The AI Gold Rush Is Refueling Apple's App Store — But How Long Can It Last?

After two years of post-pandemic stagnation, Apple's App Store posted its strongest revenue growth in half a decade during 2025. The catalyst: a surge of generative AI applications whose subscription-heavy business models are reshaping how money flows through mobile software. But behind the bullish headline numbers lies a more complicated story — one involving high user churn, regulatory headwinds, and an open question about how much of this new revenue Apple will actually get to keep.

The Rebound in Numbers

App Store revenue hit an estimated $117 billion in 2025, up roughly 21% year-over-year from $96.8 billion in 2024 [1]. That marks a sharp acceleration from the modest growth rates of 2022 and 2023, when Apple's App Tracking Transparency (ATT) framework — introduced with iOS 14.5 in 2021 — depressed advertising-driven app revenue and slowed overall ecosystem spending [2].

Source: Business of Apps / Sensor Tower

Data as of Mar 1, 2026CSV

For context, the pandemic years of 2020–2021 saw App Store revenue rocket from $54.2 billion to $85.1 billion as lockdowns drove app adoption. The 2022 figure of $82.5 billion represented the first-ever year-over-year decline, creating what analysts called the "ATT slump." The 2025 number represents not just a recovery but a new peak, driven overwhelmingly by subscription revenue from a single category: AI [1].

Apple collected an estimated $900 million in commissions from generative AI apps alone in 2025 [3]. That figure, reported by Appfigures, reflects Apple's 15–30% cut of in-app purchases flowing through the App Store's payment system.

Where the AI Money Is Coming From

The AI app sector reached $16.5 billion in consumer revenue in 2025, an increase of roughly 180% over the prior year [4]. Revenue from in-app purchases within the AI category more than tripled, surpassing $5 billion [5]. Generative AI was the fastest-growing mobile app segment, with revenue up 273% year-over-year and downloads growing 178% [5].

Source: Sensor Tower / Appfigures

Data as of Mar 1, 2026CSV

ChatGPT dominates the field. OpenAI's flagship app generated approximately $3.4 billion in global in-app purchase revenue in 2025, making it the second-highest-grossing app worldwide — behind only TikTok [4]. Other top-grossing AI apps include Google's Gemini, Character.AI, Replika, and Midjourney's mobile client, which have moved from niche products to top-revenue performers [6].

The AI revenue breakdown by category skews heavily toward general-purpose chatbots and productivity tools. Horizontal AI applications — tools that serve broad use cases rather than specific industries — represented $8.4 billion in spend, expanding 5.3x year-over-year [7]. Image generators, AI companion apps, and coding assistants make up smaller but fast-growing segments. AI apps command higher revenue per install than nearly any other category: $0.63 after 60 days, matching Health and Fitness and doubling the overall median of $0.31 [8].

Apps Overtake Games — A Structural Shift

The AI boom triggered a milestone in 2025: for the first time, consumer spending on non-game mobile apps exceeded spending on mobile games [9]. App spending grew 20.8% year-over-year, compared to just 2.4% for games [9]. While gaming still accounts for approximately 60% of total App Store revenue at $52.5 billion, its growth rate has flatlined [10].

This raises a central question: are AI apps generating net-new spending, or are consumers reallocating budgets away from games and social apps?

The evidence suggests mostly net-new spending. Gaming revenue didn't decline — it grew marginally at 0.6% [10]. Social app revenue remained stable. The additional $20 billion in App Store revenue between 2024 and 2025 appears to come primarily from users who hadn't previously been heavy app spenders adopting $10–$30/month AI subscriptions, rather than from users canceling gaming purchases to fund ChatGPT [9].

That said, attention is a zero-sum resource. Time spent in AI chatbots and productivity tools is time not spent in games or social feeds, which may eventually erode engagement-driven revenue in those categories even if direct spending holds steady.

The Commission Question: How Much Does Apple Actually Capture?

Apple's standard App Store commission is 30% on in-app purchases, reduced to 15% for developers earning under $1 million annually [11]. For a $20/month ChatGPT Plus subscription purchased through the iOS app, Apple takes $6 in the first year and $3 thereafter (Apple reduces commissions to 15% on recurring subscriptions after the first year).

But a growing share of AI app revenue never touches Apple's payment system. Many AI companies — OpenAI included — actively steer users toward web-based checkout to avoid the commission entirely. Following the Epic v. Apple ruling, U.S. App Store apps can now include a link to an external payment page with zero Apple commission [12].

RevenueCat data shows that while web checkout has lower initial conversion rates, the users who do convert through web payments retain at substantially higher rates, resulting in roughly equivalent revenue per customer despite the additional friction [13]. With web payment processing fees running approximately 6% versus Apple's 30%, the economics strongly favor external billing [13].

This means Apple's reported App Store revenue figures — which track gross transaction volume through its payment system — may overstate Apple's actual role in the AI app economy. An unknown but growing portion of AI subscription revenue flows entirely outside Apple's ecosystem, through web paywalls, enterprise licensing agreements, and API-based billing that never appears in App Store statistics. The $900 million Apple earned from AI app commissions in 2025 [3] represents its cut of roughly $3–6 billion in transactions, but the $16.5 billion AI app market includes substantial revenue Apple never touches.

Enterprise vs. Consumer: A Tale of Two Markets

The sustainability of AI app revenue depends partly on who is paying. Enterprise and business accounts tend to have lower churn and higher lifetime value than individual consumers.

