
Cisco Systems Inc. on Tuesday said it is donating the AGNTCY project, an open source infrastructure for collaboration and observability among artificial intelligence (AI) agents from different vendors and frameworks it debuted in March, to the Linux Foundation.
Cisco, Dell Technologies Inc., Alphabet Inc.’s Google Cloud, Oracle Corp., and IBM Corp.’s Red Hat Inc. will also be formative members of the project, joining more than 60 other vendors in helping build industry-standard protocols and frameworks that let AI agents find one another, communicate, collaborate and be managed across platforms, models, and organizations, according to the Linux Foundation and Cisco.
“The AGNTCY project lays groundwork for secure, interoperable collaboration among autonomous agents,” Jim Zemlin, executive director of the Linux Foundation, said in a statement. In June, the foundation launched Agent2Agent (A2A), a project designed around an open protocol created by Google for secure agent-to-agent communication and collaboration. The project boasts support from more than 100 tech companies.
“Building the foundational infrastructure for the Internet of Agents requires community ownership, not vendor control,” Vijoy Pandey, general manager and senior vice president of Outshift by Cisco, said in an interview. “The Linux Foundation ensures this critical infrastructure remains neutral and accessible to everyone building multi-agent systems.”
AGNTCY was initially open sourced by Cisco in March, with collaboration from LangChain and Galileo. It is intended to provide foundational infrastructure for the “Internet of Agents,” a new collaboration layer that lets multi-agent systems work together regardless of who built them or where they run.
A proliferation of AI agents from the likes of Cisco, Google, Salesforce Inc., Microsoft Corp., Amazon.com Inc., ServiceNow Inc., Adobe Inc., and countless others have contributed to fragmentation and vendor silos, impeding the ability for agents to securely communicate, share context and collaborate across platforms.
This has increasingly alarmed IT managers tasked with mixing and matching AI agents within vast enterprise systems while under pressure from free-spending upper management to make everything work and become more efficient. Spending on agentic AI could reach $155 billion by 2030, according to a Bank of America Global Research report last month.
AGNTCY’s open, common infrastructure provides developers and organizations with secure agent identity, reliable messaging and end-to-end observability to enhance transparency, performance, efficiency and trust, said Pandey.
Additionally, the AGNTCY project is interoperable with A2A, Anthropic’s Model Context Protocol (MCP), and other leading AI agent technologies.