Amazon Web Services

Persistent, managed EC2 infrastructure for production AI agents with multi-agent collaboration, GPU support, and sessions lasting up to 14 days.

First-party excerpt: Announcing runtime instances in Amazon Bedrock AgentCore—persistent, managed EC2 infrastructure for production AI agents with multi-agent collaboration, GPU support, and sessions lasting up to 14 days.

Native vector search capability in Amazon DynamoDB with single-digit millisecond latency at 99%+ recall, designed to scale to trillions of vectors with zero infrastructure management.

First-party excerpt: DynamoDB now supports native vector search with single-digit millisecond latency at 99%+ recall. It is designed for any scale, even trillions of vectors and requires zero infrastructure management.

Anthropic's Claude Opus 5 model made available through Amazon Bedrock.

First-party excerpt: Local Zone in Athens, Claude Opus 5 on AWS, Lambda durable execution for .NET

Multi-agent collaboration feature within the Kiro platform for coordinating multiple AI agents.

First-party excerpt: Web Search on Amazon Bedrock, Dogwood, Kiro Crew, and more

OpenAI's Daybreak model available through Amazon Bedrock.

First-party excerpt: OpenAI Daybreak on Bedrock, and more (August 17, 2026)

OpenAI's GPT-5.6 models integrated into Amazon Bedrock for use through AWS.

First-party excerpt: OpenAI GPT-5.6 models on Bedrock, and more (July 20, 2026)

Web search capability integrated into Amazon Bedrock for augmenting AI agent responses with real-time information from the web.

First-party excerpt: Web Search on Amazon Bedrock, Dogwood, Kiro Crew

Native vector search capability in DynamoDB with single-digit millisecond latency at 99%+ recall, supporting trillions of vectors at any scale with zero infrastructure management.

Amazon WorkSpaces capability extended to support AI agent deployment and execution.

Anthropic's Claude Sonnet 5 model available through Amazon Bedrock for inference and application development.

Accelerated infrastructure deployment mode enabling AI agents and developers to receive deployment confirmation in seconds, up to 4x faster than standard mode.

Latest version of Grok LLM model available through Amazon Bedrock for enterprise AI applications.

Fully managed RAG pipeline service with native data connectors, smart parsing for multi-format data preparation, and agentic retriever for complex multi-step queries integrated with AgentCore Gateway.

Enhanced security agent with STRIDE-based threat modeling, full repo and PR code scanning capabilities powered by Kiro and Claude integration.

Fully managed tool enabling AI agents to ground responses in current, cited web knowledge with zero data egress, integrated with AgentCore for building agents without manual web search integration.

Anthropic's Claude Fable 5 model delivering Mythos-level capabilities with built-in safeguards. Available on Amazon Bedrock with strong safety guardrails for broader enterprise use.

New console experience optimized for Anthropic and OpenAI-compatible APIs, enabling model comparison, project-based organization with evaluation workflows, and auto-prefilled code snippets.

Latest version of Anthropic's Claude Opus model available on Amazon Bedrock for enterprise deployments.

A rebuilt Amazon OpenSearch Serverless platform designed for agentic AI and dynamic workloads, featuring instant autoscaling and up to 60% cost savings.

A rebuilt AWS Resilience Hub featuring generative AI-powered failure mode analysis, dependency discovery assessment, modular resilience policies, and organization-wide reporting for SRE resilience journeys.

Multi-model platform (Claude, GPT-5.4 limited preview, Llama, Mistral, Titan). Now serves OpenAI post Microsoft amendment.

A tool that enables customers to optimize prompts for their current model or migrate prompts to new models with built-in evaluation feedback loops, supporting simultaneous comparison of up to 5 models.

A managed remote Model Context Protocol (MCP) server that gives AI agents and coding assistants secure, authenticated access to all AWS services, part of the Agent Toolkit for AWS.

An AI assistant for work with a desktop app and expanded integrations across enterprise workflows. Launched at What's Next with AWS 2026.

Anthropic's Claude Opus 4.7 model available in Amazon Bedrock, featuring improved agentic coding capabilities and a 1M token context window, powered by Bedrock's next-generation inference engine.

AI-powered designer for building agentic customer experience solutions within Amazon Connect, enabling automation of contact center workflows with agent capabilities.

Core framework for building and managing agentic AI applications on AWS.

GPU-accelerated EC2 instances powered by NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs, designed for AI inference, graphics, and data analytics workloads.

Q Developer (coding), Q Business (workflows), Q for QuickSight (analytics). Enterprise-friendly assistants across AWS.

New serverless compute primitive providing VM-level isolated sandboxes with no shared kernel or resources, rapid launch/resume, full lifecycle control, and state preservation up to 8 hours.

An extension of AWS Transform that enables organizations to modernize and transform code at scale using AWS-managed and custom transformations, including language version upgrades and migrations.

Anthropic's Claude Opus 5 model now available through Amazon Bedrock for deployment and inference.

Grok 4.3 large language model now available through Amazon Bedrock for AI inference workloads.

AI capability framework for building applications, mentioned in AWS Summit New York announcements.

Agent-based AI capability available through AWS, focused on automating enterprise workflows.

Cowork-class desktop agent with local file access, dashboards, slides, and connected apps. April 2026 launch.

Q1 26: AWS +28% growth (fastest in 4 years), $152B ARR. Jassy: AWS could double to $600B by 2036.

Trainium ASICs for training; Graviton 5 CPUs optimized for agentic workloads (Meta multi-billion deal April 26).

OpenAI's frontier models GPT-5.5 and GPT-5.4, plus Codex coding agent, available on Amazon Bedrock with built-in security, governance, and pay-per-token pricing.

Managed payment capabilities for AI agents enabling them to autonomously access and pay for APIs, MCP servers, web content, and other agents, built in partnership with Coinbase and Stripe.

Agent capability to assess code changes before production, reviewing release readiness and running autonomous release testing to safely ship code.

AI-powered service that automatically scans code repositories to detect, prioritize, and remediate technical debt at scale.

OpenAI models including GPT-5.5, Codex, and Managed Agents made available on Amazon Bedrock through an expanded AWS-OpenAI partnership, currently in limited preview.

Four agentic AI solutions for supply chain, hiring, customer experience, and healthcare. Announced as part of expanded Connect capabilities.