AWS News - 2026-10-13

Announcing consumption-based pricing for short-term memory in Amazon Bedrock AgentCore

Amazon Bedrock AgentCore memory now prices short-term memory (STM) based on the data your agents ingest, retrieve, and store, replacing the previous flat rate per event. AgentCore memory gives agents in-session context through short-term memory and persistent knowledge across sessions through long-term memory. With consumption-based pricing, developers building agents that write frequent, small events pay in proportion to what they use.

STM is priced across three dimensions: ingestion at $1.00 per GB, retrieval at $0.20 per GB, and storage at $0.10 per GB-month. Ingestion and retrieval are billed per event with a 12 KB minimum and 64 KB maximum billable size, while storage is billed on each event's actual size. Most workloads see lower costs, with small-payload use cases seeing the largest reductions. You can manage storage costs by setting a shorter time to live (TTL) on your memory resources. Long-term memory pricing is unchanged.

Consumption-based pricing for AgentCore Memory short-term memory is available in all AWS Regions where AgentCore Memory is offered. To get started, visit the Amazon Bedrock AgentCore pricing page.

Amazon Connect Customer can now fill in the step-by-step guides that agents complete after a contact, using information extracted from customer conversations, such as booking identifiers, claim numbers, and conversation summaries. Managers can set which information populates each form field. Agents can review and submit rather than manually typing in each field, saving time and improving data accuracy during after contact work.

Contact center agents spend time after each contact entering frequently used information into a step-by-step guide. For example, an insurance agent who finishes a claims call might record the customer’s policy number, a summary of the call, and a disposition code. Now, managers can define rules for which information to extract from the conversation and which fields to automatically fill in the guide. The agent can then check the information, make any necessary corrections, and submit the guide, shortening after contact work and reducing data entry errors.

You can use step-by-step guides with information extraction in the following AWS Regions: US East (N. Virginia), US West (Oregon), Asia Pacific (Seoul), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Canada (Central), Europe (Frankfurt), and Europe (London). Contact summaries are also available in Africa (Cape Town).

To learn more and get started, see step-by-step guides in the Amazon Connect Customer administrator guide.  

Amazon Elastic Compute Cloud (Amazon EC2) G7 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs are now available in Europe (Frankfurt) and Europe (Stockholm) Regions. G7 instances deliver up to 4.6x AI inference performance and up to 2.1 graphics performance compared to G6 instances. G7 instances also deliver faster performance for GPU-accelerated data analytics workloads.

Customers can use G7 instances for deploying AI models for language translation, video and image analysis, speech recognition and recommender systems. They also accelerate graphics workloads such as creating and rendering real-time, cinematic-quality graphics and game streaming. Additionally, G7 instances support video transcoding, spatial computing, and data analytics workloads. G7 instances feature up to 8 NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs with 32 GB of memory per GPU and custom Intel Xeon 6 processors. They support up to 192 virtual CPUs (vCPUs) and up to 700 Gbps of Elastic Fabric Adapter (EFA) networking bandwidth. They also support up to 768 GiB of system memory, and up to 7.6 TB of local NVMe SSD storage.

You can use G7 instances for Amazon EC2 in the following AWS Regions: US West (Oregon), US East (N. Virginia, Ohio) and Europe (Spain, Frankfurt, Stockholm). You can purchase G7 instances as On-Demand Instances, Spot Instances, or as part of Savings Plans.

To get started, visit the AWS Management Console, AWS Command Line Interface (CLI), and AWS SDKs. To learn more, visit the G7 instance page.

Starting today, you can deliver web applications built on AWS Lambda Web Functions through CloudFront's global edge network.  

With CloudFront in front of a Lambda Web Function, you get TLS termination at the edge for lower latency and faster connections, caching for better performance, and custom domains with managed TLS certificates. No changes to application code are needed. CloudFront Origin Access Control (OAC) restricts your Lambda Web Function so it's reachable only through CloudFront, so unintended requests that bypass your security protections do not reach your origin. With OAC in place, CloudFront's edge network protects your Lambda Web Function from network-layer DDoS attacks, and you can attach AWS WAF for additional application-layer security. For applications that authenticate their own users, OAC ensures that original credentials reach the application unchanged, so existing sign-in flows keep working.

CloudFront support for AWS Lambda Web Functions is available worldwide, except in the CloudFront China region. To get started, use the CloudFront Console, SDK, CLI, or CloudFormation to configure your Lambda Web Function as a CloudFront origin. For more information, refer to the CloudFront Developer Guide. There are no additional fees associated with this feature.

Amazon Relational Database Service (Amazon RDS) for SQL Server now supports the latest Cumulative Updates (CU) for Microsoft SQL Server. This release includes support for Microsoft SQL Server 2025 CU8 KB5104822 (RDS version 17.00.4075.5.v1).

For additional information on the improvements and fixes included in these updates, see Microsoft documentation for KB5104822. We recommend that you upgrade your Amazon RDS for SQL Server instances to apply these updates using Amazon RDS Management Console, or by using the AWS SDK or CLI. You can learn more about upgrading your database instance in the Amazon RDS SQL Server User Guide for upgrading your RDS Microsoft SQL Server DB engine.

