For two years, every major lab pushed its own standard for AI agents to talk to each other. On August 20, 2026 that fragmentation actually started closing: Google put its A2A protocol under the same neutral foundation that already governs MCP.
The Agentic AI Foundation (AAIF), run by the Linux Foundation, announced that Google's A2A (Agent-to-Agent) protocol formally moved under its governance, sitting alongside Anthropic's Model Context Protocol (MCP) under the same neutral umbrella. They're complementary protocols, not competitors: MCP standardizes how an agent connects to tools and data sources; A2A standardizes how different agents communicate with each other. Together they form the basic infrastructure for large-scale multi-agent systems.
AAIF grew from 49 to over 250 members in under a year, including Platinum signatories like AWS, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, and OpenAI. Direct competitors in the model and cloud market sitting at the same protocol-governance table is the real signal: nobody wants to be the one left outside the common standard.
If your company is evaluating building a multi-agent system (several specialized agents coordinating — one that researches, one that drafts, one that verifies), the question of "which protocol do I use" gets much simpler: MCP so your agents talk to your tools and data, A2A so your agents talk to each other, both governed by the same neutral foundation instead of whichever vendor happens to be in charge. That directly reduces the risk of your architecture being orphaned if a lab changes strategy.
Carlos Montiel is an enterprise AI solutions architect. He implements LLMs, Agents, RAG, and orchestrators for companies across Guatemala and Latin America. Reach out for a consultation.
Contact Carlos Montiel