The Economics of AI Business Ownership
The Economics of AI Business Ownership
A practitioner's look at the real P&L of running AI businesses at different scales.
The shape of the cost curve
An AI business at $1K MRR typically runs on ~$120/mo of infrastructure. Gross margin: ~88%.
At $10K MRR, costs rise non-linearly because of human oversight, observability, and client-side integrations. Gross margin: ~72%.
At $50K MRR, costs stabilize again because you've automated most human oversight. Gross margin: ~80%.
Where money actually goes
- LLM inference: 30–40%
- Payments + bank: 8–12%
- Observability + infra: 10–15%
- Human oversight / agencies: variable — usually 10–25%
- Marketing and acquisition: 15–30% depending on growth stage
Valuation math
Multiples compress as MRR climbs — but so does predictability. A stable $10K MRR business with 2% churn will routinely trade at 22x. A flashier $10K business with 10% churn trades at 12x.
Why most portfolio operators cap at 4 businesses
Cognitive load. Four WhatsApp briefings is the ceiling for most humans. Above that, you're building an agency, not running a portfolio — and agency economics are different.
Bottom line
The AI business economy is more similar to SaaS than to e-commerce. It rewards patience, documentation, and the willingness to own something small for longer than feels exciting.