サーバーレスの新プリミティブと AI エージェント運用、AWS Summit Japan 2026 に向けた記事が数多く公開された一日でした。News Blog では、ユーザーや AI 生成コードを VM レベルで分離実行できる新しいサーバーレスプリミティブ AWS Lambda MicroVMs を発表。Firecracker 上で近瞬時起動と最大 8 時間の状態保持を両立し、US East、US West、東京、アイルランドで提供されます。What's New では AWS Batch の順序指定アロケーション戦略、IAM Identity Center のアカウント/アプリ別クォータ、Network Firewall のデフォルトドロップ変更、MSK の AI Agent Skills、AWS Transform の全商用リージョン対応、OpenSearch の AI 支援マイグレーション、AWS Marketplace 経由の Claude Tag ベータ、Amazon Connect の Agentic CX designer プレビューなどが並びました。Architecture Blog は Bedrock と Verified Permissions によるマルチテナント RAG や AI 駆動のレジリエンスフレームワークを、ML Blog は多モーダル航空画像検索や AgentCore Payments を解説。Japan Blog は Summit Japan 2026（幕張メッセ）関連の展示紹介や各種機能の日本語解説を多数掲載しました。
サーバーレス: AWS Lambda MicroVMs による VM レベル分離実行の新プリミティブ
AI エージェント運用: MSK AI Agent Skills、OpenSearch AI 支援マイグレーション、AWS Transform 継続的モダナイズ
セキュリティ/ガバナンス: egress 制御、Network Firewall デフォルト変更、IAM Identity Center クォータ分離
生成 AI アーキテクチャ: マルチテナント RAG、AI レジリエンスフレームワーク、AgentCore Payments
コンタクトセンター/協業: Amazon Connect Agentic CX designer プレビュー、Claude Tag ベータ (AWS Marketplace)
AWS Summit Japan 2026: 建設・不動産や Immersive Experience などブース展示の日本語紹介
Amazon MSK now offers AI Agent Skills that give AI coding assistants expert, up-to-date guidance for operating Amazon MSK. The skills provide expert guidance for common operational tasks such as troubleshooting, sizing, configuring, monitoring, and migration from external Kafka clusters.
Teams can leverage these skills to keep their clusters healthy and performant, and to migrate their external Kafka workloads to MSK Express to take advantage of up to 3 times more throughput per broker, scale up to 20 times faster, and reduced recovery time by 90 percent as compared to Standard brokers running Apache Kafka. The skills turn tasks that once required specialized knowledge into a guided experience developers can complete quickly, on their own.
You can use the MSK skills with your existing AI coding agent - Kiro, Claude Code, or Cursor. To get started, configure the Agent Toolkit for AWS using the AWS CLI, then ask your coding agent a question, such as "What broker type and size should I use for my MSK cluster?" or "Is my Kafka cluster compatible with MSK Express?"
Amazon MSK Replicator now supports mutual TLS (mTLS) authentication for data replication from external Apache Kafka clusters - including on-premises, self-managed on AWS, or other cloud providers - to Amazon MSK Express brokers. With this capability, external Apache Kafka clusters configured with mTLS authentication can now use MSK Replicator to migrate workloads to MSK Express brokers, support disaster recovery by using MSK Express-based clusters as a failover or backup target, and enable data distribution across hybrid and multi-cloud environments.
MSK Replicator is a feature of Amazon MSK that automates data replication between Kafka clusters, eliminating the need to manage custom replication infrastructure or configure open-source tools. Previously, MSK Replicator supported SASL/SCRAM authentication only for connecting to external Apache Kafka clusters. With this launch, you can now also use mTLS authentication with MSK Replicator to replicate data from external Kafka clusters to Express brokers on Amazon MSK. Unlike self-managed replication tools, MSK Replicator lets you retain your original Kafka topic names during replication while automatically avoiding infinite replication loops. It also synchronizes consumer group offsets bidirectionally, enabling you to move producers and consumers across clusters independently, in any order, without coordination constraints or the risk of data loss.