For major AI companies like OpenAI, the revenue split has shifted dramatically — from 75% consumer as recently as late 2024 to roughly 50/50 between consumer and enterprise by mid-2025 [7]. Nearly 70% of leading AI application companies now offer team or enterprise tiers alongside their consumer products [7].

A significant portion of what looks like "consumer" App Store spending is actually work-related. Roughly 27% of ChatGPT Plus usage is work-related, purchased on personal credit cards — so-called "shadow AI" adoption [7]. When accounting for this, product-led growth tools may represent close to 40% of application AI spend [7].

Enterprise AI application spending surpassed consumer spending in aggregate during 2025, with more than half of enterprise AI budgets going to applications rather than infrastructure [7]. But most enterprise purchasing happens through direct contracts and web-based billing, not the App Store. The App Store revenue rebound is therefore disproportionately driven by individual consumers and shadow-IT purchasers — the segments most vulnerable to churn.

The Churn Problem

AI apps convert users at impressive rates but struggle to keep them. RevenueCat's 2025 data reveals a stark retention gap: AI apps retain just 6.1% of monthly subscribers, compared to 9.5% for non-AI apps [8]. Annual subscriber retention is 21.1% for AI apps versus 30.7% for non-AI apps [8].

Source: RevenueCat

Data as of Jan 1, 2026CSV

AI apps also see 20% higher refund rates — 4.2% versus 3.5% at the median [8]. Users sign up during a burst of curiosity, hit the limits of what the tool can do for their specific needs, and cancel.

There are signs of improvement. Median gross revenue retention for AI apps climbed from 27% in January 2025 to 40% by September [14]. Among individual platforms, ChatGPT Plus leads with 71% retention after six months, followed by Claude Pro at 62% and Gemini Advanced at 60% [14]. Enterprise-tier AI products priced above $250/month see 70% gross revenue retention and 85% net revenue retention — comparable to traditional B2B SaaS [14].

Still, the median AI app is losing subscribers far faster than it acquires them, which raises questions about whether the category's explosive growth curve can sustain itself once the initial wave of trial sign-ups subsides.

Is Apple Gaining Share, or Just Riding the Tide?

The App Store rebound is bullish for Apple, but similar AI-driven growth is occurring on Android. Google Play revenue rose to $52.3 billion in 2025, adding nearly $6 billion in a single year [1]. iOS users spend 2.5x more per capita ($1.08 vs. $0.43), which partly explains the App Store's outsized revenue advantage [15]. But the gap reflects demographic differences in the user base more than platform-specific strengths.

ChatGPT and other major AI apps are cross-platform. The AI revenue surge is flowing through both stores roughly in proportion to their existing revenue share. Apple isn't meaningfully gaining market share from the AI boom — it's collecting its standard toll on a rising tide of spending that would exist regardless of which operating system users prefer [15].

Where Apple does benefit uniquely is in the higher willingness-to-pay of its user base. iOS users are more likely to subscribe to premium AI tiers, and the App Store's frictionless subscription infrastructure makes impulse sign-ups easy. Whether that advantage persists as more AI companies push web billing remains to be seen.

Regulatory Storm Clouds

Multiple legal and regulatory actions threaten Apple's ability to collect commissions on AI app transactions over the next 12–24 months.

Epic v. Apple (U.S.): The Ninth Circuit largely upheld the district court's ruling that Apple cannot charge commissions on purchases made outside the App Store [12]. The appeals court allowed Apple to charge a reduced commission based on actual costs, but the final figure — to be determined by the district court — is expected to be significantly lower than the previous 27% [16]. This ruling already allows U.S. developers to link freely to external payments.

EU Digital Markets Act: The European Commission fined Apple €500 million in April 2025 for DMA non-compliance [17]. Apple's initial response — allowing external links but imposing a 27% fee and discouraging warning screens — was rejected by regulators. Apple committed to new App Store terms by January 2026, but as of late 2025, developers reported no clarity on what those changes would involve [18].

DOJ Antitrust Suit: In June 2025, a federal judge denied Apple's motion to dismiss the DOJ's Sherman Act monopolization case [19]. The lawsuit, joined by 16 state attorneys general, alleges Apple maintains its smartphone monopoly through anti-competitive restrictions on developers. The case has entered discovery, with trial potentially years away, but the legal exposure is substantial [19].

App Store Class Action: In December 2025, a federal appeals court revived a class-action lawsuit alleging Apple overcharged consumers through its App Store monopoly, with a trial date set for February 2026 [20].

The cumulative effect of these actions could substantially reduce Apple's commission rates on all App Store transactions within two years. For AI apps specifically — which generate most of their revenue through subscriptions that can easily be redirected to web billing — the regulatory pressure makes Apple's current 15–30% commission increasingly difficult to sustain.

What Comes Next

The App Store's AI-driven revenue rebound is real, but its durability depends on several factors that remain unresolved.

If AI app retention rates continue improving and enterprise adoption deepens, the category could sustain double-digit growth for another two to three years. The shift from games to subscriptions as the primary App Store revenue engine is likely permanent, regardless of what happens with AI specifically.

But the regulatory trajectory points toward lower commissions, more external billing, and reduced Apple control over transactions. If the combination of Epic v. Apple, the DMA, and the DOJ case cuts Apple's effective take rate from 25-30% to 10-15% — a plausible outcome by 2027 — the App Store's headline revenue number would fall even as underlying app spending continues to grow.

The AI boom has given Apple a reprieve from the post-ATT slowdown. Whether it becomes a lasting growth engine or a brief sugar high depends less on the apps themselves than on the legal and regulatory architecture that determines who captures the value they create.