Big Tech has been highlighting our entrance into the era of AI agents over the past year, although many of those promises remain largely theoretical. In an effort to bring imagination into reality, companies are creating a suite of tools that guide the development of generative AI. Notably, a collaboration among major AI contenders such as Anthropic, Block, and OpenAI has led to the formation of the Agentic AI Foundation (AAIF). This strategic move is set to elevate select technologies into becoming potential industry standards for future AI development.
The roadmap for developing agentic AI models is still uncertain, but significant investment from companies has led to the emergence of certain tools. The AAIF, now part of the nonprofit Linux Foundation, aims to oversee the development of three primary AI technologies: Model Context Protocol (MCP), goose, and AGENTS.md.
MCP is perhaps the most prominent of the three, having been open-sourced by Anthropic a year ago. It strives to provide standardized connections between AI agents and data sources. Anthropic, along with the AAIF, likens MCP to a "USB-C port for AI." Instead of crafting custom integrations for each separate database or cloud service, MCP offers a streamlined way for developers to access any MCP-compliant server quickly and efficiently.
Since its debut, MCP has enjoyed widespread adoption across the AI sector. At its I/O 2025 conference, Google announced it would be incorporating MCP into its developer tools, and many of its products have since integrated MCP servers to facilitate easier data access for AI agents. Shortly after its release, OpenAI also embraced MCP.
With the expanded adoption of MCP, users may have greater control over their AI experiences. For example, the new Pebble Index 01 ring features a local language learning model (LLM) responsive to voice commands, supporting MCP for enhanced customization.
Though local AI models are often restricted in comparison to their larger, cloud-based counterparts, MCP can help bridge these gaps in functionality. "A lot of tasks on productivity and content are fully doable on the edge," Vinesh Sukumar, head of AI products at Qualcomm, tells Ars. "With MCP, you have a handshake with multiple cloud service providers for any kind of complex task to be completed."