How Runlayer is Scaling Enterprise MCP Infrastructure

Runlayer, making enterprise AI secure, compliant, and fully connected, led by Andy Berman, Tal Peretz, and Vitor Balocco, has secured $11 million seed round from Khosla Ventures and Felicis.

Teams are increasingly wiring their preferred AI tools directly into production databases via MCP, just like thousands of others. But with more than 18,000 MCP servers in the wild, an estimated 10% are outright malicious, and most of the rest remain exploitable.

Enterprises face a dangerous gap: no visibility into these connections, no guardrails to enforce, and no way to curb the risk without halting AI adoption altogether.

Teams are increasingly connecting AI to production databases, internal APIs, and core company systems to drive productivity, leaving them with an impossible choice: block MCP entirely or accept the resulting chaos.

With the model race settled, the real contest now centers on controlling how AI interfaces with the world’s tools and data.

Runlayer establishes Enterprise MCP as the secure, scalable path for production AI. It combines high-accuracy threat detection, fine-grained IdP-based permissions, full observability, and a vetted MCP catalog, deployable in-cloud or in-VPC.

AI-native companies like Gusto and Opendoor already rely on it, supported by advisors including the Co-Creator of MCP at Anthropic and the Head of Security at Cursor.

As MCP becomes the backbone of enterprise AI, just as APIs defined the previous era, Runlayer provides the infrastructure needed to deploy AI agents without chaos.

Runlayer is reshaping how AI operates, shifting from passive answers to real actions inside existing tools. The company is scaling MCP as the foundation for safe, capable AI agents for every enterprise.