GitHub and Microsoft are deepening their commitment to the Model Context Protocol (MCP), a growing industry standard for linking AI models to external data sources. Announced at Microsoft Build 2025, the two companies are now formally joining the MCP steering committee, further entrenching MCP in the AI developer ecosystem.
MCP facilitates seamless, two-way communication between AI models and external systems—business tools, software platforms, content repositories, and development environments—via MCP servers and clients. With growing adoption from major players like OpenAI and Google earlier this year, MCP is becoming an essential protocol for making models truly interactive with enterprise systems.
Microsoft’s next steps include integrating MCP support across its platforms, including Azure and Windows 11. Developers will soon be able to expose app capabilities as MCP servers, enabling models to interact with native system functionalities such as the File System, Windowing, and Windows Subsystem for Linux. As Microsoft explained in its press briefing, “this will include Windows system functionalities […] for [models] to interact with” (source).
Security and identity are also being prioritized. Microsoft’s collaboration with Anthropic and the broader MCP community resulted in an updated authorization spec. This framework allows apps to securely access MCP servers using familiar, trusted sign-in methods, even when interfacing with sensitive resources like personal cloud storage or subscription-based services.
GitHub’s role centers around discoverability and infrastructure. Together with the MCP steering committee, they’ve developed a registry service for MCP servers, enabling centralized management of MCP implementations. This makes it easier for developers to share, find, and configure MCP endpoints, whether public or private.
With Microsoft and GitHub now fully onboard, the MCP initiative gains not only technical momentum but also strategic validation. Their contributions—ranging from platform integration to registry architecture—mark a significant step toward standardizing how AI connects to the real-world systems it needs to understand and act on.