New MCP specification addresses the main barrier to enterprise adoption
The Model Context Protocol (MCP) has received its largest update since launch, changing its core from a stateful, server-session-based system to a stateless request-and-response model. This is intended to improve reliability and scalability, addressing a major obstacle to deploying MCP widely in large organisations.
The update also introduces multi-round-trip requests, header-based routing, cacheable results, stronger authorisation, a formal extensions framework and updated top-tier software development kits. MCP, originally created by Anthropic and now managed through the Linux Foundation’s Agentic AI Foundation, is supported by contributors including OpenAI, Google, Microsoft and Amazon; a new policy generally requires 12 months’ notice before deprecated features can be removed.
In short
- MCP’s new stateless core aims to make enterprise deployment easier.
- The specification adds routing, caching and stronger security features.
- MCP now has broad industry support beyond Anthropic.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Moving MCP to stateless requests could make deployments more resilient and easier to scale across large organisations, while cacheable results, stronger authorisation and a formal extensions framework address practical operational and security needs. A predictable 12-month notice period for removing features may also give adopters more confidence to invest in the protocol.
The case against
A major redesign can create migration work and compatibility risks for organisations already using session-based MCP, and statelessness may not suit every workflow that benefits from persistent server state. The new features and updated software development kits could add implementation complexity, while a general 12-month deprecation notice may slow necessary changes in some cases.
Explained simply
The Model Context Protocol is a system that helps artificial intelligence applications connect with each other and access external tools and data they need. Originally created by Anthropic, it is now overseen by the Linux Foundation and supported by major technology companies including OpenAI, Google, Microsoft and Amazon.
The technology has struggled to gain adoption in large enterprises because of technical limitations that made it unreliable when deployed at scale. These architectural shortcomings have been the primary barrier to widespread business use of the protocol.
The latest specification update is intended to address these fundamental problems. If successful, it could remove a significant obstacle to the system gaining acceptance in corporate settings where reliability and scalability are essential.