The App Store just got a jolt of life. Sensor Tower data shows new app submissions grew 30% to nearly 600,000 globally in 2025, reversing a years-long decline that saw submissions fall 46% between 2016 and 2024.
The main driver is AI coding tools like Anthropic's Claude Code and OpenAI's Codex, which let anyone build apps with plain English prompts. 'We've seen an explosive growth of new apps over the past year,' said Abraham Yousef, senior insights analyst at Sensor Tower. 'It aligns with a broader release of agentic coding tools that remove prior difficulties of creating apps.'
But Apple isn't welcoming all of these tools with open arms. The company has pulled or blocked updates to apps like Anything and Replit, claiming they violate App Store guidelines by generating interpreted code that can change an app's primary purpose. It's a predictable tension. Vibe coding platforms let non-programmers spin up functional software in minutes, but they also produce code that behaves in ways Apple's review rules weren't built to handle.
Apple insists it's keeping up. A spokesperson said the review team processes 90% of submissions within 48 hours and handles over 200,000 apps per week, with AI tools now assisting human reviewers. Elon Musk complained on X about rising review delays, but Apple disputes that.
WWDC26 is where this collision gets real. Apple either updates its review rules for an agent-built world, or it watches developers drift to platforms that don't police vibe-coded software.