This new capability is supported in all AWS Regions where MSK Express brokers are available. Visit the MSK Replicator documentation, product page, pricing page, and this AWS blog post to learn more.
AWS IAM Identity Center now supports separate quotas for the number of AWS accounts and applications that can be configured in an IAM Identity Center instance. By default, you can configure up to 7,000 AWS accounts and up to 7,000 applications independently, so that using more of one does not consume capacity from the other. Quotas can be further increased by submitting a quota increase request through AWS Service Quotas console.
Customers with existing higher limits are automatically granted the same limit for both accounts and applications, with no action required. Organizations managing thousands of AWS accounts can now onboard applications without consuming account quota capacity.
This update is available in all AWS Regions where IAM Identity Center is available.
To learn more, see Quotas for IAM Identity Center. Visit the IAM Identity Center product page to get started.
AWS Batch now offers the Best Fit Progressive Ordered (BFPO) and Spot Capacity Optimized Prioritized (SCOP) allocation strategies, giving you more control over instance type prioritization in your compute environments. BFPO and SCOP enable you to manually define instance type ordering based on your workload-specific performance characteristics.
To use these features in AWS Batch, specify BEST_FIT_PROGRESSIVE_ORDERED allocation strategy for your on-demand compute environments or SPOT_CAPACITY_OPTIMIZED_PRIORITIZED for your Amazon EC2 Spot compute environments and provide an ordered list of instance types or families. These features are available via the AWS Batch API (CreateComputeEnvironment or UpdateComputeEnvironment) or the AWS Batch Management Console.
BFPO and SCOP allocation strategies are supported today in all AWS Regions where AWS Batch is available. For more information, see the AWS Batch User Guide.
Amazon Connect Customer now offers Agentic CX designer (NLX) in preview, a no-code canvas for designing and deploying AI-powered self-service experiences. You can build and launch voice and digital experiences that bring agentic and deterministic AI together to transform how you serve customers with the control and reliability enterprises demand. Your business teams can go from designing conversations and integrating with the systems that run your business, to testing and simulating, to launching production-ready experiences in weeks, not months.
AWS Network Firewall now uses "Application drop established (server-directed only)" as the default stateful action for all newly created firewall policies, replacing the previous default of "Application drop established (bidirectional)" (formerly named "Application layer drop established"). No action is required to benefit from this change when creating new policies.
AWS Network Firewall is a managed service that lets you deploy network protections across your Amazon VPCs. Previously, the “Application drop established (bidirectional)” default could silently drop legitimate server-to-client TCP packets, such as window updates, keep-alives, and resets — causing intermittent connection failures that were difficult to diagnose. With the safer default now in place, new policies avoid this issue.
If your existing environment requires “Application drop established (bidirectional)” to support post-quantum cryptography (PQC) fragmented TLS handshakes, refer to our documentation for guidance on on switching to "Application drop established (server-directed only)" or adding the “to_server” flag to your TCP drop rules so legitimate flow control packets are not blocked.
This update is available in all AWS Regions where AWS Network Firewall is offered. To get started, see Managing evaluation order for Suricata compatible rules in the AWS Network Firewall service documentation.
AWS Transform for migrations now supports all AWS commercial regions as migration targets. A migration target region is the AWS region where migrated resources are deployed, including landing zones, network infrastructure, and server rehosting. Customers can now deploy workloads in any commercial region, making it easier to meet data residency requirements.
The new migration target regions are: US East (N. California), Africa (Cape Town), Asia Pacific (Bangkok), Asia Pacific (Hong Kong), Asia Pacific (Hyderabad), Asia Pacific (Jakarta), Asia Pacific (Kuala Lumpur), Asia Pacific (Melbourne), Asia Pacific (New Zealand), Asia Pacific (Taipei), Canada (Calgary), Europe (Milan), Europe (Spain), Europe (Zurich), Mexico (Querétaro) and Middle East (Tel Aviv).
Target region selection is available in the AWS Transform for migrations workflow. For the most up-to-date availability information, see the supported migration target region list.
