasks.

Each test case pairs a natural-language query with a defined cloud resource state and a ground-truth answer, so you can score any agent or model on a consistent, verifiable basis. Researchers and model providers can use aws-bench to improve foundation model performance on AWS tasks, improve agent harnesses, and track improvement progress. The release includes an easy-to-use CLI tool to instantiate testing environments, execute and score evaluation runs, and reset resource state.

aws-bench is available now on GitHub. To get started, follow the setup instructions on the README. 

AWS Certified Machine Learning Engineer – Associate が MLA-C02 にアップデートされます。生成 AI、エージェンティック AI、基盤モデル / LLM、Amazon Bedrock のカバレッジが追加され、進化する ML エンジニアの役割に対応します。ベータ版試験（英語のみ）の受験予約は 2026 年 9 月 1 日開始、現行版 (MLA-C01) の英語での最終受験日は 9 月 28 日です。日本語を含む他言語は、ベータ期間中も現行版で引き続き受験可能です。全言語対応の標準版 MLA-C02 は 2027 年初頭に提供予定です。

2026 年 6 月 25 日(木)、26 日(金)の 2 日間にわたって幕張メッセで開催された AWS Summit Japan 2026 では、AWS Expo の Industry Zone に「AWS for Telecom」として通信業界に関する展示ブースを出展しました。

本記事では、製造業のお客様が OT をクラウドに接続する際に特に重要となるISA/IEC 62443 の様々な概念を、AWS サービスを使ってどのように実現するのかをご紹介します。お客様は、AWS サービスを活用した製造現場向けのクラウドアーキテクチャを構成することで、ISA/IEC 62443 に沿ったクラウド側のセキュリティ対策を実装できます。

2026年7月15日 (水) に開催された「Neuron Community - 2026 Vol.1」の様子をレポートします。このイベントは、「Neuron Community」の協力のもと開催しました。Neuron Community は、ユーザー間で AWS Trainium / AWS Inferentia エコシステムに関する情報や知見の共有を促進するための場として発足したものです。今回は、AWS Summit Japan 2026 開催後ということもあり、AWS Summit Japan の振り返りや、AWS Neuron のアップデート情報が多めの内容となりました。

Black Hat 2026 (Aug 1-6, 2026) brings together over 22,000 security practitioners, researchers, and CISOs who build, break, and defend enterprise infrastructure. They’re security professionals who push the limits of offensive and defensive security and demand proof over promises. As frontier security models like Mythos reshape the enterprise landscape, they need security that operates at […]

Bulletin ID: 2026-063-AWS
Scope: AWS
Content Type: Important (requires attention)
Publication Date: 07/23/2026 08:30 AM PDT

Description:

The AWS API MCP Server (awslabs.aws-api-mcp-server) is an open-source MCP server that lets AI assistants execute AWS CLI commands against a user's AWS account. It includes an optional, user-configured security policy that can deny or gate specific AWS operations. We identified CVE-2026-16584. On startup, the server loads the data used to enforce this security policy. If that initialization fails, the server continues running with the per-request policy check skipped for the lifetime of the process. When a security policy is configured without the fail-closed modes enabled, an actor could then cause AWS API operations that the policy was configured to deny or gate to execute without enforcement. IAM permissions on the configured credentials remain in effect and are unaffected.

Impacted versions: >= 0.2.13 AND < 1.3.47

Please refer to the article below for the most up-to-date and complete information related to this AWS Security Bulletin.

Bulletin ID: 2026-064-AWS
Scope: AWS
Content Type: Important (requires attention)
Publication Date: 07/23/2026 11:30 AM PDT

Description:

Smithy-RS is a Rust code generation and runtime framework that generates HTTP clients and servers from Smithy interface definitions, powering the AWS SDK for Rust and custom service implementations. We identified CVE-2026-16756 where the allocation of resources without limits in the default aws-smithy-http-server serve() path allows unauthenticated "Slowloris" denial of service.

Impacted versions: aws-smithy-http-server <= 0.66.4

Please refer to the article below for the most up-to-date and complete information related to this AWS Security Bulletin.

Bulletin ID: 2026-065-AWS
Scope: AWS
Content Type: Important (requires attention)
Publication Date: 07/23/2026 13:00 PM PDT

Description:

The AWS Bedrock AgentCore Python SDK (bedrock-agentcore) provides tools for building AI agents, including a Code Interpreter client that installs Python packages into a managed sandbox. We identified CVE-2026-16796, an improper neutralization of argument delimiters in the install_packages() method that might allow a remote authenticated user to execute arbitrary commands within the Code Interpreter sandbox via crafted package name arguments.

Impacted versions: bedrock-agentcore version <1.18.1

Please refer to the article below for the most up-to-date and complete information related to this AWS Security Bulletin.

This post focuses on why classic retrieval falls short on multi-part questions, how the AgenticRetrieveStream API works (including request construction and trace parsing), and when to choose it over the standard Retrieve API.

Amazon Bedrock AgentCore optimization surfaces silent behavioral failures in production AI agents: the ones that pass every health check but still deliver wrong outcomes. Learn how insights discovers, explains, and ranks failure patterns across sessions so you can fix the highest-impact issues first.

This post shows you how to build multi-Region carrier performance dashboards in Quick Sight using Highcharts custom visualizations to overcome native chart limitations. You will learn how to maintain data sovereignty across AWS Regions while creating unified visualizations through the Quick Sight federated dataset capability. The solution includes production-ready chart configurations and addresses security, compliance, and scalability requirements.

In this post, we explore how Jefferies overcame these challenges with a solution built on Strands Agents, an agent harness SDK for building AI agents that can reason, plan, and act by orchestrating calls to foundation models (FMs) and external tools. The solution uses large language models (LLMs), Amazon Bedrock, and Amazon Bedrock Knowledge Bases. It also uses Model Context Protocol (MCP), an open standard that helps AI agents securely connect to diverse data sources and tools through a unified interface. We cover the solution overview, the rationale for selecting the underlying technology stack, lessons learned, and the business impact the solution created at Jefferies.

Together, Motorway and AWS built an end-to-end evaluation pipeline that reduced incorrect results from 1 in 8 queries to 1 in 50 and cut issue detection time from few hours to few minutes. The pipeline combines the Strands Agents SDK with Amazon Bedrock AgentCore, a fully managed service for deploying and operating AI agents at scale. In this post, you will learn how to build this pipeline for your own agents.

In this post, we explain how Amazon Bedrock Guardrails can be configured for code generation workflows with coding assistants to overcome these constraints. With these best practices, you can build an efficient blueprint helping you with effective capacity planning with robust safety coverage.