On April 9, 2026, the Agent2Agent (A2A) protocol marked its first anniversary, and the Linux Foundation published the scorecard the same day: supporting organizations have grown from 50-plus a year ago to more than 150, the core repository has passed 22,000 GitHub stars, and the SDK lineup has expanded from a single Python implementation to five languages. A standard that started inside Google and now lives under Linux Foundation governance is becoming the default answer for agent-to-agent interop across organizational boundaries.
For developers, the birthday is not the point — the stable spec is. Version 1.0 shipped in March 2026 as the first production-ready release, with a defined migration path. The standards fight is consolidating faster than most people expected.
Year One: From 50 to 150+
Google announced A2A on April 9, 2025, with the goal of giving agents on different frameworks a common language. Two months later, on June 23, 2025, it was donated to the Linux Foundation at Open Source Summit North America in Denver, making it a vendor-neutral, community-governed standard. Founding members included AWS, Cisco, Microsoft, Salesforce, SAP, and ServiceNow.
A year on, the roster has grown past 150 organizations (IBM and others joined along the way), with production use concentrated in supply chain, financial services, insurance, and IT operations. The Linux Foundation calls it the first production-ready open standard for global agent-to-agent communication. The developer-facing change is more concrete: the SDKs went from one Python implementation to five production-ready languages — Python, JavaScript, Java, Go, and .NET — which means most teams can now consume the protocol natively instead of wrapping the reference implementation themselves.
v1.0 and Signed Agent Cards
Version 1.0, finalized in March 2026, brought multi-protocol support, enterprise-grade multi-tenancy, modernized security flows, and a migration path. The headline feature is Signed Agent Cards: cryptographic identity verification, so that before one organization delegates a task to another organization’s agent, it can confirm who that agent actually is.
The architecture also leans on web conventions — familiar load-balancing and security patterns — in exchange for reliability at scale. Google’s open source blog showed interoperability demos across ADK, LangGraph, AG2, and CrewAI: agents on different platforms delegate sub-tasks to each other without sharing internal memory.
Division of Labor with MCP, and the AP2 Payments Extension
The official positioning fits in one sentence: MCP connects agents to internal tools and data; A2A handles agent-to-agent coordination across organizational boundaries. The two protocols are complements, not competitors.
An “A2Family” is also growing on top of A2A’s open extensibility model. AP2, the Agent Payments Protocol, lets agents complete transactions securely and is backed by more than 60 payments and financial-services organizations. UCP is compatible with it through the AP2 mandates extension, which captures cryptographic evidence of user purchase consent. A2UI covers the interface layer between agents and humans.
Cloud Deployments and the Roadmap
On the deployment side, A2A is natively supported in Microsoft’s Azure AI Foundry and Copilot Studio, and in AWS’s Amazon Bedrock AgentCore Runtime. Rao Surapaneni, VP and GM at Google Cloud, put the argument plainly: AI agents are only as useful as their ability to collaborate. Luca Muscariello, Distinguished Engineer at Cisco, called A2A “the syntactic layer that makes agent-to-agent communication reliable and interoperable.”
Four items remain on the roadmap: an interoperability specification, registry consolidation, expanded testing and tooling, and security and deployment best practices. For learning, DeepLearning.AI published a short course on A2A, alongside official tutorials and samples.
Why Developers Should Care
Two takeaways. First, the agent interoperability standards fight is largely settled: an A2A-plus-MCP dual-protocol layout, under neutral Linux Foundation governance, is now the consensus among the mainstream frameworks and the large clouds. New projects can assume it exists instead of inventing their own wire protocol, and framework choice no longer locks you out of cross-organization workflows. Second, the two hardest trust problems in cross-organization agent workflows — identity and payment — are being standardized via Signed Agent Cards and AP2 respectively. That is precisely the kind of building block the next phase of enterprise AI depends on, and it arrives before most teams have had to solve those problems badly on their own.
Sources
- A2A Protocol Surpasses 150 Organizations — Linux Foundation / PR Newswire
- A year of open collaboration: celebrating the anniversary of A2A — Google Open Source Blog
- A2A Protocol Surpasses 150 Organizations — The Linux Foundation
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.