Answer · data from 744,876 iOS apps
How many new apps actually get users?
Almost none — and the odds worsen every year. Of the 149,497 iOS apps released in 2025 that survive in our July 9, 2026 dataset of 744,876 apps, 65.2% have zero ratings, only 4.7% ever reached 50 ratings, and 1.0% reached 1,000. In the fixed week of July 3–9, 2026, just 5.2% of that cohort added even one new rating — roughly 19 out of 20 apps released since January 2025 gained none. The median app released in any year since 2022 has zero ratings, period.
Ratings are the only traction signal observable outside Apple — a proxy for users, not a count of them; biases are disclosed below. Data snapshot: July 9, 2026; dataset updated daily. Check any individual app's estimated traction — free — with the app revenue checker.
The context — July 2026
A release boom with no matching usage boom
In July 2026, a16z circulated an FT graphic (by John Burn-Murdoch) of data from NBER Working Paper 35275 — “Writing Code vs. Shipping Code” by Demirer, Musolff & Yang — showing monthly iOS app releases at roughly 180 against a 2024-average=100 index, while apps with significant usage stayed flat (drifting slightly below 100) and total App Store review volume declined. The paper's production-side finding explains the mechanism: across 100,000+ GitHub developers, AI coding tools raised commit output dramatically, but the gains fall to ~30% by the time you count released software — writing an app got cheap; making it wanted did not.
Our dataset shows the same boom independently: surviving 2025 releases (149,497) outnumber 2024's (103,604) by about 44% — see how many apps are released on iOS every day. This page quantifies the other blade of the scissors: what happens to those releases after launch.
The measurement
Traction by release cohort, in a mid-2026 snapshot
Each row is the cohort of apps released that year that are still live in our 744,876-app dataset on July 9, 2026. The last column is the age-fair one: it asks the same question of every cohort over the same week — did this app add even one rating between July 3–9, 2026?
| Release year | Apps live today | Median ratings | Zero ratings | Ever reached 50 | Ever reached 1,000 | Added ≥1 rating, Jul 3–9 ’26 |
|---|---|---|---|---|---|---|
| 2015 | 20,958 | 4 | 26.2% | 25.5% | 9.2% | 11.3% |
| 2016 | 28,617 | 3 | 30.2% | 21.5% | 7.7% | 9.7% |
| 2017 | 31,013 | 3 | 30.6% | 22.0% | 7.9% | 9.6% |
| 2018 | 32,421 | 3 | 32.0% | 20.7% | 7.7% | 9.4% |
| 2019 | 38,875 | 2 | 34.6% | 19.5% | 7.4% | 9.6% |
| 2020 | 49,947 | 2 | 38.4% | 17.0% | 5.7% | 8.3% |
| 2021 | 55,529 | 1 | 45.0% | 14.1% | 4.4% | 7.6% |
| 2022 | 65,401 | 0 | 50.8% | 11.3% | 3.3% | 6.7% |
| 2023 | 87,662 | 0 | 56.1% | 9.1% | 2.4% | 6.1% |
| 2024 | 103,604 | 0 | 61.2% | 7.2% | 1.8% | 5.7% |
| 2025 | 149,497 | 0 | 65.2% | 4.7% | 1.0% | 5.2% |
| 2026 (H1, lag) | 25,375 | 0 | 73.5% | 3.3% | 0.8% | 5.4% |
Source: GetAppNiche analysis of 744,876 iOS App Store apps, July 9, 2026 snapshot; release dates are exact store metadata. Read the cumulative columns (“ever reached…”) with the age caveat below; the weekly-activity column compares all cohorts over one identical window.
The trend
More releases, thinner traction — every year
The two curves move in opposite directions with no exception in our data. Surviving releases per year roughly tripled from 2020 to 2025 (49,947 → 149,497), while the share of each cohort that ever reached 50 ratings fell monotonically — 17.0% (2020) → 11.3% (2022) → 7.2% (2024) → 4.7% (2025). Even the zero-bar weekly signal thins out: 8.3% of surviving 2020 releases added a rating in the July 3–9 week, versus 5.2% of 2025 releases. Survivorship bias flatters the older cohorts, so the true decline is at least this steep. More competitors are dividing an attention pool that — per the FT/NBER usage series — is not growing.
Read this before quoting
What these figures can and cannot say
- Ratings are a proxy, not usage. Nobody outside Apple can see installs or active users. Public ratings are the closest observable signal: an app whose rating count never moves is, with rare exceptions, an app without a meaningful user base. All traction figures here are rating-based and labeled as such.