Amazon Relational Database Service (Amazon RDS) for SQL Server now supports the latest Cumulative Updates (CU) for Microsoft SQL Server. This release includes support for Microsoft SQL Server 2025 CU8 KB5104822 (RDS version 17.00.4075.5.v1).

For additional information on the improvements and fixes included in these updates, see Microsoft documentation for KB5104822. We recommend that you upgrade your Amazon RDS for SQL Server instances to apply these updates using Amazon RDS Management Console, or by using the AWS SDK or CLI. You can learn more about upgrading your database instance in the Amazon RDS SQL Server User Guide for upgrading your RDS Microsoft SQL Server DB engine.

AWS DataSync Enhanced mode now supports Amazon FSx for OpenZFS as a source or destination location, NFS version 4.2 for on-premises NFS servers and FSx for OpenZFS file systems, and a new ObjectMetadata option for object storage transfers. Previously, customers transferring data to or from FSx for OpenZFS were limited to Basic mode.

NFS 4.2 lets you transfer data from NFS servers on the latest version without reconfiguring them to an older protocol version. The ObjectMetadata option gives you control over whether an object's user metadata is copied to the destination, which helps when destination metadata must stay intact, such as buckets served by AWS Storage Gateway, or when source metadata keys are not compatible with the destination, such as Azure Blob Storage. Existing tasks remain unaffected and continue to run as scheduled.

These capabilities are available for Enhanced mode tasks in all AWS Regions where AWS DataSync Enhanced mode is supported. To learn more, visit the AWS DataSync documentation.

Amazon Route 53 Domains now supports Premium Domain Names. Customers can now search for, register, transfer, renew, and restore registry premium names directly through Route 53.

Most domain registries set aside a selection of "premium" domains - names that are short, memorable, or in high demand. Previously, Route 53 Domains did not support these premium names, which meant customers who wanted them had to register and manage them through other registrars, fragmenting their portfolios across multiple providers. With this launch, customers can now discover and acquire premium domains directly in Route 53 Domains, consolidate their existing premium domains from other registrars, and manage premium and standard domains together through a single, unified experience. Finding and registering a premium domain works just like it does for standard domains. Premium domains appear automatically in domain search and availability results, with pricing shown upfront, so customers know the cost before they commit.

To learn more about Amazon Route 53 Domains and start registering premium domains, visit the Amazon Route 53 Domains page. For a complete list of supported top-level domains, see the Route 53 TLD list.

AWS Lambda Web Functions is a serverless compute primitive that runs your HTTP framework natively, delivering fast, consistent responses through multi-concurrent execution and automated global distribution.

このブログ記事は、AWS ソリューションアーキテクト 小塚が執筆し、ソニー銀行株式会社および富士通株式会社に監 […]

AI 主権は、データ主権と運用主権を基盤に、AI の出力で文化的規範や価値観、現地の言語を守るという新たな考慮事項を加えた概念です。本記事では、チップ、デプロイ先、モデルを自由に選べる AI スタック全体でのコントロールと選択肢を紹介します。また、Meltemi や SEA-LION などの現地の言語に対応したモデルの支援、AWS Nitro System による検証可能なデータアクセスのコントロール、ISO/IEC 42001 認証による透明性と保証についても説明します。

AWS は、コンプライアンスガイド「ISO/IEC 42001:2023 on AWS」を公開しました。このガイドでは、AWS サービスを使用して AI マネジメントシステム (AIMS) を設計、運用する方法を説明します。ISO 42001:2023 の箇条 4～10 と附属書 A の管理策を AWS サービスにマッピングし、AI ワークロードへの AWS 責任共有モデルの適用方法も解説します。エビデンス収集の自動化や、Infrastructure as Code を活用して AI コンプライアンス活動を運用に組み込む方法も取り上げ、AI ガバナンスの強化と認証審査への準備を支援します。

生成 AI の導入が急速に進む中、組織には責任ある AI ガバナンスへの体系的なアプローチが求められています。本記事では、国際規格 ISO/IEC 42005:2025 に基づく AI システム影響評価の概要と、既存のリスク管理プロセスへの統合方法を解説します。再評価のトリガーやトリアージを含む評価プロセスの構築方法も取り上げます。また、ISO/IEC 42001 認証の取得における ISO/IEC 42005 の役割や、AWS Well-Architected Responsible AI Lens をはじめとする規格との整合を支援する AWS のツールも紹介します。

今回の週刊生成 AI with AWSでは、Claude Haiku 5.5 や Amazon Nova 2.5 Sonic など、新モデルの提供開始が相次いだ 1 週間でした。 Claude Code on Amazon Bedrock を非エンジニア中心に全社展開したアテナ様の事例や、Kiro CLI V3 への切り替え案内もまとめています。 Amazon Bedrock の自動推論チェックを扱う記事 や、AI エージェントに移行作業を任せる AI Modernization Flow など今週も盛りだくさんです。

