OpenAI, Anthropic, Block Launch Agentic AI Foundation
OpenAI, Anthropic, and Block launch the Agentic AI Foundation with the Linux Foundation to standardize autonomous AI agents.
OpenAI, Anthropic, and Block Co-Found Agentic AI Foundation Under Linux Foundation
OpenAI, Anthropic, and Block have joined forces with the Linux Foundation to launch the Agentic AI Foundation (AAIF), a new open-source initiative aimed at standardizing and accelerating the development of autonomous AI agents. As part of this effort, OpenAI is donating its AGENTS.md specification, while Anthropic contributes the Model Context Protocol (MCP) and Block contributes its open-source agent framework goose. The foundation is designed to ensure that agentic AI evolves as an open, interoperable, and trustworthy ecosystem rather than a fragmented landscape dominated by proprietary systems.
The AAIF is hosted under the Linux Foundation’s neutral governance model and includes platinum members Amazon Web Services, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, and OpenAI. These companies are positioning agentic AI — systems that can plan, act, and collaborate autonomously — as the next major phase of artificial intelligence, following the era of conversational models.
What Is Agentic AI?
Agentic AI refers to AI systems that can take initiative, make decisions, and perform actions with minimal human intervention. Unlike traditional chatbots that respond to prompts, agentic systems can break down goals, use tools, call APIs, and coordinate with other agents to complete complex tasks. Examples include AI assistants that can book travel, manage workflows, or debug and deploy code autonomously.
The AAIF defines its mission as advancing open source agentic AI through shared protocols, reference implementations, and community-driven standards. The foundation explicitly avoids being a broad AI or machine learning umbrella, instead focusing narrowly on agent-centric infrastructure, interoperability, and safety.
Founding Projects: MCP, goose, and AGENTS.md
The AAIF launches with three core technical contributions:
- Model Context Protocol (MCP) – Anthropic’s protocol for enabling AI models to interact with external tools and services in a structured, secure way. MCP has already been adopted by OpenAI and others as a foundational layer for agent tool use.
- goose – Block’s open-source agent framework that provides a modular, extensible architecture for building and orchestrating autonomous agents.
- AGENTS.md – OpenAI’s specification for defining agent behavior, capabilities, and interactions. OpenAI describes this as a step toward open, interoperable standards for safe agentic AI.
OpenAI emphasized that shared, community-driven protocols are essential for a healthy agentic ecosystem. In its announcement, the company stated it has also contributed the Codex CLI, Agents SDK, and Apps SDK to support the open agentic ecosystem, building on these shared protocols.
Governance and Industry Backing
The AAIF operates under the Linux Foundation’s open governance model, ensuring that no single vendor controls the direction of the foundation. This approach mirrors the success of open standards bodies like the W3C for the web and the Open Container Initiative for cloud-native infrastructure.
Platinum members include:
- Amazon Web Services (AWS)
- Anthropic
- Block
- Bloomberg
- Cloudflare
- Microsoft
- OpenAI
Swami Sivasubramanian, VP of Agentic AI at AWS, noted that the industry is at a pivotal moment: “We are seeing AI enter a new phase, as conversational systems shift to autonomous agents that can work together.” He added that AWS believes open standards are critical to ensuring agentic AI remains interoperable and trustworthy.
Why This Matters: From Experimentation to Enterprise Infrastructure
The formation of the AAIF signals that agentic AI is maturing from experimental prototypes into enterprise-grade infrastructure. Enterprises are increasingly concerned about vendor lock-in, security, and governance when adopting agent-based systems.
For example, the Model Context Protocol (MCP) is already being used in production environments, but enterprises have been hesitant to bet on a protocol controlled by a single vendor. By moving MCP, goose, and AGENTS.md under neutral governance, the AAIF provides the transparency, stability, and security that large organizations require.
Security and governance tools like Agentgateway (also part of the Linux Foundation ecosystem) are expected to integrate closely with AAIF projects, providing enterprises with production-grade capabilities for authentication, authorization, observability, and guardrails in agentic architectures.
Industry Impact and Future Outlook
The AAIF has the potential to become the de facto standards body for agentic AI, much like the W3C for the web or the IETF for internet protocols. By fostering open collaboration, the foundation aims to:
- Prevent fragmentation across agent platforms
- Enable agents from different vendors to interoperate
- Accelerate innovation through shared infrastructure
- Improve safety, security, and transparency
If successful, the AAIF could shape how AI agents are built, deployed, and governed across industries — from finance and healthcare to software development and customer service.
The move also reflects a broader industry trend: as AI becomes more capable and autonomous, the need for open standards, interoperability, and neutral governance becomes increasingly urgent. The AAIF represents a coordinated effort by leading AI companies to ensure that agentic AI develops in a way that is open, collaborative, and aligned with the public interest.
