OpenAI, Anthropic, and Block collaborate with the Linux Foundation on a new initiative to establish standards for the AI agent era

Linux Foundation Unveils Agentic AI Foundation to Promote Open AI Agent Standards

As artificial intelligence evolves from simple chatbots to more advanced systems capable of taking action, the Linux Foundation has introduced a new initiative aimed at preventing the fragmentation of AI agents into incompatible, proprietary solutions.

This new initiative, called the Agentic AI Foundation (AAIF), will serve as an impartial hub for open-source projects focused on AI agents. The foundation’s launch is supported by contributions from Anthropic, Block, and OpenAI.

Key Contributions from Industry Leaders

• Anthropic is offering its Model Context Protocol (MCP), a standard designed to link AI models and agents with external tools and data sources.
• Block is providing Goose, its open-source framework for building AI agents.
• OpenAI is contributing AGENTS.md, a straightforward instruction file that developers can add to their repositories to guide AI coding tools in their interactions.

These resources form the essential infrastructure for the next generation of AI agents.

Broad Industry Support

Additional founding members of AAIF include AWS, Bloomberg, Cloudflare, and Google, reflecting a collective industry effort to establish common standards and safeguards that ensure AI agents are reliable and interoperable at scale.

The Importance of Shared Protocols

Nick Cooper, an engineer at OpenAI, emphasizes that shared protocols act as a universal language, enabling diverse agents and systems to collaborate without requiring developers to repeatedly build custom integrations.

“We need multiple protocols to negotiate, communicate, and work together to deliver value for people, and that sort of openness and communication is why it’s not ever going to be one provider, one host, one company,” Cooper explained to TechCrunch.

Jim Zemlin, executive director of the Linux Foundation, highlighted the initiative’s purpose: to avoid a future dominated by closed, proprietary ecosystems where agent behavior and integrations are restricted to a few major platforms.

“By bringing these projects together under the AAIF, we are now able to coordinate interoperability, safety patterns, and best practices specifically for AI agents,” Zemlin stated.

Block’s Open Source Approach with Goose

Although Block, the company behind Square and Cash App, is not traditionally associated with AI infrastructure, it is championing openness by releasing Goose as open source. According to AI Tech Lead Brad Axen, this demonstrates that open solutions can rival proprietary agents, with thousands of engineers using Goose each week for programming, data analysis, and documentation.

Open-sourcing Goose serves two purposes for Block: it invites external contributions to enhance the tool, and it allows the company to benefit from improvements made by the open-source community.

“Releasing Goose publicly gives us a platform for others to help us improve it,” Axen told TechCrunch. “We receive valuable contributions from the open-source community, and every enhancement benefits our company as well.”

By donating Goose to the Linux Foundation, Block also gains access to broader community testing and positions Goose as a practical example of AAIF’s vision—an agent framework that integrates seamlessly with standards like MCP and AGENTS.md.

Anthropic’s Commitment to Open Protocols

Anthropic is taking a similar approach by transferring MCP to the Linux Foundation, aiming to establish it as a neutral standard that connects AI models with tools, data, and applications—eliminating the need for countless custom adapters.

“The main goal is to achieve widespread adoption so that MCP becomes the default standard,” said MCP co-creator David Soria Parra. “An open integration hub allows developers to build once and deploy across any client.”

Handing MCP to AAIF ensures that no single company will control its direction.

Governance and Funding

The Linux Foundation’s decision to create AAIF stems from the need for dedicated oversight of agent standards and orchestration, including safety and interoperability guidelines. While AAIF is financed through a directed fund—allowing companies to contribute via membership dues—Jim Zemlin stresses that financial support does not equate to control. Technical steering committees, not individual members, determine project direction.

Looking Ahead: Will AAIF Deliver Lasting Impact?

A key question remains: will AAIF become a foundational part of the AI ecosystem or simply another industry consortium?

“A sign of early success, beyond adoption, would be the creation and use of shared standards by AI agents worldwide,” Zemlin noted.

For OpenAI’s Cooper, the evolution of these standards is crucial: “I don’t want these protocols to stagnate. They should continue to develop and welcome new input.”

However, even with open governance, there is a possibility that one company’s implementation could become the default due to speed or popularity. Zemlin sees this as a natural outcome, citing the rise of Kubernetes as an example where open-source merit, not vendor dominance, determined the standard.

Benefits for Developers and the Future of AI Agents

For developers and organizations, the immediate advantages are clear: reduced time spent on custom integrations, more predictable agent behavior across projects, and easier deployment in secure environments.

On a broader scale, if standards like MCP, AGENTS.md, and Goose gain widespread adoption, the AI agent ecosystem could transition from isolated, proprietary platforms to an open, interoperable landscape—mirroring the collaborative systems that underpin today’s internet.