Anthropic Donates MCP to Linux Foundation—OpenAI, Google Join the New AI Standard

Anthropic announced that it is donating its Model Context Protocol (MCP) to the Linux Foundation, establishing the Agentic AI Foundation (AAIF) as a new Linux‑Foundation‑backed project. The foundation’s co‑founders are Anthropic, Block and OpenAI, with supporters including Google, Microsoft, AWS, Cloudflare and Bloomberg.

Why the donation matters

Without MCP, developers must write separate integration code for each AI model—e.g., one set for ChatGPT and another for Claude. The author likens this fragmentation to early mobile chargers with incompatible plugs. MCP aims to be the "USB‑Type‑C" of AI, providing a universal interface that lets any model connect to data sources such as Notion, GitHub or internal databases.

By moving MCP under the Linux Foundation, the protocol becomes fully neutral, no longer a proprietary asset of Anthropic, and is no longer subject to lock‑in by any single vendor. The community will maintain and evolve it, alleviating concerns that a company might later restrict access.

MCP’s growth in its first year

Over 10,000 active public MCP servers.

More than 97,000,000 monthly SDK downloads for Python and TypeScript.

Broad adoption by editors (Cursor, VS Code, Zed) and AI products (ChatGPT, Gemini, Microsoft Copilot).

These figures show that MCP has already become the de‑facto standard for AI programming and agent development.

New initiatives from AAIF

goose : a project contributed by Block.

AGENTS.md : an OpenAI project that defines agent semantics and collaboration patterns.

The inclusion of AGENTS.md signals a growing consensus on how agents should be defined and how they should cooperate.

Implications for developers and users

When industry leaders such as OpenAI, Google, and Microsoft collaborate on a common standard, the biggest beneficiaries are developers and end‑users, who gain a stable, interoperable foundation for building AI agents and integrations.