AI's most important protocol is getting a little bit easier to use

8:00 PM · July 21, 2026

The Model Context Protocol, one of the basic building blocks underlying how AI models connect to external data sources and tools, is getting an update meant to make it meaningfully easier to run at scale, according to TechCrunch. MCP functions as a kind of plumbing that lets a chatbot or AI agent reach into a calendar, a database, or an internal company tool, and has been widely adopted since Anthropic introduced it as a shared standard. The update changes how the protocol handles session tracking on the server side, moving to a looser, stateless approach similar to how most ordinary websites already operate, rather than requiring a server to maintain detailed state about each ongoing session. That distinction matters in practice because real companies typically spread traffic across dozens of servers that do not automatically share information with each other, meaning today's MCP servers have had to do extra coordination work just to keep track of which session belongs to which user across that distributed infrastructure. That overhead has been a genuine obstacle for any company trying to run an MCP server at meaningful scale, and is part of why large, first party MCP integrations from major companies have been slower to appear than the protocol's broad conceptual adoption might have suggested. The simplified approach is expected to lower that operational barrier considerably, potentially accelerating the pace at which larger companies ship their own production grade MCP integrations rather than relying on smaller, community built connectors.