Model Context Protocol

MCP graduated from developer standard to industry standard — the Linux Foundation governance and cross-vendor support mean you can build on it with confidence.

Agentic·Infrastructure

modelcontextprotocol.io

Our Take

What It Is

The Model Context Protocol (MCP) is an open standard that defines how AI models connect to external tools, data sources, and services. Originally created by Anthropic, MCP was donated to the Linux Foundation's Agentic AI Foundation (AAIF) in early 2026, co-founded with Block and OpenAI, with support from Google, Microsoft, AWS, and Cloudflare.

Why It Matters

MCP stays in Proven, but the story has changed. It's no longer "Anthropic's protocol that others support." The Linux Foundation governance move transforms it into vendor-neutral infrastructure. When the AAIF co-founders include both Anthropic and OpenAI, the protocol wars are effectively over for tool connectivity.

The ecosystem numbers tell the adoption story: 97 million monthly SDK downloads, over 10,000 community servers, and enterprise-grade additions like SurePath AI's MCP Policy Controls and Red Hat's RHEL MCP server. Apple shipping MCP support in Xcode 26.3 was the signal that closed any remaining debate about adoption trajectory.

Key Developments

Mar 2026: Apple integrates MCP into Xcode 26.3 for agentic coding, connecting to Claude Agent and OpenAI Codex.
Feb 2026: Linux Foundation AAIF launches with MCP as its first governed protocol. Co-founded by Anthropic, Block, and OpenAI.
Feb 2026: SurePath AI launches MCP Policy Controls for enterprise governance. Red Hat releases RHEL MCP server.
Jan 2026: MCP SDK monthly downloads pass 97 million. Community server registry exceeds 10,000 published servers.

What to Watch

The 2026 MCP roadmap focuses on transport scalability, agent-to-agent communication, and enterprise readiness features. Watch for how MCP and the A2A Protocol (also in this edition) interact — MCP handles agent-to-tool, A2A handles agent-to-agent. Together, they could form the complete connectivity standard for agentic systems. The enterprise governance features (policy controls, audit trails) will determine whether MCP can satisfy compliance teams in regulated industries.

Strengths

Industry governance: Linux Foundation AAIF with Anthropic, OpenAI, Google, Microsoft, and AWS as backers. This is as close to "industry standard" as it gets.
Ecosystem scale: 97M monthly SDK downloads and 10,000+ community servers means most integrations you need already exist.
Cross-platform adoption: Works with Claude, GPT, Gemini, open-source models, Xcode, Cursor, VS Code, and virtually every AI coding tool.
Enterprise signals: SurePath AI policy controls and Red Hat RHEL server show enterprise-grade tooling is arriving.

Considerations

Governance transition: The move to Linux Foundation is recent. Governance processes, contribution workflows, and decision-making are still stabilising.
Operational overhead: Each MCP server is a separate process to deploy and maintain. At scale, this adds infrastructure complexity.
Security surface: 10,000+ community servers means quality and security vary widely. Vetting servers for production use requires due diligence.
Specification evolution: The protocol is still evolving. Servers built for earlier versions may need updates as new features land.