AI's Peace Treaty: Tech Giants Unite to Build the Internet of Agents

SAN FRANCISCO, CA – December 09, 2025 – In a move that signals a tectonic shift in the artificial intelligence landscape, the industry's fiercest competitors are laying down their arms to build a common future. The Linux Foundation today announced the formation of the Agentic AI Foundation (AAIF), an unprecedented collaboration uniting rivals like Anthropic, Block, Google, Microsoft, and OpenAI. Their shared goal: to establish the open, interoperable, and transparent infrastructure needed for the next generation of autonomous AI systems.

This alliance is not merely a symbolic gesture; it is a strategic consolidation of foundational technologies that have rapidly become critical to the development of “agentic AI.” These are not the chatbots of yesterday, but sophisticated AI systems capable of reasoning, planning, and executing complex, multi-step tasks autonomously. With the market for this technology projected to surge from roughly $7.5 billion in 2025 to nearly $200 billion by 2034, the stakes for establishing the underlying rules of the road could not be higher.

The AAIF launches with three cornerstone projects donated by its founding members: Anthropic’s Model Context Protocol (MCP), Block’s goose framework, and OpenAI’s AGENTS.md standard. By placing these crucial building blocks under the neutral governance of the Linux Foundation—a veteran steward of world-changing open-source projects like the Linux kernel and Kubernetes—the industry is making a collective bet that the future of AI must be built on collaboration, not walled gardens.

Forging the Protocols for an 'Internet of Agents'

At the heart of the AAIF's mission is the creation of what could be described as an 'Internet of Agents.' Just as protocols like HTTP and TCP/IP enabled disparate computers to communicate and form the World Wide Web, the projects within the AAIF aim to allow autonomous AI agents to interact, share context, and collaborate seamlessly across different platforms and providers.

Anthropic's Model Context Protocol (MCP), open-sourced in late 2024, has become the de facto standard for this new paradigm. It provides a universal language for AI models to securely connect with external tools, APIs, and data sources. Before MCP, connecting an AI to a new data source required building a custom, brittle integration. Now, with over 10,000 public MCP servers and adoption by major platforms including ChatGPT, Google's Gemini, and Microsoft Copilot, the protocol solves this integration challenge at scale. As Shawn Edwards, CTO of Platinum member Bloomberg, noted, MCP provides the “essential connective layer” for deploying sophisticated AI systems in complex environments like finance.

Complementing MCP is AGENTS.md, a simple yet powerful standard contributed by OpenAI. Released in August 2025, it allows developers to provide project-specific instructions to AI coding agents in a simple markdown file, much like a README.md file guides human developers. With adoption already claimed across over 60,000 open-source projects, it makes the behavior of AI agents far more predictable and reliable as they navigate unfamiliar codebases. Finally, Block's goose framework provides a practical, open-source toolkit for building and executing these agentic workflows locally, serving as a reference implementation for the entire ecosystem.

Together, these technologies form a foundational stack. AGENTS.md tells an agent what to do within a specific context, while MCP provides the universal channels for the agent to access the tools and data it needs to get the job done. This standardized architecture is designed to prevent vendor lock-in and empower developers to build complex, multi-agent systems that are not tethered to a single company’s proprietary ecosystem.

A Strategic Alliance in a High-Stakes Game

The formation of the AAIF is as much a strategic business maneuver as it is a technical one. The gathering of Platinum members reads like a who's who of the AI arms race: Amazon Web Services, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, and OpenAI. Their collaboration underscores a shared understanding that a fragmented ecosystem of incompatible agents would stifle innovation and slow market growth. By agreeing on common standards, these giants are creating a larger, more stable market for everyone.

“For the agentic future to become a reality, we have to build it together, and we have to build it in the open,” said Chris DiBona, a VP in Microsoft's Office of the CTO. This sentiment is echoed by other leaders, who emphasize that shared, community-driven protocols are essential for trust and predictability.

The motivation is clear: a standardized ecosystem allows these companies to compete on the quality of their models and specialized services rather than on controlling the underlying plumbing. An AI agent built to run on Google's infrastructure can now, in theory, seamlessly call a tool hosted on AWS or leverage a model from Anthropic. This interoperability is crucial for enterprise customers who operate in multi-cloud, hybrid environments and cannot afford to be locked into a single vendor's AI stack.

This 'coopetition' is a mature move for an industry defined by intense rivalry. It mirrors the evolution of previous technological revolutions where open standards ultimately unlocked far greater economic value than proprietary systems could alone. The Linux Foundation's proven model of neutral governance provides the necessary trust framework for these competitors to invest in shared infrastructure, confident that no single entity will seize control.

The Geopolitical Dimension of AI Infrastructure

By stewarding the future of agentic AI, the Linux Foundation is expanding its role from the bedrock of the internet and cloud computing into the core of global intelligent infrastructure. This move carries significant geopolitical and security implications. Standardized, interoperable AI agents have the potential to revolutionize everything from defense logistics and supply chain resilience to cybersecurity and intelligence analysis.

An open framework could accelerate the development of defensive AI agents capable of autonomously detecting and neutralizing cyber threats across national critical infrastructure. However, it also creates a common architecture that state and non-state adversaries will seek to understand and exploit. The governance model of the AAIF, with its emphasis on transparency and security, will be critical in managing these dual-use risks.

The broad coalition, which includes dozens of Gold and Silver members from finance, e-commerce, and enterprise software, demonstrates that the impact of agentic AI is expected to be economy-wide. The open standards being forged today are not just for developers; they are the foundational architecture upon which future autonomous systems for global commerce, manufacturing, and security will be built. As this new internet of agents begins to take shape, the decisions made within the AAIF will have a lasting impact on global economic competitiveness and stability for decades to come.