The Linux Foundation, the nonprofit supporting open-source projects such as the Linux operating system kernel, today announced the formation of the Agentic Artificial Intelligence Foundation alongside major contributions of innovative AI technology.

Major software technology companies Anthropic PBC, OpenAI Group PBC and Block Inc. joined the launch as founding members with contributions of innovative AI projects. Anthropic contributed Model Context Protocol, a universal open standard for connecting AI models to tools. OpenAI offered the AGENTS.md specification that gives AI coding agents consistent, project-specific knowledge. And Block provided Goose, an open source local-first AI framework.

Platinum members of AAIF include a showcase of the Who’s Who of the largest cloud, data and AI industry enterprise leadership including Anthropic, Block, OpenAI, Amazon Web Services Inc., Bloomberg LP, Cloudflare Inc., Google LLC and Microsoft Corp.

Agentic AI refers to a still-emerging AI technology that allows AI models and systems to set goals, plan, reason and take actions independently by breaking down long-horizon tasks into manageable actions. These AI systems go beyond simple chatbot-style question-response and become interactive software that can integrate tools and collaborate with humans or other agents. They act proactively and react to changes in their environment to execute multi-step jobs with minimal human oversight.

This new trend is changing how major large language model vendors develop AI models and how enterprise companies are adopting – and adapting to – the use of AI in their workflows. AI agents have the purposeful promise of becoming automatic and intelligent “teammates” to work alongside humans and reasoning software capable of automating tedious, dynamic tasks in the background.

The Linux Foundation stated that innovation in agentic AI will need open-source software infrastructure to thrive. Making the underlying technologies community-governed rather than controlled by a cabal of big companies will promote open communication and advancement.

The project is already a neutral home for numerous open-source projects fundamental to modern computing, including Linux, Kubernetes, PyTorch and RISC-V.

The rise of agentic AI

A report from UiPath Inc. revealed that agentic AI enterprise adoption is accelerating rapidly, with roughly 65% of organizations piloting or deploying agentic systems by mid-year 2025, and around nine out of ten executives planning to increase investment throughout 2026. Multi-agent systems have been shown to deliver substantial performance gains, with up to 60% fewer errors and 40% faster execution compared to traditional processes.

However, it cites the recent MIT report that only 5% of companies have realized any meaningful financial returns from AI efforts. The UiPath report also warned that 96% of information technology experts and security leaders harbor concerns about the increasing risk behind AI agents and called for it to be addressed swiftly.

Since the release of MCP by Anthropic, a provider of advanced AI models including Claude and Claude Code, it has rapidly become a widely adopted universal standard. With more than 10,000 published MCP servers, it is now a fundamental turnkey infrastructure for AI interoperability. The protocol has been adopted by Claude, Anysphere Inc.’s Cursor, Microsoft Corp.’s Copilot and VSCode, Google LLC’s Gemini, OpenAI’s ChatGPT and other widely used AI platforms.

“MCP started as an internal project to solve a problem our own teams were facing. When we open sourced it in November 2024, we hoped other developers would find it as useful as we did,” said Mike Krieger, chief product officer at Anthropic. “A year later, it’s become the industry standard for connecting AI systems to data and tools, used by developers building with the most popular agentic coding tools and enterprises.”

Released in August, AGENTS.md has already been adopted by over 40,000 open-source projects and coding agents to act as the internal framework for guiding safe, interoperable agentic AI. It’s a simple markdown-based convention that provides a human-readable format for reliable control of AI behaviors across multiple code repositories.

Despite its unusual name, Goose provides a standardized, open and modular framework for agentic AI systems, enabling anyone to build reasoning applications. It forms a standardized open, modular framework for agentic AI systems, allowing anyone to build reasoning applications. It offers connectivity for AI models, MCP and extensible tools in an out-of-the-box solution with a built-in desktop solution that jumpstarts development.

“We’re at a critical moment for AI,” said Manik Surtani, head of open source at Block. “The technology that will define the next decade, that promises to be the biggest engine of economic growth since the internet. It can either remain closed and proprietary for the benefit of few, or be driven by open standards, open protocols and open access for the benefit of all.”

Image: Microsoft Designer/SiliconANGLE