CROSS-CATEGORY: Both Agent Protocols Consolidated Governance in the Same 72 Hours, and Both Are Now Aiming at Enterprise Identity
A2A moved under the Agentic AI Foundation on Aug 20 and MCP published its roadmap on Aug 22, and the two documents converge on the same next problem: agent identity for cloud workloads, server-initiated events, and enterprise authorization rather than a human clicking through OAuth. MCP's roadmap explicitly lists 'agent identity and enterprise security' as a priority area and describes the shift away from browser-based authorization; AAIF's platinum roster (AWS, Anthropic, Cloudflare, Google, Microsoft, OpenAI) is the same set of vendors who need that to sell agents into regulated buyers. For builders this means the integration layer is being standardized out from under point-solution vendors, and the durable moat moves to whoever holds the identity and audit trail.
↳ Follow the thread