The enterprise is on the cusp of its next great technological transformation with AI. As organizations move beyond initial experimentation to truly scale AI across their operations, a critical challenge emerges: how to manage, secure, and govern the complex web of AI tools, agents, and data interactions. This is precisely the problem Obot is poised to solve, and why we at Mayfield are thrilled to co-lead their $35 million seed round alongside Nexus.
The Model Context Protocol (MCP) is rapidly solidifying its position as the standard for how AI tools and agents connect to applications, systems, and data within an enterprise. Think of it as the connective tissue that allows AI to truly integrate and operate across your existing infrastructure. However, with this rapid adoption comes a familiar set of enterprise needs:
· Security: How do you ensure AI agents access data and systems securely, adhering to internal policies and external regulations?
· Governance: Who can deploy what AI models? How is their usage monitored? What data are they interacting with?
· Visibility: IT and business leaders need a comprehensive view into all AI operations to ensure compliance and optimize performance.
· Fragmentation: Without a centralized approach, enterprises risk a proliferation of “shadow AI” – unmanaged AI initiatives that introduce significant risks and inefficiencies.
This is where Obot comes in.
Our vision, aligned with Sheng’s team, is for Obot to become the enterprise control plane for MCP and, by extension, for comprehensive AI management. Just as Rancher created the standard control plane for Kubernetes adoption, Obot will provide the essential infrastructure for enterprises to embrace MCP with confidence.
Here’s how Obot is addressing these critical enterprise needs:
1. Obot MCP Gateway : Unified Control and Security
The Obot MCP Gateway is an open-source control plane designed to provide IT teams with unprecedented visibility and control over their MCP deployments.
· Unified Management: IT can onboard MCPs from various sources, whether hosted internally or brokered remotely.
· Access Controls & Governance: granularly set access policies, ensuring that only authorized users and agents interact with specific MCPs and the data they access.
· Visibility & Security: All communication is proxied through the Gateway, providing a single point for auditing, logging, and applying security policies. This eliminates shadow AI and ensures compliance.
· Employee Catalog: Employees gain access to a trusted catalog of enterprise-authorized MCPs, simplifying discovery and secure interaction.
· Rich User Experience: The integrated web-based chat client allows users to connect with authorized LLMs, interact directly with one or more MCPs, and even configure sophisticated multi-MCP workflows, all within a secure and governed environment.
2. Nanobot MCP Agent Framework (Introduced Last Week): Empowering Developers
Complementing the Gateway, the Nanobot MCP Agent Framework provides developers with the “runtime + SDK” needed to build powerful, MCP-native AI agents. This framework embraces MCP + MCP-UI, enabling agents to deliver rich user interfaces directly within chat environments. This is a game-changer for agent developers, making the creation of sophisticated, interactive AI agents practical and scalable within the enterprise context.
The timing for Obot couldn’t be more critical. Enterprises want to rapidly move from proof-of-concept AI projects to full-scale deployment. Without a robust control plane like Obot, this transition will be fraught with challenges: security vulnerabilities, governance gaps, and a fragmented AI landscape that stifles innovation rather than accelerates it.
Obot provides the standard, trusted pathway for enterprises to adopt MCP with the same level of governance, security, and operational confidence they expect from all their critical infrastructure.
Sheng and his team have a unique ability to anticipate the fundamental enterprise needs of new technology waves and build the open-source infrastructure that becomes indispensable. Their past successes with Cloud.com and Rancher Labs are not just impressive product achievements; they represent new categories of enterprise control planes that enabled enterprise adoption of new paradigms. They also had the largest exits of their respective categories at that time. We believe they are poised to repeat this success in the realm of AI with MCP. Mayfield is proud to partner with Sheng and his team on the Obot journey to define the future of enterprise AI.
