Build1 publisher3 min readPublished
Subscription apps launched since 2025 share 3% of revenue, RevenueCat's 2026 report finds
RevenueCat finds apps launched since 2025 share 3% of subscription revenue while launches run at 14,700 a month. AI-written code has moved the hard part of a new app to finding paying users, though some of that gap is only because younger apps have had less time to build up renewals.
The Engineer · Build desk

What happened
- The launch count stood near 2,000 a month in January 2022, according to RevenueCat's analysis of Appfigures data.
- Consumer spending on apps rose 21.6% to $155.8 billion even as downloads fell, according to TechCrunch figures cited in the post.
- One year after launch, the median new subscription app earns about $72 a month, while the top 10% earn $2,574.
Compiled by The EngineerSomething wrong?How this is made
Why it matters
- decision A builder weighing a new subscription app is planning against a median of about $72 a month at one year. Whether to build now depends on a credible plan for reaching paying users, because building is no longer the hard step.
- constraint Each of the thousands of monthly launches competes for installs from a total about 21% below its 2020 peak, so a new app mostly has to win downloads away from apps already on the phone.
- cost For AI apps, faster churn absorbs most of the 41% per-user price premium. Under constant churn, lifetime revenue per payer ends up only about 8% higher.
The post's model gives every app four steps: someone builds it, people find it, a stranger trusts it with their data and their card, and money moves [15]. AI coding tools shortened the first step. Jacob Eiting, RevenueCat's CEO, wrote in the report's founder's letter: "AI removed a decade old supply constraint on apps... this is just a shock." [6]
Launches went from about 2,000 a month in January 2022 [2] to more than 14,700 in January 2026 [1], a factor of about 7.4 [1]. The post dates the upturn to early 2025 and ties it to AI coding tools getting good [5]. The evidence for that link is the timing of a bend in a chart.
Demand moved the other way. Appfigures counted 106.9 billion downloads in 2025, down 2.7% from 2024 and the fifth annual decline in a row [7]. The 2020 peak was 135 billion [8]. Installs are now about 21% below it [2]. Consumer spending still rose 21.6% to $155.8 billion, according to TechCrunch figures the post cites [9]. The post takes this to mean the money is going to apps people already have [17].
The 69% share for pre-2020 apps comes from RevenueCat's chart as posted by a developer on X [4]. The split needs an age correction before it proves new apps cannot find customers. Every app in the 3% group launched in 2025 or later [3]. Every app in the 69% group had been live at least six years by 2026 [6]. A subscription app's monthly revenue is its new payers plus every payer still renewing from earlier months. A young cohort's small share is therefore partly a measure of its age. For the split to show a harder market, the 2025 cohort would have to trail the 2019 cohort at the same age. The post does not publish that comparison.
I think the fixed-age numbers make the stronger case that distribution is the limit. One year after launch, the median app makes about $72 a month, the top quarter $429 and the top 10% $2,574 [10]. The top decile earns about 36 times the median [3]. Within two years, 17.3% of apps reach $1,000 a month and 4.6% reach $10,000 [11]. In 2014, VisionMobile surveyed more than 10,000 developers, set an "app poverty line" at $500 per app per month, and found half of iOS developers below it [16]. The samples and methods differ, but today's one-year median is about a seventh of that line [4].
AI apps earn 41% more per paying user and lose them 30% faster, according to the post [12]. Take "30% faster" to mean a monthly churn rate 1.3 times the baseline, and hold that rate constant. Expected subscriber lifetime then falls to 1/1.3 of normal. Lifetime revenue per payer comes out about 8% ahead (1.41 / 1.3 = 1.085) [5].
The report covers 115,000 apps, $16 billion in revenue and over a billion transactions [14]. The build step has its own revenue line. Cursor passed $2 billion in annualized revenue and Lovable $500 million, according to the post [13]. Cursor's figure alone equals a year of revenue from about 2.3 million median one-year-old apps [7].
What to watch
- A same-age cohort comparison from RevenueCat or Appfigures, such as 2019 launches in their first year against 2025 launches in theirs, would separate tenure from difficulty in the 3% figure.
- Appfigures' 2026 download total: a sixth straight decline while launches stay above 14,700 a month would further shrink the install pool new apps draw from.
- RevenueCat's next report on AI-app churn: if it falls to the non-AI rate, the 41% per-user premium turns into real lifetime revenue.