There Are Too Many Apps for That

An extraordinary number of vibe-coded apps are flooding the market. Is anyone actually using them?

One of the defining elements of AI slop, aside from an off-kilter feel and minimal human oversight, is volume. In recent years, mass amounts of low-quality content have buried original posts on social-media feeds, crowded out human-authored books in online stores, and saturated YouTube algorithms. Now, AI slop is clogging the app stores too.

Thanks to the rise of AI-enabled software development, otherwise known as “vibe coding,”  amateur developers can create hyper-personalized products, or Notes app knockoffs and Strava derivatives, in the span of an evening. In turn, app stores are being inundated with troves of submissions across categories. Nearly 1.3 million new iOS and Google Play apps were released in the first half of 2026 alone, double the amount seen in 2025 or 2024 during the same period, according to Sensor Tower, a market-intelligence firm. At Apple’s developers conference this past June, then-CEO Tim Cook revealed that the company’s App Store was seeing more than 1,000 submissions an hour as people build “more apps than ever.”

The flood of new entries has piled up in digital storefronts: Search for Cal AI, a popular nutrition tracker that logs daily intakes of calories, protein, and fiber through images of meals, and you’ll scroll past endless pages of clones that purport to do the very same thing. In addition to look-alike user interfaces and onboarding questions, some of these dupes share the same name (or variations, such as CalAI, Cal.AI, and Calio AI), and use app icons inspired by Cal AI’s signature black-and-white apple. The bulk of them are only a year or so old—in 2024, there were 330 new calorie-related apps in the Apple and Google Play stores; in 2026, there are already more than 3,000.

The amount of time mobile users are actually spending on apps, meanwhile, has remained relatively stagnant, and downloads declined this year. Whether that’s because these newer, vibe-coded apps are of lower quality, or because the avalanche of new apps makes discoverability more difficult, Mert Demirer, an MIT economics professor researching the surge in AI-coded software, said one thing is certain: “No one is really using them.”

Jonathan Briskman, the director of market insights at Sensor Tower, told me that “you’ve got more apps competing for, essentially, the same pie.” There’s been a notable surge in productivity-related apps, such as calendars and task-focused software, he said, alongside those centered on lifestyle, health and fitness, and finance—all of which saw at least a 40 percent increase in new submissions last year in the Apple App Store alone.

Apps that used to take weeks or even months to build can now be created in an hour. That’s how long Matthew Wallace, a co-founder of an AI-software firm, recently spent making a game that lets users sort cartoon soda cans and cupcakes on digital shelves. Wallace used AI tools to generate the app’s code, background music, and artwork. “I’m pretty shocked at how easy it was to get this far,” he told me. Wallace, who is testing a beta version of the app and plans to release it soon in Apple’s store, made the app for his wife, who enjoyed playing a similar shelf-shorting game but wanted one that didn’t escalate in difficulty. Even the extremely niche world of shelf-focused online puzzles—which are designed to offer a sense of peace through the process of arranging items—is seeing a boost in the vibe-coding era: Wallace’s new app will join the roughly 260 new iOS and Google Play mobile games with shelf in their name that have been released this year, up from 106 in 2025 and only 27 in 2024.

Like Wallace’s game, a large portion of these newer apps are generated for the express use case of whoever created them, or produced on the off chance they become viral hits. Take Simon Pitt, who has three apps on the App Store that he’s vibe coded. There’s Chapters, an audiobook app with an integrated sleep timer; Command Control, which allows him to view local data on bus stops and movie times; and Hippo Flippo, a hippo-themed puzzle game. None have gained much traction in the Apple App Store, although Chapters has about 300 users (some of whom occasionally email him demanding new features). Whereas Pitt, who manages a software-development team in England, has the expertise to fiddle around with code after it is AI-generated, a large portion of the creators behind new apps appear to be amateurs: The number of iOS and Google Play apps released by first-time app developers rose by more than 44 percent and 47 percent last year respectively, according to Briskman.

In some cases, vibe coders replicate apps in order to avoid paywalls, creating replacements for apps as varied as budgeting programs and recipe keepers. Pete McPherson, an indie developer, told me that tweaks of preexisting app ideas account for roughly 90 percent of the vibe-coding projects he releases, such as a recent version of a photo-library-cleanup app he enjoyed using but wanted to access for free. “I have a hard time paying for any apps now,” he said.

Vibe coders can follow “un-slop” tutorials that offer tips on how to ensure their products look less obviously AI made. But just like with other forms of AI slop, there are common tells: Developers say that many thrown-together slop apps feature small fonts, purple-gradient backgrounds, and a plethora of emojis—the app versions of constant em dashes in writing, or the It’s not X, it’s Y construction. The overload is frustrating for users, who are fatigued by constant exposure to AI-generated apps that feel interchangeable with preexisting ones. “Every day there’s a different person thinking that their AI slop is somehow going to be revolutionary to the fitness community,” one Reddit user wrote in July, in response to yet another fitness app being posted on the Strava Subreddit.

Apple is attempting to crack down on the surge—this summer, it updated its app-store guidelines to urge developers against “opportunistically creating variants of existing app categories or popular apps,” and said it would enforce tougher barriers to entry in well-established categories such as wallpapers, sound effects, and timers. Existing companies, too, are feeling the impact of the new wave of entries. Some game developers have recently seen mobile rip-offs proliferate across app stores after demos. Others have seen the influx put pressure on their resources; in June, Strava began charging developers for access to its fitness data. In a developer program update, a Strava community manager blamed “zero-code AI tools that produce apps that hammer APIs.”

A separate concern surrounding vibe-coded apps is that the first-time developers behind them may have neither the wherewithal to check AI-written code for security issues nor the capability to create bug fixes in order to maintain these apps over time. “The moment users upload personal information, or the app connects to real data sources, the creator becomes the entire security and support team for it,” Dor Zvi, the CEO of Red Access, a cybersecurity company that has researched security issues across vibe-coded applications, told me. “Those are responsibilities most of the one-person operations behind these apps aren’t built to provide.”

The Apple and Google Play app stores seem strained by the surge of new submissions. Elon Musk, responding to an X post claiming it now takes longer for an app to be reviewed than it does to create it, has described recent iOS delays as “ridiculous.” Gernot Poetsch, a developer based in Berlin, told me that review cycles that previously took 24 to 48 hours can now last a week, and attributed delays to the “avalanche” of apps that platforms must now comb through. Google did not respond to requests for comment. An Apple spokesperson told me that the average app-review time in recent weeks has been roughly 36 hours, although some new apps can take longer if they need additional scrutiny. To keep pace with growth, Apple said it is prioritizing reviews of updates for existing apps, expanding its use of automation, and strengthening safeguards for abuse detection and repeated submissions.

Whatever happens with slop apps could be a sign for where the rest of the internet is heading. Mediocre, AI-generated content is flooding nearly everything, including academic journals, courtroom filings, and even Zillow listings. That’s forcing us to wade through a plethora of cheap content and often shoddy copycats. “The cost of producing something is falling, but the cost of trusting gets passed on to everybody else,” Paul Shovlin, a professor at Ohio University studying AI, told me. For now, at least, “the onus is on us to sift through all the detritus.”