OpenAI Just Called Out Apple in Public. This Was Never About the App Store
Companies rarely attack each other in public. They especially don’t attack the company that owns the storefront selling their app. So when OpenAI put up a post on its own blog arguing that Apple is getting this wrong, that’s not a tantrum. It’s a declaration of war over the most valuable piece of real estate on a smartphone for the next decade.
One caveat up front: this is fresh. The usual venues where these things get chewed over — Hacker News, the relevant subreddits, tech X — haven’t fully weighed in yet. What follows is the confirmed facts plus my read on how platform fights like this actually play out.
The Default Assistant Slot Is Not Like Other Slots
What’s the most valuable position on a phone? Most people picture the home screen’s first row. It’s lower than that. It’s the layer that catches the words when a user says “hey, sort this out for me.”
Here’s what makes it different from every other slot. An app requires intent — the user has to decide to open it. An assistant fires when the user has chosen nothing at all. Being on without being picked is the whole point of a default. Microsoft understood this when it bundled Internet Explorer into Windows and buried Netscape. Google understands it now, paying Apple somewhere north of $20 billion a year to stay the default search engine in Safari. Same math, different decade.
OpenAI’s position is easy to read. ChatGPT has hundreds of millions of monthly users, but those people still open an app. Siri doesn’t need to be opened. It’s preinstalled on well over a billion iPhones, and reaching it takes a long press on the side button. No download, no decision, no funnel.
Apple’s Argument: An Assistant Isn’t an App, It’s the OS
Apple’s logic holds together, at least on its own terms. The company has spent years insisting Siri is not an app but a component of the operating system. That distinction is doing enormous work.
If it were an app, users should be able to swap it. Apple already caved on that elsewhere: under pressure from the EU’s Digital Markets Act, it opened up default browser, default mail, and default maps. But an assistant, Apple argues, is a different animal. To be useful, it needs contacts, calendar, photos, messages, health data, and location history. That’s a request for permission to read a person’s entire life.
So Apple plays the privacy card. On-device processing, Private Cloud Compute, and the claim that piping all of that to a third party’s servers is structurally unsafe. The Siri-ChatGPT integration reflects it: separate consent flow, and a prompt every time a request is about to leave for OpenAI.
The awkward part is how neatly this argument aligns with Apple’s revenue. The harder Apple protects privacy, the more control Apple keeps. That’s why critics keep describing privacy as Apple’s moat rather than Apple’s principle. Both things can be true at once, which is exactly what makes the argument so hard to dislodge.
What OpenAI Is Actually Afraid Of
Why now, and why this loud? The timing gives it away.
Apple has been papering over its own AI gaps with third-party model integrations. The moment Apple’s in-house foundation models are good enough, the outside partner becomes unnecessary. OpenAI is working with a clock: if the door doesn’t open now, it never opens. Ask to be considered as a default assistant after Apple has swapped Siri’s brain for its own, and you’re asking a question nobody at Apple has any reason to answer.
Hardware is the other variable. Reports that OpenAI is building its own AI device have been circulating for a while. The strategic logic is obvious — better to own the house than rent a room in someone else’s. But new hardware categories take years to land, if they land at all. You need a bridge to cross in the meantime, and the bridge is the iPhone.
The public callout is itself a tactic. Instead of pressing the demand quietly across a negotiating table, OpenAI seated regulators and the press in the audience. Apple’s App Store policies are already under active scrutiny in both Brussels and Washington. Against that backdrop, “we just want fair competition” is a frame that travels remarkably well.
The Part Developers Should Be Watching
However this ends, the fallout reaches anyone shipping an app.
Open up the assistant layer and you change how users find software at all. Today a user searches the App Store and installs something. In an assistant-mediated world, the user says “find me a flight” and the assistant decides which service gets called. Assistant selection replaces App Store ranking as the thing that routes traffic. Whoever builds the assistant becomes the new gatekeeper — the gate just moved.
For users it’s not simple either. More choice is good. But deciding which company gets your message history and your photo library is a categorically heavier decision than picking a browser. Apple’s concern is not 100% self-serving, and pretending otherwise makes the debate dumber than it needs to be.
Cold read: neither side is fighting for you. OpenAI wants its model to be the default. Apple wants its ecosystem to stay closed. Openness and privacy are the flattering names each company gives to the outcome it already wanted.
The Question That Sticks
The browser wars took two decades to resolve. The search default contracts are still being litigated. This fight over the assistant layer is not ending with one blog post. What the post does tell us is that OpenAI concluded quiet negotiation wasn’t going to get there.
Here’s what I keep coming back to. When you talk to your phone, who do you want listening? The company that made the device, or the company that made the model? How that question gets answered — by users, by courts, or by neither — sets the shape of the platform landscape for the next ten years.
Comments
Loading comments...