AWS introduces Lambda MicroVMs, a new serverless compute primitive that provides VM-level isolation, near-instant launch and resume speeds, and state preservation for executing user or AI-generated code. You can now give each user or job their own compute environment to securely run code without managing virtualization infrastructure or choosing between isolation, speed, and state retention.
Developers are increasingly building multi-tenant applications that execute code supplied by end users or AI for use cases such as interactive coding environments, data analytics platforms, coding assistants, and vulnerability scanning platforms. For these applications, developers need to allocate a separate, isolated execution environment per user or session to limit the impact of incorrect or malicious code on other concurrently running users or jobs. Previously, developers needed to choose between strong isolation, fast launch times, and state retention when building these applications. Starting today, Lambda MicroVMs provides you these capabilities without any trade-offs. You get VM-level isolation, near-instant launch speeds, and the ability to suspend and resume execution for up to 8 hours. Lambda MicroVMs is built on Firecracker virtualization, the technology powering more than 15 trillion monthly Lambda Function invocations.
To get started, create a MicroVM image from your Dockerfile, then launch MicroVMs from that image. Give each user or job their own MicroVM with a dedicated HTTPS URL that supports popular connectivity protocols such as HTTP/2, gRPC, and WebSockets.
Lambda MicroVMs is available today in the following AWS Regions: US East (N. Virginia), US East (Ohio), US West (Oregon), Asia Pacific (Tokyo), and Europe (Ireland). To learn more, visit the AWS Lambda MicroVMs developer guide and the launch blog post. Get started with MicroVMs through the AWS Lambda console, AWS CloudFormation, AWS Cloud Development Kit, or use the Agent Toolkit for AWS with your preferred Agentic development tools. You pay for baseline compute resources while your MicroVM is running, and only for the active duration of additional resources consumed when your workload exceeds the baseline. To learn more about pricing, see Lambda MicroVMs pricing.
AWS HealthOmics now supports Nextflow profiles, enabling customers to activate predefined execution settings at run time. Nextflow profiles allow customers to define reusable settings and select them at the point of execution, making it easy to switch between execution settings without modifying workflow source code. AWS HealthOmics is a HIPAA-eligible service that helps healthcare and life sciences customers accelerate scientific breakthroughs at scale with fully managed bioinformatics workflows.
With Nextflow profiles, you can cleanly separate platform-specific settings such as resource limits or execution options from core workflow logic. You can switch between development and production settings without creating separate workflow definitions. This reduces errors from manual edits, accelerates workflow portability, and saves time when scaling from development to production. If you use nf-core workflows, you can now activate the built-in and institutional profiles those pipelines already ship with.
You can now specify one or more Nextflow profiles in your workflow runs in all AWS HealthOmics Regions: US East (N. Virginia), US West (Oregon), Europe (Frankfurt, Ireland, London), Israel (Tel Aviv), and Asia Pacific (Singapore, Seoul). To learn more, visit the Nextflow Profiles section on HealthOmics Nextflow engine settings documentation.
Amazon G7e instances feature up to 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, with 96 GB of memory per GPU, and 5th Generation Intel Xeon processors. They support up to 192 virtual CPUs (vCPUs) and up to 1600 Gbps of Elastic Fabric Adapter networking bandwidth. G7e instances support NVIDIA GPUDirect Peer to Peer (P2P) that boosts performance for multi-GPU workloads. Multi-GPU G7e instances also support NVIDIA GPUDirect Remote Direct Memory Access (RDMA) with EFAv4 in EC2 UltraClusters, reducing latency for small-scale multi-node workloads. Customers can use G7e instances to deploy large language models (LLMs), agentic AI models, multimodal generative AI models, and physical AI models. G7e instances offer the highest performance for spatial computing workloads as well as workloads that require both graphics and AI processing capabilities.
Amazon EC2 G7e instances are available for SageMaker Studio notebooks in the AWS US East (N. Virginia and Ohio) and US West (Oregon) regions.
Visit developer guides for instructions on setting up and using JupyterLab and CodeEditor applications on SageMaker Studio. For pricing information on these instances, please visit our pricing page.
