- Published on
AI is creating an era of app glut
- Authors
- Name
- DeepHorus
AI programming tools are turning software development into a one-sentence affair, but app stores have not seen a corresponding surge in demand. a16z's latest data weekly report and academic research both show that while app supply has doubled, user downloads and payments remain unchanged—the era of 'app oversupply' has arrived.
In May this year, researchers from MIT and the University of Pennsylvania published a paper titled 'Writing Code vs. Shipping Code,' marking February 2025 as the starting point of the 'agentic programming era.' Since then, the number of new iOS apps per month has increased from about 40,000 to 120,000, and new apps on Android and Chrome have also doubled to quadrupled, with the curve rising almost vertically.
The researchers observed the performance of these new apps in their first three months after launch and found that overall engagement was flat or declining. The proportion of 'zero/low-usage' apps with fewer than 10 ratings and fewer than 100 downloads increased significantly, while the share of apps achieving 'escape velocity' dropped sharply. Researcher Sternstein named this phenomenon 'App-Slop,' reminiscent of the 'AI Slop' of the content farm era.
Sensor Tower data confirms this judgment from the commercial side. With December 2024 as the baseline of 100, by August 2026, the U.S. app total revenue index was only 102.0, and the total usage time index was 107.0. In a year and a half, revenue grew by only 2% and usage time by 7%—the app economy pie has not grown because of AI.
By category, productivity apps are the only category in the U.S. market where both revenue and usage time are growing rapidly, with revenue up about 100% and usage time up about 55%, driven by ChatGPT, Claude, Gemini, and Grok. Meanwhile, the largest category, games, still has negative revenue growth.
Consumer Edge data cited by a16z shows that as of the first quarter of 2026, only about 3% of U.S. consumers personally paid for AI services, with about 5.9% among 18-24 year olds and only 1.4% among those over 65. Users are willing to pay for AI capabilities, but they won't be interested in every app made with AI.
What AI programming truly changes is the supply curve of the software world, but users only have 24 hours a day and won't download four times as much software just because the number of apps has doubled. The value of Vibe Coding lies in reducing the cost of 'validating an idea,' not the cost of 'building a business.' Independent developers face more brutal competition: more and more developers have become users of AI applications, but the applications developed with AI have no new paying users.
