App Store downloads have been quietly climbing for the past three quarters, and the conventional wisdom — that the golden age of apps was over, that everyone already had everything they needed on their phone — is looking pretty embarrassed right now.
Enjoying this? Get stories like this delivered daily.
The catalyst isn't a new social network or a viral game. It's AI. Specifically, a wave of AI-native apps that are pulling in users, charging real money, and doing something the app economy hasn't managed in years: making people download something new on purpose.
Introduction
The App Store boom is one of the more quietly significant tech stories developing right now. After years of stagnation — where the top-grossing charts were dominated by the same games, the same streaming apps, the same tired utilities — the market is moving again. And it's moving fast.
According to data from app intelligence firm data.ai (formerly App Annie), global consumer spending on iOS and Android apps hit roughly $171 billion in 2023, and 2024 closed even higher. Early 2025 numbers suggest the trajectory hasn't slowed. The driver, consistently, is a new category of AI-powered apps: writing assistants, image generators, voice cloners, coding helpers, and a dozen other things that didn't exist in any meaningful form three years ago.
Here's what's actually happening: the App Store isn't just recovering from a post-pandemic hangover. It's being restructured around a new type of product. And if you're a developer, a consumer, or anyone who watches where tech money flows, that matters more than any single headline. This article is going to tell you why the boom is real, what's actually driving it, and what it means for the apps you'll be paying for next year.
The Slump Nobody Wanted to Talk About
Before the current surge, the App Store was in a rut that the industry was remarkably reluctant to name out loud.
From roughly 2019 through 2022, download growth plateaued. The number of apps available in Apple's App Store peaked at around 2.2 million in 2017 and had actually declined to roughly 1.8 million by 2022 as Apple culled low-quality listings. Revenue growth was real but concentrated — the top 1% of apps captured the vast majority of consumer spending, mostly through in-app purchases in mobile games like Candy Crush Saga and Pokémon GO.
Discovery was broken. The average smartphone user downloaded zero new apps per month, according to a Comscore study that made the rounds in 2017 and remained depressingly accurate for years afterward. The app economy wasn't dying, exactly — it was calcifying. The winners had won, and the market for new winners had essentially closed.
Why Nobody Fixed It
The problem wasn't effort. Thousands of developers launched new apps every week. The problem was that smartphones had reached saturation — most people had apps for everything they needed, and switching costs were high enough that displacing incumbents was nearly impossible.
What the market needed wasn't a better version of an existing app. It needed a new category of thing people actually wanted to do on their phones. Enter 2023.
ChatGPT Broke the Seal
When OpenAI launched the ChatGPT iOS app in May 2023, it hit 500,000 downloads in its first day. That's not a typo. Half a million downloads in 24 hours, for a productivity app, in a market where productivity apps had been an afterthought for years.
The ChatGPT app crossed 1 million downloads in its first five days — faster than any other app in the App Store's history at the time, including TikTok. It wasn't just a hit. It was a signal flare.
What followed was predictable in retrospect but genuinely surprising in the moment: every major AI lab and a thousand startups rushed to build iOS and Android apps. Google launched Gemini as a standalone app. Anthropic shipped Claude for mobile. Microsoft embedded Copilot deeper into its mobile Office suite. The floodgates were open.
The Subscription Model Did the Heavy Lifting
Here's the part that actually matters for the App Store's financial health: AI apps don't just get downloaded. They charge money. Real money, monthly, through Apple and Google's billing infrastructure.
ChatGPT Plus costs $20 per month. Claude Pro is $20 per month. Midjourney, which processes through Discord but has been building toward a standalone app, charges $10 to $60 per month depending on usage tier. These aren't the $2.99 one-time purchases that defined the early App Store. These are recurring subscriptions that generate meaningful revenue — and Apple and Google take 15-30% of every dollar.
That cut is not small. Apple's services revenue, which includes App Store commissions, hit $26.3 billion in its fiscal Q1 2025 alone. A meaningful slice of that growth traces directly to AI app subscriptions.
The New Faces on the Top Charts
Spend five minutes on the App Store's top-grossing list today and you'll see something genuinely different from three years ago. AI apps now regularly appear alongside the usual suspects — TikTok, YouTube, Spotify, the mobile gaming juggernauts.
Apps like Perplexity AI, which positions itself as a search engine replacement and recently crossed 15 million monthly active users, are pulling in subscription revenue at a scale that would have seemed implausible for a two-year-old startup in 2022. Perplexity raised at a $9 billion valuation in early 2025. (I wrote about a similar story of an under-the-radar AI company blowing past its valuation benchmarks — see The $2 Billion AI Startup Nobody Has Actually Heard Of.)
Then there are the AI companion and creative apps. Character.AI, which lets users chat with AI personas, reported 20 million daily active users in 2024. Lensa AI, the portrait generator, had already demonstrated in 2022 that people would pay for AI-generated images — it hit 5 million downloads in a single week during its viral moment. The pattern held and scaled.
Who's Actually Paying
The demographic breakdown here is important and often misread. Early AI app adoption skewed heavily toward tech workers and early adopters. But by late 2024, the paying user base had broadened significantly.