Migration Assistant for Amazon OpenSearch Service now includes an AI-assisted experience that simplifies moving your self-managed Apache Solr, Elasticsearch, or OpenSearch deployments to OpenSearch Serverless or Managed Clusters. With the new assistant, you can use your preferred AI tools like Kiro, Claude Code, and others to plan a migration, deploy necessary infrastructure, and execute both historical and live traffic migration.
Migrations are often complex and require weeks of planning before any data movement can begin and even then, the process can be error-prone. We launched Migration Assistant in December 2023 to simplify migrating existing and live data from self-managed clusters to Amazon OpenSearch Service by automating manual migration tasks. The new AI-assisted experience takes this further: it provides an agent-guided workflow that helps you structure, execute, and validate your data migration faster and more reliably. Additionally, Migration Assistant for Amazon OpenSearch Service now supports live traffic capture and replay for Solr. To get started, see Migration Assistant documentation.
Migration Assistant supports migrations to OpenSearch Serverless and Managed Clusters from various Solr, Elasticsearch, and OpenSearch versions. For more details about the versions supported, see the documentation. Migration Assistant is available in all commercial AWS Regions and AWS GovCloud (US) Regions where Amazon OpenSearch Service is available.
Anthropic is launching Claude Tag — bringing Claude directly into the channels where your team already works, starting with Slack. Claude Tag is available today in beta to AWS customers who access Claude Enterprise through AWS Marketplace.
Claude Tag is a new way for teams to work with Claude. Grant Claude access to selected channels, and connect it to whichever tools, data—and even codebases—you choose.. It's multiplayer, so anyone in the channel can tag @Claude in, and delegate tasks to it while they focus on other work. Claude builds context by remembering relevant information from the channels it’s in, and can plan out tasks to complete in the future. And, for security and governance teams, Claude Tag operates under its own identity, scoped per channel, with spend controls and ambient mode off by default.
Getting started with Claude Enterprise on AWS Marketplace
The experience for Claude Enterprise in AWS Marketplace customers is identical to first-party Claude Enterprise: same setup, same capabilities, same controls. Consumption-based pricing tracks usage rather than headcount, with org-wide budget visibility and per-channel limits. Customers use their existing Claude Enterprise on AWS entitlement — an admin provisions the agent identity in the Claude admin console (approximately one hour) and scopes it per channel.
To learn more, see the Claude Enterprise in AWS Marketplace
AWS launches a new serverless compute primitive, AWS Lambda MicroVMs. VM-level, isolated sandboxes with no shared kernel or resources between sessions. Rapid launch and resume, full lifecycle control, state preservation up to 8 hours, no infrastructure to manage.
IoT やコネクテッドデバイスの普及により、製造業の製品はハードウェア単体からソフトウェアで価値を提供する「ス […]
日本最大の “AWS を学ぶイベント” AWS Summit Japan が 2026 […]
6 月 25 日 (木)・26 日 (金) の 2 日間、千葉・幕張メッセにて開催される AWS Summit […]
Amazon Elastic Container Service (Amazon ECS) サービス自動スケー […]
2026 年 6 月 17 日、プレビューで提供が開始となった AWS DevOps エージェントに新しいリリ […]
2026 年 6 月 17 日、AWS Transform の新機能である AWS Transform – 継 […]
Kiro Web に、繰り返し発生するタスクを自動化するクラウドで動作する Automations 機能を追加しました。スケジュールを設定すると、Kiro が自律エージェントとしてサンドボックス上でタスクを実行し、変更を含むプルリクエストを自動的に開きます。ドキュメント更新、テスト生成、TODO 整理など、定期的な作業をまかせられます。
2026 年 6 月 17 日、Amazon Bedrock AgentCore での Web Search […]
2026 年 6 月 17 日、Amazon Bedrock マネージドナレッジベースを発表しました。これは、 […]
2026 年 6 月 18 日、AI 推論、グラフィックス、データ分析のワークロード向けに高性能な GPU ア […]
re:Invent 2025 では、開発ライフサイクル全体を通じてあらゆる環境でアプリケーションをプロアクティ […]
エンジニアリング組織で Claude Code のような AI コーディングエージェントを利用している場合、そ […]
本記事では、AWS サンプルアセットである AI エージェント SMART（Store Manager Agent for Retail Tech） についてのご紹介と、それを活用した株式会社ユナイテッドアローズ（以下、ユナイテッドアローズ）の取り組みについてご紹介します。小売業にとって、店舗の声をどう本部に届けるかは永遠のテーマです。売上数字の裏には、現場スタッフだけが感じている気づきが必ずあります。しかし店舗の日報や週報のフォーマットだけでは、その気づきを届けるのは難しいのが実情です。SMART は、店舗の気づきを AI の力で引き出し言語化して、本部に届けることを支援するために誕生しました。
