OpenAIAnthropic and Block have launched a modern open source organization, the Agentic AI Foundation, to promote standards for artificial intelligence agents.
The three companies are also transferring ownership of some commonly used agent technologies to the foundation. This includes anthropic Model Context Protocol (MCP)that enables agents to connect and interact; OpenAI Agents.mdwhich allows programs and websites to specify rules for coding agents; AND Gooseagent building framework developed by Block. These technologies were already available for free, but thanks to the modern foundation, others will be able to contribute to their development.
“MCP is used by many companies, but there are others [who don’t use it]” says Nick Cooper, who leads the protocol effort at OpenAI. Cooper says making MCP an open standard should encourage developers and companies to adopt it and build systems that integrate agent-based AI. “This open interoperability – this open standard – really means that companies can communicate between vendors and agent systems.”
The Agentic AI Foundation is established within the Linux Foundation, which oversees the development of the widely used open-source Linux operating system, as well as other projects. The foundation provides legal and technological support in creating an open source foundation. Other companies that have joined AAIF, beyond the three founding members, include Google, Microsoft, AWS, Bloomberg and Cloudflare.
The new foundation reflects an emerging shift from chat-based AI systems to the greater use of programs that take actions on behalf of users. This type of agent-based AI creates the opportunity for a potentially lucrative new paradigm in which AI agents use networks and negotiate with each other to support all kinds of applications. For example, consumers can use AI assistants for shopping and bookings, while businesses use AI agents to manage transactions and customer interactions.
Srinivas Narayanan, director of B2B application technology at OpenAI, predicts a time when large numbers of AI agents will routinely communicate with each other as part of business operations. An AI industry working to the same open standards should help ensure these interactions run smoothly. “Open source will play a very big role in shaping and implementing AI in the real world,” says Narayanan.
The issue of openness seems crucial for AI today. American companies make money mainly by offering access to productive closed models via application programming interfaces, or APIs. Meta has already made the weights of its top model, the Llama, available for anyone to download and run, although the company recently signaled a shift to a more closed approach. A number of Chinese artificial intelligence companies, including DeepSeek, Alibaba, Moonshot AIAND Z.aithey provide mighty open source models that have become popular among developers, startups, and AI researchers. Some fear that this image could provide Chinese companies with a major strategic advantage over time.