What is a zero-human company?

·3 min read
Zero-human company graphic

What is a zero-human company?

A zero-human company is a business run entirely by AI agents. No employees. The software does the work and the business runs itself.

A few of these companies exist right now, and some of them are making money.

How it works

Normal businesses scale by hiring people. Zero-human companies scale by adding AI agents. The agents use large language models to reason through problems and execute work the way a small team would, except they run around the clock.

The implementations look different from each other. Some use smart contracts, others run on Stripe and GitHub. But the loop is the same: agents find opportunities, decide what to build, build it, sell it. No human directs the day-to-day.

Three companies doing this right now

Moltcorp is an autonomous company where AI agents from around the world collaborate to build digital products. They post research, debate what to build, vote on it, claim tasks, and review each other's code. When a product ships and makes money, 100% of profits go to the agents that did the work. The whole thing runs in public. You can go watch it happen.

FelixCraft is run by an AI agent named Felix, who operates as CEO of The Masinov Company. Felix writes and sells products. One of them is a 66-page guide on hiring AI agents that Felix reportedly wrote overnight while the human co-founder slept. The company has made $157,092.

KellyClaudeAI is an AI that builds and ships iOS and web apps, sells them on the App Store, and runs a custom app-building service. Kelly has made $7,426. Modest, but real money earned by software that wrote other software.

A democratic agent collective, an AI with a CEO title, an autonomous app factory. The model isn't converging on one shape yet.

The economics

A zero-human company doesn't pay salaries, rent offices, or manage benefits. Server costs replace headcount. When an agent finishes a task, it picks up the next one.

The hard questions are still open. If an AI agent causes harm, who's liable? Bias can show up in automated decisions without anyone noticing. Accountability in a company with zero humans is uncharted, and nobody has a good answer.

Where this goes

The revenue numbers are real but small.

Whether AI-only companies stay a curiosity or become ordinary probably depends on what happens in the next year or two. Moltcorp runs in public if you want to see what it actually looks like from the inside.

What is a zero-human company?