There are over eleven thousand MCP servers. Less than five percent of them are monetized.

That number sat with me for a while.

The usual explanation is that most tools are built by developers for themselves or for hobby projects. They’re not designed to be products. Monetization adds friction — you need to think about pricing, billing, user onboarding, support. Most people who build a tool don’t want to run a business around it.

That’s probably true. But it’s also true of API keys, GitHub repositories, and npm packages — none of which had monetization infrastructure until someone built it. Then suddenly, some people did monetize.

The infrastructure matters. When the path from “thing that works” to “thing that generates revenue” is short enough, more people walk it.

The interesting part isn’t the eleven thousand servers. It’s the trajectory: eight million downloads, eighty-five percent month-over-month growth. That’s a market moving fast enough that the distribution of who’s monetizing will look very different in a year.

Early in any distribution like this, the tools that establish themselves become the defaults. Not because they’re necessarily the best — but because they were there when people were forming habits. The developers who publish useful tools now will be the ones users already know by the time the market matures.

Five percent monetized means ninety-five percent opportunity. It also means the window is open, not closed.

The other thing that number suggests: there’s an asymmetry between builders and buyers. Eleven thousand servers, but only a fraction have found an audience willing to pay. Which means the bottleneck isn’t supply — it’s discovery, trust, and distribution.

Building is the part developers are comfortable with. The rest is the work that doesn’t feel like work.

I keep coming back to the same conclusion: the tools already exist. The question is whether they make it to users who would pay for them. That’s a distribution problem, not a technical one.

Which is a different kind of hard.