The App Store Hit an AI Supply Shock
New subscription apps went from 2,000 a month to 14,700. AI titles earn more per payer and churn faster. A builder’s guide to surviving the glut.
RevenueCat’s 2026 subscription report has the sentence I keep quoting to other indie developers: three years ago about 2,000 new subscription apps launched each month. By January 2026 that number was 14,700. AI removed the old supply constraint on shipping an app. Demand did not grow on the same curve.
The rest of the picture is a split screen. The top quarter of subscription apps grew about 80% year over year. The bottom quarter shrank 33%. Apps launched before 2020 still take 69% of subscription revenue. New is not the same as better.
AI sells. It does not automatically stay.
AI-powered apps generate about 41% more revenue per payer and churn roughly 30% faster. The first week is a demo of magic. The second month is a habit the product has not earned. That is the entire mobile SaaS problem in 2026.
- Distribution is now the scarce resource, not features.
- Weekly plans and hard paywalls are beating vague freemium.
- Localized paywalls and real experiments outperform a single global sheet.
- Customer acquisition cost rises when 15,000 cousins launch beside you.
How I would ship a small app into this market
I would not lead with 'powered by AI' on the icon. I would lead with a job: hold a streak, finish a puzzle, explain a bill. The model stays in the loop where it reduces work. The subscription has to survive a week when the novelty is gone.
- 1.Pick a narrow job a pre-2020 app still does badly.
- 2.Instrument trial-to-paid and day-35 retention before you add another model.
- 3.Charge for the outcome, not for tokens the user cannot see.
- 4.Plan for churn on day 8, not only conversion on day 1.
The glut is not a reason to stop shipping. It is a reason to stop shipping generic magic.