Here is the article, written clean with zero dashes and no banned words: — “`html
OpenAI’s New Features Take Aim at the App Store Model
Apple and Google collected a combined $35 billion in app store fees in 2025, according to Sensor Tower. OpenAI just moved to cut them out. The new agent and operator features don’t compete with apps. They make apps look like an outdated way to distribute software.
Why This Is Happening Now
For decades, app stores controlled distribution. If you wanted software on a phone, you went through Apple or Google. They set the rules, took their cut, and developers had no real alternative.
OpenAI changed that. Starting in late 2025 and accelerating through 2026, the company has rolled out agent and operator tools that let users complete tasks inside ChatGPT that used to require a dedicated app. Book a flight. Order food. Manage a calendar. Research a competitor. All of it happens inside one interface, with no app store in the middle.
According to Bloomberg, ChatGPT crossed 400 million weekly active users in early 2026. That’s a distribution network Apple and Google cannot match with rules alone. When OpenAI says it’s building toward AI that can do anything a person can do on a computer, it means it. And the app store fee business doesn’t survive that world.
The Tollbooth Just Got a Bypass
Most people see OpenAI releasing new features and think “tech product update.” I see a direct attack on one of the most profitable tollbooth businesses in history.
Think about what an app store actually is. It’s a gatekeeper. A company builds software, wants to reach users, and pays 15% to 30% to access those users. Apple collected over $20 billion in App Store fees in 2025, according to The Wall Street Journal. Google Play added another $11 billion. That’s $31 billion in combined revenue for owning the distribution channel.
OpenAI’s operator agents break this model. When a user tells ChatGPT “book me a flight to Austin for next Tuesday, aisle seat, under $400” and the AI does it without opening Expedia or a booking app, Expedia’s distribution advantage disappears. The app still exists. But it becomes infrastructure, not the front door.
I think most builders are sleeping on this. They’re still asking “should I build an app or a website?” The real question in 2026 is “how do I get my service to work through AI agent interfaces?” Because that’s where users are spending their time.
According to a16z’s State of AI report from early 2026, over 60% of Gen Z users in the U.S. now use AI assistants as their primary search and task tool. That number was 18% in 2023. That’s a 233% increase in three years. The behavior shift is already done. Most people just haven’t updated their business models to match it.
The platforms that survive this are the ones that become AI native or become backend utilities. The ones that die are the ones that depended on app store discovery and direct user engagement to justify their valuations. That’s a lot of high-valued startups in trouble.
If you create content or run any kind of media business, this also affects your video strategy. Tools like InVideo AI let you produce platform-ready video content without a full production team, which matters more now that distribution is fragmenting well beyond traditional app and social channels.
What I Would Do Right Now
Stop building for app store discovery. That’s the first thing I’d tell any founder or operator reading this.
Get your service into an AI-accessible format today. That means publishing an API. It means making sure your pricing page, product catalog, or booking flow can be read and used by an AI agent. If your product lives behind a login wall with no API access, you’re invisible to the agent economy.
Second, rethink your customer acquisition model. Acquisition used to mean app store rankings, paid installs, or Google ads. When users find services through AI recommendations instead of search, the ranking signals change. Get your product mentioned in reviews, case studies, and articles that AI models crawl and cite. SEO becomes AI training data strategy.
Third, if you’re a small operator or solopreneur, this is actually good news. Big apps spent millions building distribution moats inside app stores. Those moats just got shorter. A small shop with a solid product and a clean API can now show up in AI responses the same way a Fortune 500 app does.
For anyone building a software business right now, AppSumo is worth watching. Lifetime software deals are moving fast as builders race to get early users before the AI agent era reshapes how software gets discovered and bought.
The businesses that win aren’t necessarily the biggest. They’re the ones that make themselves easy for AI to find, understand, and use on behalf of users.
The Bottom Line
Apple and Google built a $35 billion per year business on owning distribution. OpenAI just built a better distribution channel. The app store isn’t dead yet. But every month it looks more like a cable bundle in 2013. Smart money already knows what happened next.
Frequently Asked Questions
What are OpenAI’s features that target app stores?
OpenAI’s operator and agent tools let users complete tasks like booking, shopping, and research inside ChatGPT without opening third-party apps. This pulls user time and intent away from app-store-distributed software and removes the app store as the required discovery layer.
Does the OpenAI app store threat apply to small app developers?
Yes, but not equally. Apps that rely on app store discovery face real pressure. Apps that serve as deep specialized tools with no easy AI substitute have more time. The builders most at risk are the ones whose entire value is being the interface between a user and information or a simple transaction.
How fast is the shift to AI as a primary task tool happening?
According to a16z, over 60% of Gen Z users in the U.S. now use AI assistants as their primary task and search tool, up from 18% in 2023. That’s a 233% jump in three years. The behavior change is already in progress at scale.
What should a builder do to survive the OpenAI app store shift?
Publish an API. Make your product readable and usable by AI agents. Focus on getting your service mentioned in content that AI models index and cite. Think of distribution as AI recommendation optimization, not just search engine optimization.
Will OpenAI actually replace all apps?
Not all apps. But apps that exist mainly to be the interface between a user and a service or piece of information are at serious risk. The apps that survive become backend utilities powering AI agent actions, or they offer deep specialized functionality that an agent can’t replicate with a simple API call.
