OpenAI said it is co-founding the Agentic AI Foundation (AAIF) under the Linux Foundation with Anthropic and Block, with support from Google, Microsoft, AWS, Bloomberg and Cloudflare, in a move to establish open standards for agentic AI systems as they move from experimentation into real-world use.
The foundation will provide neutral governance for open, interoperable infrastructure designed to prevent agent ecosystems from fragmenting into incompatible silos. As part of the effort, OpenAI is donating AGENTS.md, an open Markdown-based format that gives AI agents project-specific instructions, alongside contributions from Anthropic and Block.
Agentic systems are increasingly being used for coding, workflow automation and customer service, making portability, safety and predictability more critical. Placing AAIF under the Linux Foundation mirrors how widely adopted technologies such as Kubernetes and PyTorch are governed.
The initiative signals industry alignment around shared standards and reflects a broader shift toward agentic AI as foundational infrastructure rather than experimental tools.