Linux Foundation forms Agentic AI Foundation: open standards

The Linux Foundation announced the Agentic AI Foundation (AAIF) on December 23, 2025, to develop open standards and best practices for agentic AI. InfoQ and AFP reported the launch and early membership, describing the AAIF as a neutral home for donated protocols and governance work. (InfoQ, AFP)

InfoQ says OpenAI and Anthropic donated AGENTS.md and Anthropic's Model Context Protocol (MCP) to AAIF. AFP reported that Liquid Reply joined as an early participant. The foundation will accept protocol donations and convene technical working groups to improve agent-to-tool interoperability and trust practices. (InfoQ; AFP)

For implementers, AAIF could centralize standards work now handled by custom integrations between agents and APIs, reducing duplicate effort across vendors. Next steps likely include formal governance rules, published technical specifications, and invitations to additional vendors and open-source projects to donate protocols and join working groups.

Why It Matters

• Reduce integration effort: standardized agent protocols can cut custom connector work when linking agents to tools, APIs, and databases.
• Lower vendor lock-in and improve auditability: a neutral home for donated specs (AGENTS.md, MCP) makes multi-vendor agent deployments easier to evaluate and govern.
• Prepare architecture and procurement: IT teams should track AAIF specs and working-group RFCs to anticipate changes to agent runtimes, connector patterns, and security controls.