Google's Agent2Agent protocol is becoming the basis for a broader Linux Foundation project focused on open communication between AI agents. The move brings several major technology companies into a shared effort to reduce fragmentation and create a vendor-neutral path for agent cooperation.
What A2A is meant to solve
A2A, short for Agent2Agent, was originally developed by Google. Its purpose is to give AI agents from different providers a common way to recognize each other, share information, and coordinate tasks.
That matters because the source describes the current landscape as one with few technical standards for this kind of collaboration. Without a common approach, AI agent systems could develop as separate proprietary environments that do not work cleanly together.
The Linux Foundation project is designed to push in the opposite direction. By placing A2A in an open project, the companies involved are trying to create a shared communication layer that is not controlled by one vendor.
Google hands over the protocol
Google has transferred A2A to the Linux Foundation. The transfer includes the protocol specification, development tools, and software libraries.
The stated reason is vendor-neutral development. In practical terms, that means the protocol is intended to evolve under neutral governance rather than remain solely inside Google's control.
Other companies involved in the initiative include Amazon Web Services, Cisco, Microsoft, Salesforce, SAP, and ServiceNow. According to Google, more than 100 companies already support the protocol.
AWS and Cisco are also joining as new validators. The Linux Foundation is taking over management in order to support neutral access and open participation.
The project has four immediate goals
The new Agent2Agent project is not only a handoff of existing technology. It also sets out a defined agenda for how the standard should grow.
The project has four main goals:
- Establishing an open industry standard.
- Building a broad developer community.
- Ensuring vendor-neutral governance.
- Enabling secure cooperation between agents.
These goals are closely connected. An open industry standard needs developers to adopt it. A broad developer community needs confidence that the project is not locked to one provider. Secure cooperation matters because the protocol is about agents exchanging information and coordinating activity across systems.
The group also plans to develop additional standards over time. The source names agent identity, delegated decision-making, security rules, and reputation systems as examples of future work.
Anyone interested can get involved through the project's GitHub repository. That detail is important because it frames the initiative as an open participation effort, not only a private agreement among large companies.
How A2A fits with MCP
A2A is not the only open standard emerging around AI agents. The source also points to Anthropic's Model Context Protocol, or MCP.
MCP addresses a different layer of the agent stack. It defines how a single AI agent connects data sources, tools, and memory to a language model.
A2A focuses on communication between agents. MCP focuses on context within an agent. The two protocols therefore solve different but related problems.
Taken together, the source says they could form an open, vendor-neutral infrastructure for AI agents. A2A would help agents work with one another, while MCP would help an individual agent connect to the resources it needs.
Why this handoff matters
The significance of the Linux Foundation move is governance. A communication standard for AI agents becomes more useful when many providers can trust the process behind it.
If A2A remained only a Google-led protocol, other companies might still support it, but the structure would carry a different set of expectations. Moving the specification, tools, and libraries to the Linux Foundation is meant to support neutral access and open participation.
The broader question is whether AI agents will develop around shared standards or around separate proprietary systems. The companies involved in Agent2Agent are choosing the standards path, with A2A as the foundation for communication and coordination between agents.
For developers and companies building agent-based systems, the practical takeaway is straightforward: the push for open AI agent infrastructure is becoming more organized. A2A is now positioned as a Linux Foundation project, backed by major technology companies, with a roadmap that reaches beyond basic communication into identity, decision-making, security, and reputation.