OpenAI, Anthropic & Block Join Linux Foundation AI Agent Effort

The Rise of Agentic AI and the Need for Open Standards

As artificial intelligence evolves beyond simple chatbots and begins to encompass systems capable of independent action, the Linux Foundation is establishing a new organization. This initiative aims to prevent the fragmentation of AI agents into a collection of incompatible and proprietary products.

Introducing the Agentic AI Foundation (AAIF)

The newly formed group, known as the Agentic AI Foundation (AAIF), will serve as a neutral hub for open-source projects centered around AI agents. Initial funding for the AAIF is being provided by prominent companies including Anthropic, Block, and OpenAI.

Key Contributions from Founding Members

Anthropic is contributing its Model Context Protocol (MCP), a standardized method for connecting AI models and agents to essential tools and data sources.

Block is donating Goose, its open-source agent framework, to the foundation.

OpenAI is contributing AGENTS.md, a straightforward instruction file that developers can integrate into repositories to guide the behavior of AI coding tools.

These contributions represent the foundational elements for the emerging agent era.

Industry-Wide Support and Shared Guardrails

Additional members of the AAIF include AWS, Bloomberg, Cloudflare, and Google. This broad participation demonstrates a collective industry effort to establish shared safety measures, ensuring AI agents are reliable and trustworthy at scale.

The Importance of Shared Protocols

According to OpenAI engineer Nick Cooper, protocols function as a common language, enabling diverse agents and systems to collaborate effectively without requiring developers to repeatedly create integrations from scratch.

Cooper emphasized that the need for multiple protocols to facilitate negotiation, communication, and value delivery ensures that no single provider will dominate the field.

Avoiding Proprietary "Closed Wall" Stacks

Jim Zemlin, executive director of the Linux Foundation, articulated the goal more directly: to avoid a future characterized by proprietary stacks where tool connections, agent behavior, and orchestration are restricted to a limited number of platforms.

By consolidating these projects under the AAIF, the foundation aims to coordinate interoperability, safety protocols, and best practices specifically for AI agents.

Block's Openness Play with Goose

Block, the fintech company behind Square and Cash App, is contributing Goose as a demonstration that open-source alternatives can rival proprietary agents in performance and scalability. Thousands of engineers currently utilize Goose weekly for coding, data analysis, and documentation.

Dual Benefits of Open Sourcing Goose

Open sourcing Goose serves a dual purpose for Block. It allows for community contributions to enhance the framework, and it provides access to community-driven stress testing.

Furthermore, it positions Goose as a practical example of the AAIF’s vision – an agent framework designed to integrate with shared building blocks like MCP and AGENTS.md.

Anthropic's Contribution of MCP

Similarly, Anthropic is contributing MCP to the Linux Foundation with the objective of establishing it as the standard infrastructure for connecting AI models to tools, data, and applications, eliminating the need for numerous custom adapters.

The goal is widespread adoption, making MCP the de facto standard for integration.

Governance and Control

The AAIF’s structure is funded through membership dues, but Zemlin stresses that funding does not equate to control. Project roadmaps are determined by technical steering committees, ensuring no single member has unilateral authority.

Measuring Success and Future Evolution

The success of the AAIF will be measured by the adoption of these standards and the implementation of shared standards by vendors globally.

Cooper envisions a continuous evolution of these standards, with ongoing input and adaptation to ensure they remain relevant and effective.

The Potential for a Dominant Implementation

There is a possibility that one company’s implementation could become dominant due to speed of deployment or widespread usage. However, Zemlin points to the history of open source, citing Kubernetes as an example of a project that achieved dominance through merit rather than vendor control.

Benefits for Developers and Enterprises

The immediate benefits for developers and enterprises include reduced time spent building custom connectors, more predictable agent behavior, and simplified deployment in secure environments.

The broader vision is to transform the agent landscape from closed platforms to an open, interoperable software ecosystem, mirroring the principles that drove the development of the modern web.

• MCP: Standardized connection between models, agents, tools, and data.
• Goose: Open-source agent framework for coding, analysis, and documentation.
• AGENTS.md: Instruction file for guiding AI coding tools.