Anthropic Donates MCP to Linux Foundation Agentic AI Foundation
Summary
Anthropic donated the Model Context Protocol to the Linux Foundation, which simultaneously announced the formation of the Agentic AI Foundation (AAIF) to govern MCP and related agent standards. Anthropic and OpenAI co-founded the AAIF alongside Block; AWS, Google, Microsoft, Cloudflare, and Bloomberg joined as platinum sponsors. By this point, the MCP ecosystem had grown to over 10,000 public servers and 68 million monthly SDK downloads.
What Happened
Less than thirteen months after MCP's initial open-source release, Anthropic transferred stewardship of the protocol to the newly created Agentic AI Foundation under the Linux Foundation's umbrella. The move ended Anthropic's unilateral control over the specification and placed it under a multi-stakeholder governance structure with formal voting rights, working groups, and a standardization process.
The co-founding membership was notable for including OpenAI — a competitor that had adopted MCP in March 2025 and was now co-governing the standard it had initially sought to compete with. Platinum sponsors brought in the major cloud providers: AWS, Google, and Microsoft. Cloudflare's inclusion reflected MCP's growing importance in edge-deployed agent architectures. Bloomberg's presence signaled financial services adoption.
The AAIF announced three working groups at launch: Protocol Evolution (managing the MCP specification), Security (addressing prompt injection and authentication standards), and Ecosystem (server certification and compatibility testing). The 10,000+ public MCP server figure and 68 million monthly SDK download figure were cited as evidence that standardization, rather than proprietary lock-in, had produced the larger ecosystem.
Why It Matters
The AAIF donation marked the end of MCP's period as a single-company standard. The governance structure created conditions for long-term stability: the protocol could evolve through consensus, security vulnerabilities could be addressed across the ecosystem simultaneously, and no single vendor could unilaterally break compatibility. For enterprise buyers, Linux Foundation stewardship provided the institutional durability that justified building long-term infrastructure on MCP.
The co-founding by OpenAI was symbolically significant. It acknowledged that the agent tooling layer had become too important to be governed by any one lab, regardless of who had originated the standard. The AAIF also formalized the security work that had been ad hoc until then — the Security working group's focus on prompt injection standards acknowledged one of the most serious unresolved risks in deployed agent systems.