In a significant discussion at the recent MCP Dev Summit held in New York, key maintainers of the Model Context Protocol (MCP) from leading technology organizations, including Anthropic, AWS, Microsoft, and OpenAI, provided a comprehensive update on the protocol’s development and its strategic direction. The panel, hosted by The Linux Foundation’s Agentic AI Foundation (AAIF), aimed to reassure stakeholders and the broader developer community about the protocol’s future, emphasizing its robust governance and its increasing importance in addressing critical enterprise requirements for security, reliability, and governance in the rapidly evolving field of agentic artificial intelligence.
The AAIF, which began its journey in December with initial contributions of MCP, the Goose project, and AGENTS.md, has experienced remarkable growth, now boasting 170 members. MCP has emerged as the flagship project within the foundation, rapidly establishing itself as an industry standard for enabling AI agents to seamlessly connect with data sources and applications. This swift adoption underscores the growing demand for standardized, interoperable solutions in the AI landscape.
Evolution and Governance: A Foundation for Trust
A core message from the summit was the continuity in the project’s governance model. Maintainers reiterated that MCP continues to operate as a bottom-up, open-source project, a structure that has fostered trust and innovation. The AAIF’s role is not to dictate protocol development but to serve as a crucial bridge, connecting the technical evolution of MCP with the practical needs and stringent requirements of enterprise users. This feedback loop is instrumental in ensuring that MCP is production-ready and addresses the complex challenges faced by organizations deploying AI at scale.
Claire Liguori, Sr. Principal Engineer at AWS, highlighted the palpable enthusiasm from customers regarding the Foundation’s neutral stance. "We see customers excited about the Foundation and about this being a neutral place to work on MCP and related projects," Liguori stated. "It’s great to be around the community and work within the entire developer ecosystem, and not just within our own companies." This sentiment reflects a broader industry trend towards collaborative development of foundational AI technologies in neutral, community-driven environments.
Addressing Enterprise Demands: Security, Scalability, and Governance
The widespread adoption of MCP has illuminated specific areas ripe for enhancement, particularly concerning the needs of enterprise applications. Maintainers acknowledged that the protocol’s evolution must prioritize strict security protocols, robust scalability to handle massive data flows, unwavering reliability for mission-critical operations, and comprehensive governance frameworks to ensure compliance and accountability.
Nick Cooper, Technical Staff at OpenAI, articulated the vision for MCP’s role within the burgeoning agentic AI ecosystem. "MCP is the seed," Cooper explained. "The foundation has a broad mandate beyond just MCP. It’s open to new protocols and technologies, just like the early Cloud Native Computing Foundation (CNCF) was. But MCP itself should stay narrow: Connecting AI to data sources. Identity, observability, and governance should come in as other projects." This strategic focus ensures that MCP remains a specialized, high-impact protocol, while the AAIF can foster a broader ecosystem of complementary projects.
The AAIF is actively soliciting new project proposals related to agentic AI, a process that Cooper emphasized requires careful curation. "We need to be careful that the first accepted projects set the right direction," he noted, underscoring the importance of strategic alignment in the foundation’s early stages.
David Soria Para, Technical Staff at Anthropic and a co-creator of MCP, echoed the sentiment regarding the critical need for robust security and authorization mechanisms. "We see open challenges in security and authorization, and we’re happy to have AAIF bring the industry together and talk about the right solutions," Para commented. He further elaborated that authorization has been a focal point of intense development within the MCP specification over the past year, with ongoing collaborations with industry leaders like Okta to enhance authentication capabilities. However, Para cautioned that no single protocol can solve all security challenges; a holistic evolution of the surrounding ecosystem, including gateways, registries, sandboxing, and interceptors, is paramount.
The Growth Trajectory of MCP
Stephen O’Grady of RedMonk, who moderated the panel, provided striking data on MCP’s adoption rate. He declared MCP to be the fastest-growing standard RedMonk has ever tracked, citing an analogy that it took Docker approximately 13 months to achieve a similar level of establishment that MCP has reached in roughly 13 weeks. This rapid ascent is a testament to the urgent need for standardized communication protocols in the AI domain.
Interoperability and Future Convergence: MCP and A2A
In the evolving landscape of agentic AI, the Agent2Agent (A2A) protocol has also emerged as a significant standard, focusing on enabling direct communication between AI agents. The panel addressed the relationship between MCP and A2A, clarifying that they are not in direct competition but are rather complementary large protocols that can learn from each other. While future convergence is a possibility, it is not a certainty, as the current approaches differ slightly. "Approaches are slightly different at the moment," noted Para, "But we are open to anything that makes the industry easier to work with through open standards." This openness signifies a commitment to industry-wide interoperability and the advancement of open standards.
Addressing Misconceptions: CLI vs. Protocol
A recent social media debate questioned the relevance of MCP, with some claiming it was obsolete due to the availability of command-line interfaces (CLIs) with comparable functionality. Catie McCaffrey, Partner Software Engineer at Microsoft, addressed this misconception by framing CLIs and protocols as complementary tools serving different purposes and developer experiences.
"We ship APIs, we ship SDKs, and we ship CLIs all to interact with Azure for a concrete experience with Microsoft, and that’s because we want to meet developers where there are, and we want to meet them at the scenario that they’re working in," McCaffrey explained. "For local development scenarios, having an agent just interact with the Azure CLI or the GitHub CLI is a really wonderful use case." This perspective highlights that CLIs offer a valuable, direct interaction method for specific development workflows, while MCP provides a standardized, scalable mechanism for broader system integration.
Cooper reinforced this point, emphasizing the enduring value of MCP’s core function: "The focus of MCP going forward has to be on its utility in connecting things. MCP can evolve as long as it preserves the utility of what’s important. Where the value lies for me is that there’s real utility in using MCP to connect these different systems. MCP should grow, evolve, and focus on that. And that’s why it’s important to behave neutrally and focus on the utility that MCP is delivering." This strategic direction ensures that MCP remains focused on its primary objective of facilitating robust agent-to-data and agent-to-application connections.
Enhancing Developer Experience and Best Practices
The panel also acknowledged the need for continued attention to the MCP client and the dissemination of best practices. Maintainers stressed the importance of clear documentation and communication to guide developers in effectively utilizing the protocol. A common anti-pattern identified was simply wrapping hundreds of API endpoints, which can lead to suboptimal agent behavior. Instead, the recommendation is to design MCP interfaces with the agent as a new class of consumer, distinct from traditional developer-to-developer interactions. This approach emphasizes the creation of carefully designed server interfaces rather than naive API wrappers, thereby enhancing the quality and utility of agentic interactions.
The MCP Dev Summit served as a critical forum for reaffirming the protocol’s strategic importance and its commitment to evolving alongside the dynamic agentic AI landscape. With a strong foundation, active community involvement, and a clear focus on enterprise needs, MCP is well-positioned to remain a pivotal standard in the development of intelligent, interconnected AI systems. The AAIF’s role as a neutral steward of this technology, fostering collaboration and addressing industry challenges, further solidifies confidence in the protocol’s long-term viability and impact.