In 2025, mobile app downloads were broadly flat by Sensor Tower’s estimate while consumer spending and time spent rose; Appfigures, using a different estimation approach, reported falling downloads alongside a sharp rise in spend. The useful reading of the 2026 “state of mobile” reports is not that every app is booming, but that mature reach is being monetized differently and category behavior diverges.
Who should read what: everyday users can start with what growth and spending mean for useful daily apps; developers should focus on product and measurement choices; analysts should read the methodology and comparison sections before comparing totals.
Key takeaways
- Sensor Tower estimated almost 150 billion 2025 downloads across the App Store and Google Play, up 0.8% year over year; its estimate of store consumer spend reached $167 billion, up 10.6%.
- Appfigures estimated 106.9 billion downloads in 2025, down 2.7%, and $155.8 billion in consumer spend, up 21.6%. The two vendors do not measure identical universes or model data identically.
- Time spent is increasingly concentrated in habitual categories: Sensor Tower counted nearly 2.5 trillion hours on social apps in 2025.
- Growth is selective: AI and short-form drama downloads surged, while retail downloads and time spent declined in Sensor Tower’s 2025 category findings.
- Use these reports to frame questions and test local evidence, not to assume a global category trend predicts the success of an individual app.
What changed in mobile apps in 2025?
The January 2026 Sensor Tower report, “2026 State of Mobile: AI Moves Mobile into Its Next Phase”, describes a large but mature market. Its estimate says global downloads across iOS and Google Play edged up 0.8% to nearly 150 billion in 2025, while in-app purchases and paid apps and games reached $167 billion, a 10.6% annual increase. The report estimates 5.3 trillion hours across those stores, up 3.8%, or roughly 3.6 hours per day per mobile user.
These are estimates, not a census of every installation, payment, or minute on every device. The store scope matters: web use, advertising revenue, enterprise contracts, and activity on third-party Android marketplaces are not represented as equivalent store purchases. “Consumer spend” also describes gross consumer outlay rather than a developer’s net proceeds.
“Total downloads across iOS and Google Play edged up 0.8% YoY to nearly 150 billion”
— Sensor Tower, 2026 State of Mobile
Appfigures’ annual estimate, reported in TechCrunch’s “App downloads declined again in 2025, but consumer spending soared to nearly $156B”, puts 2025 downloads at 106.9 billion, down 2.7%, and consumer spending at $155.8 billion, up 21.6%. Its article says app downloads had declined for a fifth consecutive year. Those figures are useful as a second estimate, not a correction to Sensor Tower: differences in store coverage, definitions, deduplication, and modeling can move the totals and growth rates.
“global downloads of all mobile apps and mobile games via the App Store and Google Play reached an estimated 106.9 billion”
— Sarah Perez, TechCrunch
There is a striking gap between the vendors’ spend growth rates: 10.6% for Sensor Tower versus 21.6% for Appfigures, although both put spending at a very large scale. Rather than average the two into a false precision, keep each estimate attached to its publisher. Appfigures explicitly labels its totals estimates; Sensor Tower also describes its report as estimated app-market data. The safe conclusion is directional: both describe higher spending in 2025, while the magnitude is method-sensitive.
Everyday users: what does the trend mean for useful apps?
For people choosing apps, a growing category is a clue about attention and product availability, not proof of quality. AI, short-form entertainment, retail, finance, and food delivery can meet real needs, but the same category can contain excellent tools and disposable products. Look for transparent pricing, a clear reason to return, useful privacy controls, and a graceful free experience before treating chart momentum as a recommendation.
Sensor Tower’s press release, “Boosted by Gen-AI Services, Consumers Spent More Money in Apps than Games for First Time”, reports that non-game app in-app purchases overtook games for the first time in 2025. It says non-game IAP increased 21% year over year and was nearly three times its level five years earlier. That does not mean all non-game apps are paid, or that a category leader is representative of typical products.
“IAP revenue from non-gaming apps surpassed games for the first time”
— Sensor Tower, State of Mobile 2026 press release
Attention is another lens. Sensor Tower says social platforms accumulated nearly 2.5 trillion hours in 2025. That is an aggregate across a huge population and many services; it does not say that any one person spends that much time, nor does time alone tell us whether a service helped someone. For daily-life apps, repeated use can reflect convenience, necessity, habit, or friction in leaving. Users can audit notifications and subscriptions, and remove apps that no longer earn their place.
“Time spent on social media apps continued to increase in 2025, with nearly 2.5 Trillion combined hours spent on social platforms”
— Sensor Tower, State of Mobile 2026 press release
Developers: what should you build, ship, and measure?
For developers, the reports are a prompt to solve a recurring problem well, then prove that users receive value. A market-wide spend increase is not a forecast for a new product. Start with a specific audience and job: plan a commute, track a bill, learn a small skill, or take a short restorative break. Reduce setup steps, make the core action understandable, and let a user reach a useful outcome before asking for a subscription.
Sensor Tower’s category examples show why broad labels can mislead. In its 2025 data, credit and lending app downloads rose 18% year over year, while it reported declines in investing and financial management and cryptocurrency apps. Restaurant and food delivery apps grew 14%; global retail app downloads and time spent fell. The report attributes some retail decline to slower expansion by major global players, but that is an interpretation, not proof that the whole retail problem has one cause.
“Credit and Lending apps saw downloads climb 18% YoY”
— Sensor Tower, State of Mobile 2026 press release
Build a measurement plan around the product’s promise. For a practical utility, track successful task completion, time to first value, repeat use at meaningful intervals, cancellation reasons, and support issues. Separate organic acquisition from paid campaigns where your analytics allow it; label platform, country, store, new installs, and returning users. Monetization metrics need the same care: trial starts, refunds, renewal, ad exposure, and gross versus net receipts are different events.
Our experience making casual games at NabiGame Studio in Da Nang since 2016 is that a small, clear interaction deserves more attention than a fashionable category label. That is an opinion from a game studio, not a claim that puzzle design transfers unchanged to finance or health. If an app is a short daily break, for example, it should make that break feel complete; our Juice Fruit merge puzzle and Blot Up Ink color puzzle illustrate the kind of focused interaction we enjoy making. Developers working across mobile and browser distribution can also consider the practical tradeoffs in our HTML5 web games guide.
Analysts: how should you compare market estimates?
Start by writing down the unit and universe before comparing a chart. “Downloads” may mean first-time installs, may handle reinstalls differently, and may cover only named storefronts. “Spend” may include paid downloads and in-app purchases but exclude ads, direct web payments, or enterprise sales. Time estimates may be modeled from panels or other signals rather than directly observed for every device.
A useful comparison checklist is:
- Match the period. Compare calendar 2025 with calendar 2025, not a fiscal year or a rolling twelve-month window.
- Match the stores and regions. Check whether both totals cover iOS and Google Play, and whether mainland China or third-party stores are treated differently.
- Match the metric definition. Look for first installs versus all downloads, gross consumer spend versus net publisher proceeds, and the stated scope of time spent.
- Keep publisher estimates separate. Report the range or two labeled estimates when methods differ; do not average them without a defensible harmonization model.
- Test a category claim locally. Global growth can coexist with regional contraction, shifting subcategories, and a weak product-level outcome.
The numbers also describe different parts of the funnel. Downloads approximate acquisition volume; time spent is a coarse engagement signal; spending measures one monetization path. None demonstrates retention quality, customer satisfaction, profitability, or causal impact. A category can gain spend while shrinking in installs if existing users pay more, and it can gain usage without generating corresponding revenue.
Sensor Tower’s category data is useful for observing direction, not for explaining every cause. Its 2026 release says downloads and time spent in retail apps declined in 2025, while short-form drama and AI apps had unusually high download growth of 278% and 148%, respectively. It also reports global sports betting downloads up 24%, and US sports betting time spent up 7%. Each signal has its own geography and category definition; keep those qualifiers visible when summarizing the result.
“Global downloads and time spent in retail apps declined in 2025”
— Sensor Tower, State of Mobile 2026 press release
What should teams do next?
Everyday users can judge apps by the need they meet, what they cost over time, and how much attention they ask for. Developers can use category reports to choose hypotheses, then validate onboarding, repeat value, and sustainable monetization with their own cohort data. Analysts should publish the vendor, year, scope, and estimate status alongside every figure, especially when two reputable reports disagree.
For a compact app team, our rule of thumb is to make one recurring task noticeably easier before expanding into a wider platform story. That is a recommendation, not a market statistic. NabiGame Studio builds casual mobile and web experiences; if you are exploring a small, focused app or game concept, you can learn about our studio, browse our apps and community tools, or contact us to start a conversation.
FAQ
How many apps were downloaded worldwide in 2025?
Sensor Tower estimated nearly 150 billion downloads across the App Store and Google Play, up 0.8% year over year. Appfigures estimated 106.9 billion across those stores, down 2.7%, so keep the publisher and estimate attached to the number.
How much did consumers spend on mobile apps in 2025?
Sensor Tower estimated $167 billion in IAP revenue and paid app and game spending, up 10.6%. Appfigures estimated $155.8 billion in consumer spending, up 21.6%; the totals and growth rates reflect different vendor methodologies.
Which app category had the most time spent?
Sensor Tower reported nearly 2.5 trillion combined hours on social media apps in 2025, making social a major attention category. Aggregated time is not a measure of user benefit or individual screen time.
Did every mobile app category grow in 2025?
No. Sensor Tower reported unusually strong download growth for short-form drama and AI apps, while global retail app downloads and time spent declined. Other categories also diverged, so inspect the specific segment and geography.
Why do Sensor Tower and Appfigures report different market totals?
They produce estimates using proprietary methods and category definitions, which can differ in coverage, treatment of installs, and modeling. Compare like-for-like scope where possible, and present vendor figures separately if definitions cannot be reconciled.
Comments
Loading comments…