At the MCP Dev Summit in New York, maintainers from Anthropic, AWS, Microsoft, and OpenAI discussed the future of the Model Context Protocol (MCP) under the Agentic AI Foundation (AAIF). Key themes included enterprise security, authorization improvements (in collaboration with Okta), governance, and reliability. The panel emphasized MCP should stay narrowly focused on connecting AI to data sources, while identity, observability, and governance should be handled by separate projects. They addressed the 'MCP is dead' narrative, noting both MCP and CLI approaches serve different use cases. The panel also highlighted the fastest-growing standard RedMonk has ever tracked — MCP reached Docker-level adoption in 13 weeks vs. 13 months. The A2A protocol was discussed as complementary rather than competing. A key anti-pattern flagged: naively wrapping 500 API endpoints instead of designing MCP interfaces specifically for AI agents.
Sort: