a16z Reports Surge in AI Applications but Revenue Growth Lags Behind
a16z's latest research indicates that since February 2025, there has been a structural surge in the submission of new AI-driven applications, but the overall revenue and growth of paying users in the application ecosystem have not kept pace, showing a divergence pattern of "quantity increase, quality lag" in market prosperity.
According to data cited by a16z, the monthly submission of new applications on the iOS platform has jumped from about 30,000 to 50,000 to approximately 120,000, while new Chrome extensions surged from about 3,000 to over 20,000. The submission volume of Android applications has also rebounded from previous lows to about 100,000, with all three major platforms experiencing several-fold expansions on the supply side.
In contrast to the explosive supply side, user engagement and revenue in mature application categories have weakened: gaming app usage time has decreased by about 3%, with revenue down by about 15%; lifestyle app revenue has dropped by about 30%; and revenue from book reading apps has seen the largest decline, reaching about 40%.
Independent data also confirms the extreme concentration of revenue structure—according to Sensor Tower's "State of Mobile 2026" report, in 2025, the top 1% of applications captured 92.2% of in-app purchase revenue in the App Store and Google Play (approximately $154 billion), while the remaining 99% of applications shared only about $13.1 billion; during the same period, global in-app purchase revenue from application categories surpassed that of games for the first time, reaching about $85.6 billion compared to $81.8 billion.
a16z believes that the new revenue generated by a few leading AI applications is offset by declines in mature categories such as games, lifestyle, and books. The market is not experiencing a "collective prosperity of all AI applications," but rather growth is highly concentrated among a small number of winners, with the vast majority of newly submitted applications failing to convert into sustainable users and revenue.
From a market mechanism perspective, the limited resources for app store reviews and user attention mean that the exponential expansion on the supply side has marginalized many new applications that lack user acquisition channels and monetization capabilities; funding and user attention continue to concentrate on leading AI applications that already have network effects and distribution advantages, further increasing customer acquisition costs for latecomers and creating a "Matthew effect" where the strong get stronger.
Source: Public Information
ABAB AI Insight
a16z has long tracked the landscape of AI consumer applications through its "Top 100 Gen AI Consumer Apps" series of reports. Analysts like Olivia Moore have previously pointed out the structural challenges in user retention and payment conversion for AI applications. This report continues a16z's consistent "data-driven, bubble warning" style, similar to its previous warnings about SaaS valuation bubbles and the uncontrolled costs of user acquisition in mobile gaming.
In 2025, Apple's App Store collected nearly $900 million in revenue from generative AI applications, rising from about $35 million at the beginning of the year to a peak of about $101 million in August, indicating that the platform itself is profiting from the AI application boom. Meanwhile, venture capital continues to bet heavily on leading AI applications, while a vast number of long-tail AI applications are effectively "working for" the platform and a few winners due to their inability to secure distribution resources and funding.
This pattern is highly reminiscent of the early "gold rush" in the iOS App Store from 2008 to 2012—during that time, the number of applications exploded, but revenue was also concentrated in a few blockbuster hits, leaving the vast majority of developers with nothing. The current AI application industry is in an early, rough expansion phase characterized by "explosive availability of tools, but distribution and retention barriers have yet to be broken," which is highly comparable to the early developer ecosystem of the mobile internet.
Essentially, this is a concentration of capital—widespread availability of AI large model APIs and development tools has significantly lowered the "production" threshold for applications, leading to intensified homogenous competition on the supply side, while user attention, app store traffic distribution, and payment habits have not expanded in tandem, ultimately forcing revenue and user resources to concentrate heavily on leading applications that possess first-mover advantages, brand recognition, and distribution channels, making it difficult for long-tail applications to break through growth bottlenecks even if the technology is available.
ABAB News · Law of Cognition
The easier the tools are to obtain, the more valuable the moat becomes.
Where there is an explosion in quantity, there is often a dilution of value.
Prosperity does not mean everyone shares in it; it is another way of saying that the winners take all.