Picture the early 1980s, when the internet's plumbing was still a political question: would one company's proprietary stack become the connective tissue of the world, or would a neutral, open standard win? The answer, once IETF codified TCP/IP, shaped everything that followed. Something structurally similar happened this week, and most developers missed it. Google's Agent2Agent Protocol, a standard for AI agents to talk to one another, is moving to the Agentic AI Foundation, joining Anthropic's Model Context Protocol under the same vendor-neutral roof. The Axios scoop landed August 17. By the next morning, the infrastructure community was quietly recalibrating.
To understand why this matters, you need the two-layer picture. MCP handles how an agent connects to tools and data sources; A2A handles how agents communicate with each other across organizational and platform boundaries. Think of MCP as the wiring that gives a single AI agent its hands, and A2A as the postal system that lets millions of those agents address and route work to one another. According to Google's official documentation, A2A defines how AI agents discover each other's capabilities, negotiate task delegation, exchange structured data in JSON format, and maintain conversation context across multiple turns; the protocol operates over HTTP with JSON-RPC 2.0, making it compatible with existing web infrastructure. Before A2A existed, engineers building multi-agent systems faced a brutal combinatorial problem: agents built on different frameworks couldn't hand off work to each other without custom integration code written specifically for each pairing, and every new vendor relationship required the same integration work from scratch; the cost wasn't in the agents themselves but in the integrations between them.
The governance move is the story. A2A will become a hosted project of the Agentic AI Foundation, moving from the Linux Foundation's broader portfolio into its foundation focused specifically on agentic AI. The protocol joins MCP, as well as projects such as Block's Goose and OpenAI's AGENTS.md. The AAIF operates under a two-tier governance model: a Governing Board responsible for strategy, budget, and membership policy, and a Technical Committee that approves projects and conducts technical reviews; the Linux Foundation provides the legal and operational structure without dictating the technical direction itself. In practical terms, that means no single company can unilaterally change the spec, deprecate features to advantage their cloud, or quietly fork it into a proprietary variant. Anthropic can't lock down MCP; Google can't gatekeep A2A; enterprise adoption accelerates when governance is neutral, and companies that wouldn't bet on a competitor's proprietary protocol will bet on a Linux Foundation standard. The numbers bear this out: from its launch in December 2025, AAIF says it has grown from fewer than 40 members to more than 250, with key backers including Google, Microsoft, Amazon, Anthropic, OpenAI, Bloomberg, Shopify, and Block.
The A2A protocol itself has moved fast for a standard. Google launched it in April 2025 with fifty-plus partners, donated it to the Linux Foundation in June 2025, and then watched IBM's Agent Communication Protocol merge into A2A in August 2025, showing that the field was moving toward a single shared standard rather than competing ones. A2A v1.0, the first stable specification, shipped in March 2026; it added multi-protocol bindings and version negotiation, multi-tenancy, and signed agent cards for cryptographic identity verification. Real deployments followed: A2A is already running in production across mobile platforms, cloud AI infrastructure, financial services, and supply chain; Huawei has standardized A2A as the protocol between Celia, its OS-level AI assistant, and in-app agents across the HarmonyOS developer platform. Google Cloud VP Rao Surapaneni told Axios: "When we first envisioned A2A, the hypothesis was customers are deploying agentic systems from multiple technology providers and platform providers."
Skeptics exist, and their objections are not trivial. A2A arrived before most teams had even solved basic tool access, prompt injection, observability, cost control, and security for single-agent systems; in that environment, agent-to-agent interoperability sounded ambitious but also premature, and many AI agent demos in 2025 did not need A2A at all; they needed better prompts, better tools, better permissions, better retry logic, and better logs. Mahesh Shanmugasundaram, Seekr's lead AI solutions architect, believes A2A brings with it the risk that unverified claims will travel and gain apparent authority through an agent chain. A common envelope still leaves hard questions about authentication, authorization, capability discovery, task ownership, cancellation, data classification, and audit evidence. And there is the institutional realpolitik: whether 250-plus members translates to active contributors or mostly badge-collecting remains unresolved, and whether the foundation model works for AI agent standards, where the pace of change is faster, the competitive dynamics more intense, and the technology less mature than Linux or Kubernetes, is the open question that 2026 will begin to answer. Still, the consolidation signal is hard to dismiss; the fact that Anthropic's protocol and Google's protocol now share a technical committee and overlapping membership is, structurally, unprecedented in AI infrastructure history.
The question is no longer whether the multi-agent web gets open standards; it is whether the companies now co-governing those standards will hold the line when their own cloud margins are what is actually at stake.