OpenAI, Anthropic Back Linux Foundation's AI Agent Standard

As artificial intelligence evolves beyond simple chatbots to sophisticated systems capable of taking autonomous actions, the Linux Foundation is spearheading a new initiative to prevent this emerging field from fragmenting into a maze of incompatible, proprietary products. The organization has launched the Agentic AI Foundation (AAIF), a neutral home for open-source projects focused on AI agents.

The AAIF kicks off with crucial contributions from industry giants Anthropic, Block, and OpenAI, signaling a concerted effort to standardize the burgeoning AI agent era.

Leading Tech Companies Contribute Core Technologies

Each founding member brings a vital piece of the puzzle to the AAIF:

• Anthropic is donating its Model Context Protocol (MCP), a standard designed to connect AI models and agents to various tools and data sources.
• Block, the fintech company behind Square and Cash App, is contributing Goose, its open-source agent framework.
• OpenAI is providing AGENTS.md, a straightforward instruction file that developers can add to repositories to guide the behavior of AI coding tools.

These contributions are envisioned as the foundational "plumbing" necessary for the agent era, ensuring basic compatibility and functionality across different systems.

Beyond the initial contributors, the AAIF boasts a strong roster of members including AWS, Bloomberg, Cloudflare, and Google. This broad industry support underscores a collective push for shared guardrails, aiming to make AI agents trustworthy and scalable.

The Imperative for Interoperability

Nick Cooper, an OpenAI engineer, emphasizes that protocols serve as a shared language, enabling diverse agents and systems to collaborate without developers constantly reinventing 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 told TechCrunch.

Jim Zemlin, executive director of the Linux Foundation, articulated the goal more directly: to avert a future dominated by "closed wall" proprietary stacks, where tool connections, agent behavior, and orchestration are locked within a few 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 Openness Play with Goose

While Block isn't typically known for AI infrastructure, its contribution of Goose highlights a commitment to open alternatives. Brad Axen, Block's AI Tech Lead, views Goose as proof that open-source agents can rival proprietary solutions at scale. Thousands of engineers reportedly use Goose weekly for coding, data analysis, and documentation.

Open-sourcing Goose offers Block a dual advantage. "Getting it out into the world gives us a place for other people to come help us make it better," Axen explained. "We have a lot of contributors from open source, and everything they do to improve it comes back to our company." Donating Goose to the Linux Foundation also provides Block with community stress-tests and positions it as a working example of AAIF's vision—an agent framework designed to integrate with shared building blocks like MCP and AGENTS.md.

Anthropic's Protocol Layer Contribution

Anthropic's donation of MCP to the Linux Foundation follows a similar strategy at the protocol layer. The objective is to establish MCP as the neutral infrastructure for connecting AI models to tools, data, and applications, eliminating the need for countless one-off adapters.

"The main goal is to have enough adoption in the world that it's the de facto standard," MCP co-creator David Soria Parra told TechCrunch. "We're all better off if we have an open integration center where you can build something once as a developer and use it across any client."

This move signals that MCP's governance will not be controlled by a single vendor, fostering broader trust and adoption.

Governance and the Path Forward

The Linux Foundation, already home to major AI and developer infrastructure projects like PyTorch, Ray, and Kubernetes, created the AAIF specifically for agent standards and orchestration, including shared safety patterns and interoperability. The AAIF's structure is funded through a "directed fund," where companies contribute via membership dues. However, Zemlin asserts that funding does not equate to control; project roadmaps are determined by technical steering committees, ensuring no single member dictates direction.

A key question remains: will the AAIF evolve into genuine infrastructure or simply another industry alliance? Zemlin suggests that an early indicator of success, beyond standard adoption, would be the development and implementation of shared standards by vendor agents globally. OpenAI's Cooper hopes for an evolving set of standards. "I don't want it to be a stagnant thing... They should evolve and continually accept further input," he noted.

Even with open governance, there's a subtle risk that one company's implementation could become the de facto default due to rapid deployment or widespread usage. Zemlin, however, views this as potentially positive, citing open-source history—like Kubernetes "winning" the container race—as evidence that "dominance emerges from merit and not vendor control."

Benefits for Developers and the Future of AI

For developers and enterprises, the immediate appeal of the AAIF's efforts is clear: less time spent building custom connectors, more predictable agent behavior across codebases, and simpler deployment in security-conscious environments.

The larger vision is more ambitious: if tools like MCP, AGENTS.md, and Goose become standard infrastructure, the AI agent landscape could transform from closed platforms to an open, "mix-and-match" software world, reminiscent of the interoperable systems that underpin the modern web.