Apple faced a flood of AI apps in the App Store and tightened app review

Apple faced a side effect of the vibe coding boom: the App Store was flooded with new applications, and the review queue surged sharply. In a single quarter, the number of new submissions jumped by 84%, and now the company is simultaneously clearing the backlog while applying stricter enforcement of old rules against applications that generate code on the fly.

Where the Surge Came From

Vibe coding turned mobile app development from a week or month-long effort into a process where an idea is described in plain language and the model itself assembles a working prototype. According to The Information, as cited by TNW, in just one quarter the number of new applications submitted to the App Store grew by 84% — the sharpest jump in a decade. Separate Sensor Tower metrics showed similar dynamics: the pace of iOS app launches accelerated in late 2025 and early 2026, and over all of 2025 the store received 557,000 new submissions. Behind this stand several platforms that made releasing applications almost a mass service:

• Cursor — an AI-IDE for developers that has already become mainstream
• Lovable — a constructor for people without strong technical background
• Replit — an environment where you can quickly assemble and deploy a product
• Bolt.new — a tool for ultra-fast transition from idea to prototype

They all have one thing in common: they dramatically lowered the barrier to entry and made publishing an application almost a routine task. For Apple, this means that the App Store funnel now receives not only studios and indie developers, but thousands of people who hadn't written code at all the day before. The cheaper and faster the assembly, the more experimental, raw, and disposable applications arrive for review.

Why Reviews Are Stuck

The problem is not just volume. The App Store was originally built on a different model: the developer assembles a static version of the application, sends it for review, Apple checks the specific build and after approval it goes to users as-is. But vibe coding often assumes different behavior — the service generates or executes new code on demand from the user after publication.

And this conflicts with rule 2.5.2 from the App Review Guidelines: an application must be self-contained and must not download or execute code that changes its functionality after review.

Hence the sharp spike in delays. In March 2026, developers reported that updates and new applications waited not for the usual 24–48 hours, but for one to 30 days or more. According to TNW, most of the time the submissions spent waiting even before the review began — in the Waiting for Review queue.

That is, the bottleneck emerged not from detailed legal assessment but at the very intake: the stream of new submissions became too large for an infrastructure designed for a much slower development cycle.

How Apple Is Responding

Apple's response has moved from passive waiting to direct pressure on category leaders. In mid-March, reports emerged that the company quietly blocked updates for Replit and Vibecode without making public announcements beforehand. Developers received rejections citing the same rule 2.

5.2 and complained that Apple had tightened its interpretation without explicit warning. Formally, the company is not introducing new rules for AI applications, but simply enforcing old ones more strictly.

Practically, the difference is slight: the key feature of many vibe coding services is precisely the dynamic generation of new behavior within the application. The most notable case is Anything, an application for creating small utilities and automations through text queries. According to its co-founder, Apple had been slowing down updates to the service since December 2025, and on March 30, 2026, removed it from the App Store entirely.

The team tried to ease the conflict by showing the generated result not inside the application but in a browser, but even that compromise didn't work. Apple's position is nonetheless predictable: the company claims it is fighting not vibe coding as an idea, but the mechanics that allow an application to change its capabilities after review.

What This Means

The situation demonstrates the first major conflict between the speed of AI development and old platform distribution rules. If Apple leaves everything as is, developers will migrate to web, Android, and other channels where dynamic code is more loosely restricted. If it relaxes the rules, it will have to rebuild the review and quality control model itself. For the market this is an important signal: mass application generation has already run up against not the quality of models, but the rules for accessing the platform.