When securing an Amazon Web Services (AWS) environment, teams naturally prioritize inbound controls, firewalls, WAFs, and access policies, because that’s where the most visible threats originate. Outbound traffic, on the other hand, tends to get less attention. It’s often left open by default to avoid breaking application dependencies and because the risk feels less immediate. […]
In this post, you’ll learn how to architect and implement a five-layer AI-powered resilience framework that automatically discovers dependencies, generates targeted experiments, and integrates with your existing Continuous Integration/Continuous Deployment (CI/CD) pipelines. First, we’ll explore the key challenges in resilience testing. Then, we’ll walk through the five-layer architecture that solves these challenges. Finally, we’ll show you how to implement this, with phased rollout guidance for pilot, expansion, and organization-wide deployment.
Avanse Financial Services, India’s leading education loan providers, migrated to a cloud-native lakehouse architecture using Amazon SageMaker Unified Studio, which unified their data engineering, analytics, and artificial intelligence (AI) workflows in a single governed environment on AWS. In this post, we walk through their migration journey so you can adapt their approach to your own environment.
This post walks you through a two-layer, defense-in-depth authorization pattern for granular, intra-tenant access control in RAG applications. Defense in depth is a security strategy that uses multiple independent layers of protection. Each layer operates independently. If one layer is misconfigured, the other layer still enforces access control. The pattern runs on Amazon Bedrock, a fully managed service that offers a choice of high-performing foundation models (FMs) from Amazon and AI companies through a single API, along with a broad set of capabilities you need to build generative AI applications with security, privacy, and responsible AI.
In this post, we walk you through how to deploy ComfyUI workflows on Amazon SageMaker AI processing jobs to generate hundreds of high-quality images in a single batch. You learn how to set up the infrastructure using AWS Cloud Development Kit (AWS CDK), configure GPU-accelerated processing, and automate image generation at scale. You can then adapt this solution to your ComfyUI workflows specific to your needs. We will guide you through a practical, step-by-step process to automate ComfyUI workflows to generate hundreds of high-quality images in a single batch empowering you to scale your creative pipeline.
In this post, we walk through the problem space, our architecture on Amazon Bedrock and Amazon OpenSearch Serverless, the evaluation methodology we built on OpenStreetMap ground truth, four experiments that compared embedding models, fusion strategies, captioning, and search methods, and the practical guidance you can apply when building a similar system. You’ll learn which design choices move the needle for geospatial semantic search, including why Amazon Nova Multimodal Embeddings delivered the highest F1 scores across both benchmark queries in our evaluation. The work described here evolved into Vexcel Intelligence, a searchable imagery product.
In this post, you will learn how Ampersend built a pay-per-intelligence routing layer on top of Amazon Bedrock AgentCore Payments. AI agents autonomously route tasks to the most effective model, pay per request, and operate within spending budgets. You will also see how the two-hop payment pattern works end-to-end and how to get started with your own implementation.
In this post, we introduce new self-service capabilities for managing the full AWS Outposts lifecycle: configuration and quoting, subscription visibility, and end-of-term renewal and decommissioning. These capabilities reduce the time and coordination previously required, giving you direct control over your Outposts from evaluation through end of term. These tools are available now in all commercial AWS Regions that support AWS Outposts.