NEW YORK, NY – July 16, 2025 – Amazon Web Services (AWS) today announced the launch of Amazon Bedrock AgentCore, a significant expansion of its Amazon Bedrock service designed to empower organizations to build, deploy, and operate secure and scalable AI agents at an enterprise level. This new offering marks a pivotal moment in the evolution of generative AI, providing a comprehensive suite of services that go beyond foundational model access to enable sophisticated, multi-step AI workflows.

Amazon Bedrock, which achieved general availability in September 2023, has rapidly become a cornerstone for businesses looking to leverage generative AI without managing underlying infrastructure. With Bedrock AgentCore, AWS is deepening its commitment to enabling more complex and automated AI solutions.

AgentCore: Orchestrating Intelligent Actions

At the heart of Bedrock AgentCore is the ability to build and orchestrate AI agents capable of performing multi-step tasks across various systems and data sources. This moves beyond simple prompt-response interactions, allowing AI to take actions, integrate with business applications, and manage complex workflows. Key components and features of Bedrock AgentCore include:

• AgentCore Runtime: A secure and dynamic runtime environment engineered to handle variable workloads, supporting interactive experiences with low latency and complex asynchronous workloads running up to eight hours. It is framework-agnostic, offering complete session isolation for enhanced security.

• Enhanced Amazon Nova Customization: New capabilities for Amazon Nova models on Amazon SageMaker AI, including support for SageMaker HyperPod, enable higher accuracy and flexibility for specific business needs. The Nova Act SDK also gains additional security and access controls for production-ready agent deployment.

• Amazon S3 Vectors (Preview): A groundbreaking development, S3 Vectors introduces native vector support for AI workloads directly within cloud object storage. This promises to drastically reduce the cost of storing and querying vectors (up to 90%), making it more cost-effective to leverage large vector datasets for RAG (Retrieval Augmented Generation) and semantic search. It seamlessly integrates with Bedrock Knowledge Bases and OpenSearch Service.

• Kiro IDE: A new Integrated Development Environment (IDE) specifically designed to simplify the developer experience for building AI agents. Kiro allows developers to "vibe code" and collaborate with agents on detailed plans, automatically triggering agent workflows for tasks like testing and documentation generation.

• New MCP Resources: The Model Context Protocol (MCP) is emerging as a standard for agents to connect to data sources, tools, and memory banks. New local AWS API MCP and AWS Knowledge MCP servers provide comprehensive, up-to-date knowledge of AWS APIs and documentation, making it easier for AI assistants to become AWS experts.

• TwelveLabs AI Models in Bedrock: Expanding the diverse foundation model offerings, TwelveLabs AI models are now available in Amazon Bedrock. These models excel at transforming video libraries into searchable, actionable assets by simultaneously processing visual, audio, and text elements, enabling content discovery through natural language prompts.

Benefits for the Enterprise

Bedrock AgentCore offers significant advantages for enterprises looking to scale their AI initiatives:

• Accelerated Development: Simplifies the creation and deployment of sophisticated AI agents, allowing developers to focus on business logic rather than infrastructure management.

• Enhanced Automation: Enables automation of complex, multi-step business processes, from customer service to data analysis and content generation.

• Improved Accuracy and Relevance: By integrating with proprietary data through features like RAG and leveraging enhanced customization capabilities, agents can deliver more accurate and contextually relevant responses and actions.

• Cost Efficiency: Features like S3 Vectors significantly reduce the cost of managing and querying large datasets for AI workloads.

• Security and Compliance: Built on AWS's robust security infrastructure, ensuring data privacy, encryption, and adherence to industry compliance standards.

• Flexibility and Choice: Continues Bedrock's commitment to offering a wide array of leading foundation models from various providers, preventing vendor lock-in.

Looking Ahead

The introduction of Amazon Bedrock AgentCore underscores AWS's strategic vision for generative AI, moving towards a future where AI agents seamlessly integrate into enterprise operations, driving efficiency and innovation. This release is expected to fuel a new wave of AI-powered applications, enabling businesses to unlock unprecedented levels of automation and intelligence. As organizations increasingly seek to leverage AI for more than just content generation, Bedrock AgentCore provides the foundational capabilities to build truly transformative AI solutions.