Catenaa, Tuesday, December 09, 2025- Major US technology companies formed a new Agentic AI Foundation under the Linux Foundation as concerns grow over China’s rapid gains in open-source artificial intelligence.
Anthropic, OpenAI and Block launched the group with backing from Google, Microsoft, Amazon Web Services, Bloomberg and Cloudflare.
The effort marks one of the industry’s broadest attempts to coordinate standards for autonomous AI agents.
The foundation brings together three core projects. Anthropic contributed its Model Context Protocol, which has seen wide adoption across developer platforms.
OpenAI added its AGENTS.md specification used by tens of thousands of code repositories. Block submitted its Goose framework for local agent development. All three now fall under neutral governance.
The decision comes as China expands its influence in open-source AI. A recent MIT analysis of global model downloads found Chinese developers now outpace US counterparts.
Companies such as DeepSeek and Alibaba have released high-performance models that developers worldwide increasingly rely on.
Executives involved in the new foundation said common standards would help maintain US competitiveness by ensuring developers can build agents across platforms without restrictions.
The group aims to strengthen shared infrastructure rather than rely on closed systems that risk losing ground to overseas providers.
The White House recently warned that open-source and open-weight models carry geostrategic value, with broad adoption shaping academic research and commercial use.
The foundation said no single member will direct policy and published a tiered membership list that includes Amazon, Cisco, IBM, Oracle, SAP and others.
Supporters view the coordinated approach as a step toward stabilizing the base layer of agentic AI while US companies work to regain momentum in the open-source arena.