A survey by Sensor Tower from Q3 2024 found that AI app subscribers skewed toward 25-44 year olds with household incomes above $75,000 — the same demographic that drove early streaming adoption. These are people who are already comfortable paying for digital subscriptions and who have identified genuine professional utility in AI tools. That's a much stickier user base than the teenagers who download a game, play it for two weeks, and churn.
Is This a Bubble? Depends on Who You Ask.
Is this a bubble? Depends on who you ask. Then let me answer it.
Some of the enthusiasm is clearly frothy. There are AI apps on the App Store right now that are charging $9.99 per week — not per month, per week — for features that amount to a thin wrapper around OpenAI's API. (The company calls this "AI-powered." What it actually does is forward your question to ChatGPT and mark it up 400%.)
The FTC has already begun scrutinizing some of these apps for misleading pricing practices, and Apple updated its App Store guidelines in 2024 to require clearer disclosure of subscription terms specifically because AI apps were generating an unusual volume of refund requests and complaints.
But the Core Demand Is Real
Critics will point out that many AI apps have high churn rates — users subscribe for a month, experiment, and cancel. That's true. But it's also true of every subscription product at launch, including Netflix, which had enormous early churn before its content library matured.
The apps that are building retention are the ones with genuine workflow integration. Notion AI, which added AI features to its existing productivity platform, reported that users who engage with AI features have 40% higher 90-day retention than those who don't. That's not a gimmick metric. That's evidence of actual utility.
The bubble framing misses the point. Yes, some AI apps will fail. Most startups fail. The question isn't whether every AI app survives — it's whether the category creates durable new revenue for the App Store ecosystem. And the early evidence says yes.
What Apple and Google Are Actually Doing About This
Apple and Google are not passive beneficiaries here. Both companies have been actively courting AI developers and, in some cases, building AI features directly into their platforms in ways that will reshape which apps succeed.
Apple Intelligence, announced at WWDC 2024 and rolling out through iOS 18, integrates generative AI directly into core system functions — writing suggestions in Mail, image generation in Notes, a smarter Siri. The implications for third-party AI apps are genuinely ambiguous. On one hand, Apple is validating the category. On the other, it's competing with it.
Google's situation is similar. Gemini is now the default assistant on new Android devices, replacing Google Assistant. That's a significant distribution advantage for Google's own AI products — and a significant headwind for any third-party AI assistant trying to gain traction on Android.
The 30% Problem Isn't Going Away
There's a structural tension that AI app developers are increasingly vocal about: Apple's 30% commission (reduced to 15% for developers earning under $1 million annually under the Small Business Program) takes a substantial bite out of subscription revenue.
For AI apps, which have unusually high infrastructure costs — running large language models at scale is expensive in ways that serving a static game asset is not — that commission is painful. OpenAI reportedly pays tens of millions of dollars annually in server costs just to run ChatGPT. Handing Apple 15-30% on top of that compresses margins significantly.
This is part of why several AI companies have been nudging users toward web subscriptions, where Apple's cut doesn't apply. It's also why the Epic v. Apple antitrust saga, which resulted in Apple being ordered to allow developers to link to external payment options in the U.S., matters so much to this new generation of AI developers. Every percentage point of commission avoided goes directly to profitability.
What Developers Should Actually Do Right Now
If you build apps, or you're thinking about it, here's the practical read on this moment.
The window for building an AI app that gets discovered organically is still open, but it's closing. The App Store charts are increasingly competitive in the AI category, and Apple's own AI features will commoditize the most basic use cases — writing polish, simple image edits, basic Q&A — within 18 months.
The durable opportunity is in vertical AI: apps that serve a specific professional or creative workflow with genuine depth. An AI app for radiologists reviewing imaging. An AI tool for independent musicians managing rights and royalties. An AI assistant for immigration lawyers. These are areas where general-purpose tools like ChatGPT are demonstrably inadequate, where users have high willingness to pay, and where the competitive moat comes from domain expertise rather than raw model capability.
The developers who will win in the next phase aren't the ones who slapped a chat interface on top of GPT-4o and called it a product. They're the ones who understood a specific user's workflow well enough to build something that actually fits inside it.
The Bottom Line
The App Store boom is real, it's being driven by AI, and it's not a temporary blip. Consumer spending on apps is rising because AI has created a genuinely new category of product that people are willing to pay for on an ongoing basis — something the app economy hadn't produced at scale since the early days of mobile gaming.
The froth is real too. There are AI apps charging predatory prices for commoditized features, and some of them will be regulated or rated into oblivion within a year. That cleanup is healthy, not alarming. Markets sort themselves out. The underlying demand — for tools that make people meaningfully more capable at things they actually need to do — is not going away.
What we're watching is a structural reset of the app economy, not a cycle. The apps that dominate the top-grossing charts in 2027 will look very different from the ones that dominated in 2020. Some of the incumbents will adapt. Many won't. The developers who treat this moment as a genuine platform shift — not just a trend to ride — are the ones who will still be here when the dust settles. The App Store had a decade-long problem with discovery and new value creation. AI just solved it. What happens next is entirely up to the people building.