The split is wrong

The split is wrong.

We still talk about AI as enterprise or consumer. The money didn’t follow that line. It went upstream, to the people building things.

Menlo Ventures found 27% of AI application spend now comes in through one person signing up first, nearly four times the rate in traditional software. Count the people on personal cards and it may be close to 40%. Cursor reached 200 million in revenue before it hired a single enterprise sales rep. Lovable went from its first million a year to 100 million in eight months. Replit’s annualized revenue grew from 2.8 million to 150 million dollars in under a year.

Those customers are indie developers, small studios, a designer at a big brand with a personal account. They found the tool, used it, then dragged their company in after them. It isn’t only at work: in Menlo’s survey with Morning Consult this summer, 55% of American AI users paid for at least one AI product, and 48% of those payers covered it entirely themselves.

Retail has seen this before. The early adopter is rarely head office. It’s the person on the floor who figured it out first.

Who’s the builder in your company, and do they know you know?

What are your thoughts?