AI-powered apps struggle to keep subscribers, report finds
Users cancel annual subscriptions to AI-powered apps 30% faster than they do for non-AI apps
AI-powered mobile applications are growing rapidly, but a new industry report suggests they may face a significant challenge: keeping users subscribed over time.
The findings come from the 2026 State of Subscription Apps Report published by subscription management firm RevenueCat, which analysed more than 1 billion in-app transactions across iOS, Android and web platforms.
The company’s tools are used by more than 75,000 developers, generating over $11bn in annual revenue for app makers.
According to the report, while AI features can help apps attract paying customers, they do not necessarily ensure long-term loyalty.
At the median level, users cancel annual subscriptions to AI-powered apps 30% faster than they do for non-AI apps.
RevenueCat’s data shows that annual retention, the share of users who remain subscribed after a year, was 21.1% for AI apps, compared with 30.7% for apps without AI features. The gap also appears in shorter-term figures. Monthly retention for AI apps stood at 6.1%, compared with 9.5% for non-AI apps.
The only category where AI apps performed better was weekly retention, where they recorded 2.5%, compared with 1.7% for non-AI apps. However, weekly subscription plans are relatively uncommon among AI services.
Researchers say the rapid pace of development in AI may partly explain the trend.
As new AI models and features emerge, users are more likely to experiment with different apps, the report suggests, meaning customers may switch services more frequently in search of the latest capabilities.
The report also found that AI-powered apps experience higher refund rates than other subscription apps. At the median level, refunds accounted for 4.2% of purchases for AI apps, compared with 3.5% for non-AI apps.
Strong early monetisation
Despite these challenges, AI-powered apps appear particularly effective at converting new users into paying customers.
Trial users of AI apps converted to paid subscriptions 52% more often than those using non-AI apps.
They also generated higher realised lifetime value, with the median AI app producing $18.92 per paying user each month, compared with $13.59 for non-AI apps.
Overall, AI apps monetised downloads around 20% more effectively than traditional apps.
Despite the surge of interest in AI, the majority of subscription apps are still not AI-driven. RevenueCat found that 27.1% of apps on its platform describe themselves as AI-powered, compared with 72.9% that do not.
Adoption varies widely by category. Photo and video apps have the largest share of AI-powered services, at 61.4%, while gaming apps have the lowest at 6.2%.
Travel and business apps also show relatively low levels of AI integration.
Growing gap between top apps and the rest
The report also highlights a widening divide between the most successful subscription apps and the broader market.
Driven partly by AI-assisted development tools that lower barriers to entry, 2025 saw a record number of new apps released.
The top 25% of apps increased monthly recurring revenue by 80% year-on-year, while the top 10% grew by 306%. Meanwhile, the share of new apps reaching $1,000 per month in recurring revenue fell from 19% in 2024 to 17% in 2025.
The proportion surpassing $10,000 per month also declined, from 5.3% to 4.6%.
Among categories, photo and video apps were the most likely to reach the $1,000 monthly mark, while gaming apps were the most likely to exceed $10,000.
By contrast, developers face tougher odds in sectors such as education, productivity, travel, shopping and business, where fewer new apps break through financially.
