Major Tech Firms Build Interoperable 'Agent Web'
An emerging “agent web” is being built by companies including Stripe, Coinbase, Cloudflare, and OpenAI, creating a cross-company ecosystem where AI agents can interoperate via open protocols. This development aims to shift engineering and business workflows from being ticket-based to outcome-based. The trend suggests a future where autonomous agents can collaborate across organizational boundaries to complete complex tasks.
- The convergence of infrastructure for AI agents occurred rapidly, with Coinbase launching "Agentic Wallets," Cloudflare releasing "Markdown for Agents," and OpenAI detailing "Skills" and "Shell" for agent execution all within hours of each other. This parallel development highlights a widespread move toward creating an economy for autonomous software. - A key enabler for this new ecosystem is the emergence of open protocols for agent communication and interoperability, such as Agent2Agent (A2A), Agent Network Protocol (ANP), and Model Context Protocol (MCP). These standards provide a common language for AI agents to discover each other, negotiate, and collaborate across different platforms and companies. - In the fintech space, Stripe's x402 payment protocol, launched on Coinbase's Base blockchain, allows AI agents to autonomously handle payments and manage assets using USDC. This infrastructure is designed for the high-volume, low-latency microtransactions expected in an agent-driven economy. - For SRE and DevOps, this agent-based approach supports a shift from reactive, ticket-based work to proactive, outcome-driven operations. AI agents can autonomously handle tasks like incident response, CI/CD optimization, and infrastructure provisioning, allowing engineers to focus on higher-level system architecture and reliability. - This transition to an "agentic" model is poised to significantly impact engineering efficiency by automating complex, multi-step workflows that traditionally require extensive human coordination. Success is measured less by standardized process metrics and more by engineering outcomes like faster project delivery and more explored design alternatives. - Projects like Fetch.ai and Olas are building decentralized networks where autonomous agents can interact, transact, and collaborate. Olas, for example, has created a "Mech Marketplace" where AI agents can hire other agents to perform tasks, fostering a self-sustaining digital economy. - The security and identity of these autonomous agents are being addressed through technologies like Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs), creating a foundation of trust for inter-agent transactions and collaborations.