Amazon Nova 2.5 Sonic が GA、Claude Haiku 5.5 が AWS で利用可能に、Amazon GuardDuty RDS Protection が Aurora / RDS のデータ流出と破壊を検知、AWS Security Hub が検出結果の S3 エクスポートに対応など

「今収集し、後で偽造する」脅威に備え、AWS KMS、AWS Private CA、IAM Roles Anywhere を組み合わせて、ML-DSA による耐量子ソフトウェアサプライチェーンを構築する方法を紹介します。ML-DSA PKI 階層の構築から、CI/CD パイプラインでのバイナリ、コンテナイメージ、SBOM への署名、デプロイ時の検証、SCP による予防的コントロールまでを解説します。さらに、CNSA 2.0 や EU サイバーレジリエンス法の期限に向けた推奨アクションもまとめます。

この記事では、組織の暗号アルゴリズムポリシーを AWS の予防的コントロールと発見的コントロールに変換する方法を紹介します。ML-DSA-65 と ML-DSA-87 を承認し、ML-DSA-44 を除外する例を用いて、kms:KeySpec 条件キーと SCP、IAM ポリシー、アクセス許可の境界、キーポリシーを組み合わせた多層防御を解説します。AWS Private CA や TLS 証明書に対するプロビジョニング制御、AWS CloudTrail と AWS Config によるドリフト検出、暗号アジリティを実現する設計のポイントも説明します。

Amazon GuardDuty 拡張脅威検出の攻撃シーケンスの検出結果に、5 つの新しいインジケーターが追加されました。Amazon Inspector、Amazon VPC Reachability Analyzer、Amazon Macie、AWS Security Hub から得られる脆弱性、悪意のあるパッケージ、到達可能性、設定ミス、機密データの情報を、弱いランタイムシグナルと相関付けます。これにより、ノイズは重要度が CRITICAL の対応可能な検出結果に変わります。この記事では、Amazon EKS と Amazon EC2 での実例に加えて、EPSS を重視した優先順位付けによるインシデント対応の手順を解説します。

これまで、管理対象デバイスでエンタープライズユーザーが Kiro にどの方法でサインインするかを制御する手段はありませんでした。組織として IAM Identity Center の利用を必須にしていても、ユーザーは個人アカウント（Google、GitHub、Builder ID）で Kiro にサインインできてしまっていました。Kiro では、利用できるサインイン方法を管理者が決められるようになりました。ユーザーには組織が許可した選択肢だけが表示されます。

「フィジカル AI 開発支援プログラム by AWS ジャパン」第 2 期の応募受付を開始 AWS ジャパンは […]

Software supply chain attacks have grown in both frequency and sophistication. Recent high-profile compromises of build systems and open source dependencies demonstrate that attackers increasingly target the trust mechanisms that verify software integrity. Those trust mechanisms rely on digital signature algorithms, and today many use traditional algorithms such as Rivest-Shamir-Adleman (RSA) and the Elliptic Curve […]

You want to set a cryptographic algorithm policy and stop non-compliant choices from being made in your Amazon Web Services (AWS) environment. This post shows you how. As you adopt post-quantum cryptography, one of the first decisions you face is which algorithms and parameter sets fit your business needs, balancing security against signature size, key […]

Amazon GuardDuty Extended Threat Detection correlates weak GuardDuty signals into a single, critical-severity attack sequence finding when their composition crosses a threshold of credibility. On May 20, 2026, the service started attaching five new indicator types to those sequences: MALICIOUS_PACKAGE, MISCONFIGURATION, REACHABILITY, SENSITIVE_DATA, and VULNERABILITY. For more information, see Document history for Amazon GuardDuty and […]

In Part 2 of our Bulk Executor for Amazon DynamoDB series, we share how we built the tool: the design tradeoffs, surprises, and workarounds we hit while working with DynamoDB and AWS Glue. Consider it a travel guide for anyone building their own bulk tooling on AWS Glue.

Learn how VIPRE Security built an AI-powered email security solution on AWS to stop business email compromise. VIPRE Integrated Email Security uses machine learning and a serverless architecture on Amazon ECS, Amazon SQS, and Amazon S3 to detect and block sophisticated email impersonation in real time, without MX record changes.

Learn how to deploy one large language model across two GPU nodes on Amazon EKS using the multi-node Ray LLM Deep Learning Containers, with separate prefill and decode nodes behind one OpenAI-compatible endpoint. The pre-built container ships the full inference stack, so you configure the deployment instead of assembling it.

Learn how to deploy decision models with llama.cpp on Amazon SageMaker AI. This technical how-to serves OpenAI-compatible chat completions with the AWS Deep Learning Container for llama.cpp, then extends the runtime with SemIf to return probabilities over declared options on CPU or GPU.

Learn how to run OpenAI GPT-6 Sol, Luna, and Astra on Amazon Bedrock through global cross-Region inference from Singapore, Jakarta, Thailand, Malaysia, and Taipei. This post covers the Converse and OpenAI-compatible APIs, IAM and SCP permissions, quota sizing, and monitoring inference from your source Region.