OpenAI, Anthropic and Block form Agentic AI Foundation to set shared industry standards

OpenAI, Anthropic and Block have announced the formation of the Agentic AI Foundation (AAIF), a new open-source body created under the Linux Foundation to develop interoperability standards for agentic artificial intelligence. The initiative is backed by Google, Microsoft, AWS, Bloomberg and Cloudflare, indicating broad industry alignment as agentic systems begin moving into real-world deployment.

As part of the launch, the three founding companies have transferred key technologies to the new foundation. Anthropic’s Model Context Protocol, which enables agents to connect and interact; OpenAI’s AGENTS.md format, which provides project-specific instructions for coding agents; and Block’s Goose framework for building agentic tools will now be governed through an open, community-led process. All three technologies were already freely available, but the shift to a neutral body allows broader developer participation in their evolution.

AGENTS.md, introduced in 2025, has been adopted across more than 60,000 open-source projects and frameworks. Its inclusion within AAIF is aimed at ensuring long-term, vendor-neutral stewardship as agentic AI progresses from experimentation to large-scale implementation.

Under the oversight of the Linux Foundation, AAIF will maintain and advance shared protocols to prevent ecosystem fragmentation and support the safe, scalable deployment of agentic AI across industries.