App installs stopped growing.
Spending did not.

Feature 14 of this edition ended with a table containing three deliberately blank rows. One of them read "Download growth rate 2024→2025 - not obtained", and the accompanying note said the missing figure was the single most useful number absent from the analysis, because without it there was no way to tell whether app store spending grew from more users or more money per user.

The number has since been published. Downloads across iOS and Google Play rose 0.8% in 2025, to nearly 150 billion. Store revenue over the same period rose 10.6%, to $167 billion.¹

That is the decomposition. It is not close.

Figure 01

Volume against price

Year-over-year growth, 2025, across iOS and Google Play.

Source: Sensor Tower, State of Mobile 2026, published January 2026, covering calendar 2025. All five series are the vendor's; the juxtaposition is ours. Conflict: Sensor Tower sells mobile market intelligence and has a commercial interest in the market appearing measurable and consequential. These are its own panel estimates, not audited figures from the platforms.

What a thirteen-to-one ratio means

Revenue grew roughly thirteen times faster than installs. Run the arithmetic the other way and revenue per download rose about 9.7% while the number of downloads was essentially unchanged.

Nothing about 2025 was an acquisition story. Almost the entire increase came from people who had already installed something.

This has a consequence that most app growth plans are not written for. If total installs are flat, the install market is a fixed pool. An app that grows its install base in a flat market is not expanding the category - it is taking share. Every install won is an install another product lost.

User acquisition in a growing market and user acquisition in a flat one are different activities with different economics. In the first, spending more brings in genuinely new demand. In the second, spending more bids against competitors for a supply that does not respond to price. Feature 01 of this edition documents what happens to ad prices when demand rises against fixed supply, and the mechanism here is the same one.

Time spent did rise - 5.3 trillion hours across both stores, up 3.8%, or roughly 3.6 hours per day per mobile user.¹ So engagement grew faster than installs and slower than revenue. Existing users are spending somewhat more time and considerably more money.

The crossover nobody announced

Underneath the aggregate is a structural change. In 2025, in-app purchase revenue from non-game apps exceeded revenue from games for the first time.¹

Games IAP revenue approached $82 billion, up 1.3% - a third consecutive year of growth, and a recovery from the correction that followed the 2021 peak. Non-game IAP revenue grew 21%, reaching nearly three times what it was five years earlier.¹

Figure 02

The crossover

In-app purchase revenue by category, 2025, iOS and Google Play combined.

Source: Sensor Tower, State of Mobile 2026. The games figure of approximately $82bn is the vendor's. The non-game figure of approximately $85bn is our subtraction from the $167bn total, because the source states the crossover and the games number but does not print the non-game total. The gap between the two bars is therefore a residual and should be read as approximate; the crossover itself is the source's claim, not our inference.

Eighteen years into app stores, the thing they were mainly for stopped being the thing they mainly earn from. Games did not shrink. They stopped compounding, and something else did not.

Sensor Tower's own framing of the games position is worth quoting for its precision: with installs down and time spent up, growth depends less on adding users and more on expanding lifetime value from the existing base.¹ That is a description of a mature category. It now also describes the whole market.

Where the 0.8% came from

A flat aggregate can hide a great deal of movement underneath it, and in 2025 it did.

Downloads of generative AI apps doubled year over year to reach 3.8 billion. In-app purchase revenue from those apps nearly tripled to exceed $5 billion. Time spent in them reached 48 billion hours, roughly 3.6 times the 2024 total and close to ten times the 2023 level, and sessions passed one trillion.¹

Now put the download figure against the market. A 0.8% increase on a base of roughly 150 billion is a net gain of about 1.2 billion downloads. Generative AI apps, doubling from roughly 1.9 billion to 3.8 billion, added about 1.9 billion on their own.

One category appears to have supplied more than the whole market's net download growth, which would mean everything else, taken together, slipped.

We would not push that further than the arithmetic supports. Two rounded figures divided into each other produce a fragile result, and a small error in either the base or the market total flips the sign of the residual. But the direction is not fragile: the growth in installs was concentrated in one category, and that category is nine years younger than the app stores it arrived in.

Feature 47 of this edition examines what happens to those users afterwards.

Where the money is

The United States remained the largest market by revenue, with consumers spending nearly $60 billion in 2025 - roughly 36% of the global total by our arithmetic. US time spent rose 4%, a modest rebound after signs of fatigue in 2024. Western European growth was led by the United Kingdom, Germany and France.¹

What follows from a flat install market

Stop reporting installs as a growth metric. In a market where total installs moved 0.8%, an install number tells you about share and nothing about expansion. Revenue per install, payback period and cohort value are the series that carry information now.

Price your acquisition against a fixed pool. If category installs are flat, competitor bidding is the dominant force in your acquisition cost, not category growth. That changes the payback assumption underneath every campaign plan.

Look at the crossover before you assume your category's ceiling. Non-game revenue grew 21% and tripled over five years. Whatever ceiling a category was assumed to have, the aggregate says the constraint is monetisation design rather than audience size.

Do your own volume-versus-price split every quarter. It takes an afternoon, and it determines whether your growth plan should be about reach or about depth. At market level the answer for 2025 is unambiguous, and it is depth.

Read the games number as a preview, not a laggard. Games hit flat installs first and responded by shifting to lifetime-value expansion. The rest of the market arrived at the same position in 2025. The playbook already exists in the category everyone stopped paying attention to.