AWS launches Amazon Bedrock AgentCore Payments with Coinbase and Stripe for autonomous agent transactions
AWS announced Amazon Bedrock AgentCore Payments (preview), enabling AI agents to autonomously discover and pay for APIs, web content, MCP servers, and other agents. Built with Coinbase and Stripe, the service supports micropayments through the x402 protocol with per-session spending limits and full transaction observability.
AWS launches Amazon Bedrock AgentCore Payments with Coinbase and Stripe for autonomous agent transactions
AWS announced Amazon Bedrock AgentCore Payments (preview), the first managed payment infrastructure purpose-built for autonomous AI agents. The service enables agents to discover, access, and pay for resources including APIs, web content, MCP servers, and other agents without developers manually configuring billing relationships.
Technical implementation
AgentCore Payments is built on the x402 protocol, an HTTP-native payment standard for instant stablecoin micropayments. When an agent encounters a paid resource returning HTTP 402 "Payment Required", the platform handles wallet authentication, payment execution, and content delivery within the agent's execution loop.
Developers connect agents to either Coinbase wallets or Stripe Privy wallets through the AgentCore SDK or console. End users fund wallets via stablecoin or fiat using debit cards. The system enforces spending limits per session, with agents requiring explicit user authorization before accessing funds.
The payment manager orchestrates the complete flow:
- Protocol negotiation and retries
- Credential authentication and token lifecycle management
- Transaction routing through appropriate providers
- Real-time spending tracking against session budgets
All transactions integrate with AgentCore's existing observability infrastructure, appearing in the same logs, metrics, and traces developers use for monitoring agent behavior.
Discovery and integration
AWS is making the Coinbase x402 Bazaar MCP server available through AgentCore gateway. The bazaar provides a curated directory of x402 endpoints that agents can search and discover autonomously, eliminating the need for developers to hardcode each service integration.
According to AWS, the platform is designed to support any framework and protocol. While the preview supports x402, additional payment protocols are on the roadmap. New protocol support will be added at the platform level without requiring agent rebuilds.
Early adopters
Companies including Cox Automotive, Thomson Reuters, and PGA TOUR currently use AgentCore for building agents. Warner Bros. Discovery is evaluating the payments capability for potential agent-driven commerce experiences around premium content including live sports.
Heurist AI is building a financial research agent using AgentCore Payments to perform analysis and provide investment advice, with the agent autonomously paying for real-time market data feeds and paywalled publications.
What this means
This represents the first integrated payment infrastructure for autonomous agents at the cloud platform level. The significance is in managed orchestration: developers no longer need to build custom billing relationships, handle credential management, or write payment logic across fragmented services. The per-session spending limits and mandatory user authorization address the core risk of autonomous systems with financial access. However, the preview only supports x402 and stablecoin micropayments under $1. Broader adoption depends on protocol ecosystem maturity and whether merchants implement x402 endpoints at scale. The Coinbase bazaar directory is a forcing function for merchant adoption, but the agent economy remains nascent. Pricing for AgentCore Payments was not disclosed.
Related Articles
AWS Adds OS-Level Control to Bedrock AgentCore Browser for Native UI Automation
AWS announced OS Level Actions for Amazon Bedrock AgentCore Browser, extending agent automation beyond the browser's Document Object Model to interact with native operating system UI. The feature enables agents to control system dialogs, security prompts, and context menus through direct mouse and keyboard commands at the OS level.
AWS Launches AgentCore Optimization: Automated Performance Loop for Production AI Agents
Amazon Web Services released AgentCore Optimization in preview, introducing an automated performance loop that generates configuration recommendations from production traces, validates them through batch evaluation and A/B testing, and enables continuous agent optimization. The system targets the quality drift problem where AI agents degrade as models evolve and user behavior shifts.
AWS launches agent-guided workflows in SageMaker AI to automate model fine-tuning
Amazon Web Services has released agent-guided workflows in SageMaker AI that use AI coding agents to automate model customization. The feature includes nine pre-built skills covering use case definition, data preparation, fine-tuning technique selection (SFT, DPO, RLVR), evaluation, and deployment to Amazon Bedrock or SageMaker endpoints.
AWS SageMaker adds automatic instance fallback to prevent GPU capacity failures
Amazon SageMaker AI now supports capacity-aware instance pools that automatically try alternative GPU instance types when primary choices lack capacity. The feature works across endpoint creation, autoscaling, and scale-in operations, eliminating the manual retry cycles that previously left endpoints stuck in failed states.
Comments
Loading...