- Cumulative columns favor old cohorts. A 2015 release has had eleven years to accumulate 50 ratings; a 2025 release has had months. The “ever reached 50 / 1,000 ratings” columns therefore overstate old cohorts relative to new ones. The age-fair column is the last one — did the app add even one rating in the same fixed week (July 3–9, 2026)? — where every cohort is measured over an identical window.
- Survivorship. Apps removed from the store vanish from a today-snapshot, and older cohorts have lost more members. Surviving old apps are disproportionately the ones that worked, which flatters their numbers. This bias works against our headline finding, not for it — and the decline shows up anyway.
- 2026 discovery lag. Very recent releases take time to enter any tracker's dataset, including ours. The 2026 row is an undercount of releases and its traction shares will shift as the cohort ages — read it as an early sample, not a verdict.
None of this is an official Apple statistic: Apple publishes no per-app usage, install, or removal data. Our dataset is a 744,876-app actively-tracked research subset of the full store — see how many apps are in the App Store for how it relates to the ~1.9–2.4M total.
What to do with this
Launching into the boom without joining the graveyard
When writing code is cheap and attention is fixed, the scarce skill moves from building to picking the market. The odds are not uniform: in our State of iOS App Niches 2026 study, the odds that an app clears an estimated $1k/month differ by 160x between categories — and most apps earn nothing precisely because they launch where demand is already saturated. Researching which niches still convert releases into revenue — before writing a line of code — is the whole premise of GetAppNiche: 1.1M+ tracked iOS apps, daily estimated revenue/download data, and niche-level odds. Plans and a 3-day trial are on the pricing page.
Want the list, not the theory? See which iOS niches still turn a launch into revenue.
Free · sent straight to your inbox.
The iOS niches where launches still earn
Free — in your inbox in a minute.
- The niches from our State of iOS App Niches 2026 study, ranked by the odds a launch clears est. $1k/month — CSV
- Plus the Monday drop of 10 iOS niches worth building in
Free · arrives in a minute.
Check your inbox — the CSV is on the way.
While it lands: this table is a preview. The full access database behind it is 1.1M+ iOS apps — each with pricing model, estimated revenue and downloads, keyword difficulty and reviews. Free to explore for 3 days.
Full access database — free for 3 days →FAQ
New-app traction — quick answers
How many new apps actually get users?
Very few, by the only public signal available (ratings). Of 149,497 iOS apps released in 2025 that survive in a 744,876-app dataset as of July 9, 2026, 65.2% have zero ratings, only 4.7% ever reached 50 ratings, and 1.0% reached 1,000. In the fixed week of July 3–9, 2026, just 5.2% of the 2025 cohort added even one new rating — roughly 19 out of 20 apps released since January 2025 gained none that week.
What percentage of new apps get zero ratings?
The zero-rating share has risen every single release year in our data: 26.2% of surviving 2015 releases have no ratings at all, versus 50.8% for 2022, 61.2% for 2024, 65.2% for 2025, and 73.5% for the early-2026 cohort so far (July 9, 2026 snapshot of 744,876 iOS apps). Survivorship bias flatters older cohorts, but the direction is unambiguous: the median app released since 2022 has zero ratings.
Is the App Store oversaturated in 2026?
Release volume is accelerating faster than traction. Surviving 2025 releases (149,497) outnumber 2024's (103,604) by about 44%, yet the share of a cohort that ever reaches 50 ratings has fallen monotonically from 25.5% (2015) to 4.7% (2025). An FT graphic of NBER Working Paper 35275 data, circulated by a16z in July 2026, shows the same scissors at store level: monthly iOS releases at ~180 vs a 2024=100 index, while apps with significant usage stayed flat and total review volume drifted down. More apps are competing for a user-attention pool that is not growing.
Did AI coding tools cause the app release boom?
That is the leading explanation, with direct evidence on the production side. NBER Working Paper 35275 (Demirer, Musolff & Yang, 2026, “Writing Code vs. Shipping Code”) tracked 100,000+ GitHub developers and found AI tools raised coding output dramatically (cumulative effects up to ~180% on commits) — but the gains shrink to ~30% for actual software releases, and the FT/a16z chart built on the paper's App Store data shows released-app volume surging while usage stays flat. In short: AI made writing and shipping an app much cheaper; it did nothing to make any given app wanted.
How do I launch an app that actually gets traction in 2026?
The data points one way: distribution-aware niche selection before code. With ~410+ apps released per day and under 5% of recent cohorts reaching even 50 ratings, building in a crowded category with no differentiated demand is statistically a zero-rating outcome. Study which niches still convert release into revenue (per-category odds differ by 160x in our State of iOS App Niches 2026 study), validate demand against apps that already monetize, and pick markets where the existing winners are weak — that research is exactly what app-intelligence tools exist for.