In a significant move, major technology firms such as Anthropic, OpenAI, and Block have teamed up to establish the Agentic AI Foundation under the Linux Foundation's umbrella. This initiative aims to develop open standards for AI agents, leveraging contributions from three core projects: Anthropic's Model Context Protocol, OpenAI's AGENTS.md specification, and Block's Goose framework. The foundation has attracted prominent platinum members like Amazon Web Services, Bloomberg, Cloudflare, Google, and Microsoft, signaling a unified industry effort towards neutral governance in agentic AI infrastructure.
The formation of this group comes at a time when China is rapidly advancing in global open-source AI, prompting America's leading tech companies to collaborate. Cloudflare's CTO, Dane Knecht, emphasized the importance of open standards and protocols like MCP in fostering a vibrant developer ecosystem without the risk of vendor lock-in. U.S. companies face a dilemma: they desire the recurring revenue from proprietary APIs, but risk becoming obsolete if China dominates the foundational layer. Thus, standardizing on MCP and agentic AI is seen as a strategic move to retain competitiveness and capture value through superior models rather than relying on ecosystem lock-in.
This pragmatic collaboration marks a victory for the open-source community, with competitors recognizing the collective benefits of standardization over fragmentation. The U.S. AI industry's efforts to reclaim open-source dominance could prove advantageous for users worldwide, including those in China, by potentially enhancing open-source development.
Anthropic has contributed its Model Context Protocol to the Linux Foundation, which allows AI models to use tools creatively beyond mere API calls. Since its launch, MCP has gained substantial traction with over 10,000 active servers, support from platforms like ChatGPT, Gemini, Microsoft Copilot, VS Code, and 97 million monthly SDK downloads. OpenAI's AGENTS.md, a lightweight specification adopted by 60,000 repositories, standardizes project instructions for AI agents. Meanwhile, Block's Goose offers a local-first agent framework. Together, these contributions operate under the Linux Foundation's neutral governance.
The timing of this initiative is crucial. A December 2025 MIT study revealed that China now accounts for 17.1% of global open-source AI downloads, surpassing the U.S.'s 15.8%. Companies like DeepSeek and Alibaba have saturated the market with high-performance models, while American firms have largely focused on closed APIs in pursuit of profitability. According to Jim Zemlin, executive director of the Linux Foundation, AI is entering a new phase as conversational systems evolve into autonomous agents capable of collaboration. The foundation addresses a strategic vulnerability: reliance on Chinese open-source models could diminish dependency on U.S. cloud services and APIs.
The foundation's responsive governance and vendor neutrality are underscored by its diverse membership. Platinum members include Amazon, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, and OpenAI. Gold members feature Cisco, Datadog, Docker, IBM, Oracle, SAP, Snowflake, and Twilio, while Silver members include Hugging Face, Uber, SUSE, and others. The organization stresses that no single company directs its path.
China's edge comes from a strategic approach that prioritizes adaptability and modular innovation over large-scale AI operations. By providing open weights, Chinese companies allow developers to build upon their technologies, enhancing their adaptability. The Trump administration's AI Action Plan acknowledges the threat, noting that open-source and open-weight models could become global standards in certain business sectors and academic research, thus holding geostrategic significance.