---
title: "Agent2Agent (A2A) Joins the Agentic AI Foundation, Putting It Under the Same Neutral Governance as MCP"
url: https://xcube.enlightcorp.com.tw/en/news/a2a-joins-agentic-ai-foundation-growth-stage
category: "Agent Protocols / Standards"
source_type: official
published: 2026-08-27
updated: 2026-08-27
lang: en
---

# Agent2Agent (A2A) Joins the Agentic AI Foundation, Putting It Under the Same Neutral Governance as MCP

Published 2026-08-27 · Updated 2026-08-27 · Official release · Source: [A2A Protocol Blog](https://a2a-protocol.org/latest/blog/2026/08/27/a-new-chapter-for-a2a-joining-the-agentic-ai-foundation/)

**Key answer:** Agent2Agent (A2A) is an open protocol that lets AI agents built on different frameworks and by different vendors discover each other, delegate tasks, and collaborate. Its acceptance into the Linux Foundation-directed Agentic AI Foundation (AAIF) in August 2026 means the agent-to-tool layer (MCP) and the agent-to-agent layer (A2A) are now governed by the same neutral body.

On August 27, 2026, the A2A project blog announced that the Agent2Agent (A2A) protocol has been accepted as a Growth Stage project at the Agentic AI Foundation (AAIF), the Linux Foundation-directed foundation that already hosts the Model Context Protocol (MCP), goose, and AGENTS.md.

The announcement frames the division of labor explicitly: MCP is the vertical integration layer that connects agents to internal tools and databases, while A2A is the horizontal layer for peer-to-peer collaboration. An agent publishes a structured agent card describing its capabilities and contact methods, and other agents can read it, negotiate modalities, and delegate tasks without custom integration code for every framework pairing.

According to the project, A2A has been backed by more than 150 organizations since its stable v1.0 specification, with native support in Google Cloud, AWS Bedrock AgentCore Runtime, and Microsoft Azure AI Foundry, adoption by enterprise SaaS platforms including ServiceNow, Salesforce, Atlassian, and SAP, and support in frameworks such as LangGraph, CrewAI, Pydantic AI, AG2, and IBM BeeAI. These figures are the project's own claims rather than independent measurements.

This is a governance move, not a specification change: the post announces no new version or breaking changes. For enterprises the value is that A2A's roadmap is no longer controlled by a single vendor, which lowers the risk of betting on the wrong standard. Growth Stage status, however, also signals that cross-vendor implementation consistency and the security model are still maturing.

From a deployment standpoint, A2A addresses cross-boundary handoff. The project notes that before A2A, handing work between agents built on different frameworks required custom integration code for every pairing, while Google Cloud, AWS Bedrock AgentCore Runtime, and Microsoft Azure AI Foundry have now built A2A support directly into their platforms. For enterprises using multiple clouds or SaaS products, internal agents may increasingly delegate tasks to agents on external platforms, and identity verification, permission scope, and data-egress control at that boundary are harder to govern than tool calls inside a single platform.

What to watch next is whether A2A and MCP converge on shared identity, authorization, and discovery mechanisms inside AAIF, and whether the major clouds' A2A implementations actually interoperate rather than merely claiming support.

## X Cube view

Enterprises building multiple agents on a private AI platform should treat agent discovery and delegated authority as a governance question: inventory which agents will collaborate across systems, decide whether to describe them with A2A agent cards, and make delegation events auditable instead of letting each team wire integrations ad hoc.

Tags: Agentic AI, A2A, MCP, Interoperability, Governance, Linux Foundation, Agent2Agent (A2A), Agentic AI Foundation (AAIF), Model Context Protocol (MCP), goose, AGENTS.md