AWS launches comprehensive platform for building enterprise AI agents at scale

Amazon Web Services has unveiled Bedrock AgentCore, a comprehensive platform for building and deploying enterprise AI agents at scale. The new offering, announced at AWS’s New York Summit, represents a significant expansion beyond the company’s existing Bedrock Agents framework by supporting any agent framework or foundation model—not just those hosted within Bedrock—positioning AWS to compete directly with OpenAI’s Agents SDK and Google’s Gemini-based platforms in the rapidly growing enterprise AI agent market.

What you should know: AgentCore is a modular stack of services designed to move AI agents from prototype to production environments with enterprise-grade security and scalability.

• The platform includes seven core modules: Runtime (serverless execution), Memory (long- and short-term learning), Identity (OAuth-based access management), Observability (debugging and telemetry), Gateway (API integration), Browser (headless web interaction), and Code Interpreter (secure code execution).
• Unlike the previous Bedrock Agents framework, AgentCore supports open-source toolkits like CrewAI, LangChain, LlamaIndex, LangGraph, and AWS’s own Strands Agents SDK.
• The platform integrates with AWS Marketplace, enabling teams to discover and deploy pre-built agents and tools.

Why this matters: AWS is positioning itself at the center of what executives describe as a “tectonic change” in software development, where agents represent the shift to “service as a software.”

• “Agents are the most impactful change we’ve seen in ages,” said Swami Sivasubramanian, AWS Vice President of Agentic AI. “With agents comes a shift to service as a software. This is a tectonic change in how software is built, deployed and operated.”
• The platform supports emerging industry standards like Model Context Protocol (MCP) and Google’s Agent-2-Agent (A2A) protocol, emphasizing interoperability across different systems.

Early customer adoption: Several companies with early access are already building production-grade agentic applications across finance, healthcare, marketing, and content management.

• Box, a cloud document and file storage company, is exploring ways to extend its content management tools using Strands Agents and Bedrock AgentCore Runtime, with CTO Ben Kus citing “top tier security and compliance” benefits.
• Brazil’s Itaú Unibanco is using AgentCore to develop hyper-personalized, secure digital banking experiences.
• Healthcare company Innovaccer has built a new Healthcare Model Context Protocol (HMCP) on top of AgentCore Gateway, with CEO Abhinav Shashank calling Gateway a “game-changer” for converting existing APIs into agent-compatible tools at scale.

Competitive landscape: AWS’s launch puts it in direct competition with other major tech companies building end-to-end agent development platforms.

• OpenAI’s Agents SDK and Google’s Gemini-based Agents SDK are pushing similar visions of comprehensive agent development environments.
• Writer’s AI HQ and startups like Cognition (maker of Devin) are also building tools for managing autonomous software agents.
• “The line between an agent and an application is getting blurrier,” said Deepak Singh, AWS Vice President of Databases and AI.

Pricing and availability: AgentCore is now available in preview in select AWS regions including US East (N. Virginia), US West (Oregon), Asia Pacific (Sydney), and Europe (Frankfurt).

• The platform is free to try until September 16, 2025, with consumption-based pricing beginning thereafter.
• Each module is billed independently: Runtime, Browser, and Code Interpreter cost $0.0895 per vCPU-hour and $0.00945 per GB-hour.
• Gateway charges $0.005 per 1,000 tool API invocations, while Memory costs range from $0.25 per 1,000 short-term memory events to $0.75 per 1,000 long-term memories stored.
• Identity costs $0.010 per 1,000 token or API key requests, though it’s included at no extra charge when used via Runtime or Gateway.