The Surprising Companies Building MCPs Right Now

There’s a lot of talk about Model Context Protocols (MCPs) right now.

At Tadata, we have a front-row seat. As the creators of the open source project FastAPI-MCP and the team behind the Tadata platform, we’ve seen over 2,000 organizations engage with our tools to create and host MCPs. That gives us a unique window into who's actually building, what they're building, and why.

And some of what we’re seeing may surprise you.

It’s not just startups.

Sure, there are plenty of scrappy, AI-native startups using MCPs to power copilots and context layers. But the adoption curve is far wider than that.

12% of the organizations engaging with Tadata are 10,000+ person companies.

In some ways, it makes perfect sense—large companies have more developers, and developers are hungry to explore the latest AI tooling. The big question is whether these enterprises can go beyond prototypes—whether they can get MCPs into production, scale them, and manage the complexity of security, compliance, and change management along the way.

It’s not just tech companies.

When you think of early AI infra adopters, you probably think of companies like Wiz, Scale AI, or MongoDB. And yes, they’re all using our tools.

But some of the most active MCP builders are companies you wouldn’t typically associate with bleeding-edge tech:

While tech-native companies can adopt MCP faster thanks to their modern tech stacks, the upside can be greater for legacy organizations. MCP servers are helping these companies leapfrog decades of accrued complexity and get value from AI, fast.

It’s not just for external APIs.

A lot of the hype around MCPs has focused on public-facing endpoints—“turn your OpenAPI spec into an AI agent!”

But we’re seeing just as much momentum around internal use cases.

One of my favorite stories: a developer at Cisco started using FastAPI-MCP for an internal tool. A few weeks later, he found out that someone else in a peer team at Cisco was also using the project – they had each discovered it independently.

While we certainly have CEOs, CTOs, and CPOs driving tops-down MCP adoption initiatives to avoid being leapfrogged by AI-native startups, there’s also a huge wave of bottoms-up AI infra adoption.

So what does this all mean?

MCP servers are becoming commonplace across companies, industries, and use cases–fast.

They’re being used to close gaps between product and support, developer and data, team and tool. They're quietly becoming the connective layer for AI in the enterprise.

And if you’re exploring how to get started yourself, you don’t have to reinvent the wheel.

Two good entry points:

• Tadata: Convert any OpenAPI spec into a hosted, authenticated MCP server with built-in analytics—no infra work required.
• FastAPI-MCP: Our open-source converter for FastAPI apps. Lightweight, flexible, and free to use.

Whether you're prototyping an internal copilot or deploying an AI-facing interface for your customers, it’s easy to stand up an MCP in minutes.