Tim Cook is stepping down as Apple CEO on September 1. His replacement is John Ternus — the guy who runs Apple Silicon and the device hardware team. Not a services guy. Not a software platform guy. A hardware engineer.

Everyone covering this appointment is asking what it means for the App Store, for Apple TV+, for Services revenue. That's the wrong question.

The right question is: what does a hardware engineer running Apple see when he looks at the agentic commerce race?

He sees two billion devices. Devices with Face ID, Secure Enclave, and Apple Pay already embedded. Devices where every major AI model wants to run. Devices that Apple controls end to end — the chip, the OS, the wallet, the biometric authentication layer.

He doesn't need to build a shopping agent. He already owns the room where every agent competes.


There are three things Apple did in the last 90 days that nobody has connected into a single story.

In March, Apple hired Lilian Rincon as VP of product for Apple Intelligence and Siri. Her last job was VP of product at Google Shopping. That's a commerce hire dressed in a product title — someone who ran AI-powered discovery and shopping surfaces for the world's second-largest shopping platform. Apple brought that expertise in-house right before WWDC.

At WWDC in June, Apple is expected to announce that Siri can route queries to Claude, Gemini, Grok, and other third-party AI models depending on the task. OpenAI's exclusive arrangement ends. Siri becomes a routing layer, not a destination.

Put those two things together: a device-level interface that can pass any shopping query to any AI model, managed by someone who built Google Shopping's AI layer. Apple hasn't announced a single commerce product. That's the tell.


Here's what Apple's device-layer position actually means in practice.

When an agent buys something on your iPhone, Apple already has the authentication. Face ID confirms you're you. Apple Pay holds the payment credentials. The Secure Enclave handles the cryptography. None of that requires Apple to build a competing shopping agent — it just requires them to open APIs to agents that already exist.

Google is assembling a coalition around UCP. OpenAI has retailer apps for Walmart and Target. Amazon has Rufus and 250 million monthly users. All of them need to solve identity and payment from scratch, or build relationships with Visa, Mastercard, and every bank.

Apple solved that problem fifteen years ago. They just haven't turned it on for agents yet.


The honest counter: Siri has been bad for a long time, and Apple Intelligence launched last year to mostly underwhelming reviews — so it's completely possible they fumble this the way they fumbled Siri's original promise.

And the multi-model routing idea — Siri as a neutral broker passing queries to whichever AI handles it best — is elegant in theory and a nightmare in practice. Who decides which model gets the query? How does that work when three different AI companies are competing to influence the same purchase? Apple has never been great at the messy platform politics that kind of system requires.

So this is a bet, not a certainty.

But the ingredients are there in a way they weren't twelve months ago. A hardware engineer at the top who thinks in terms of device capabilities. A Google Shopping veteran running the AI product team. A multi-model Siri that becomes a device-level routing layer. And two billion devices with payment credentials already on file.

Everyone's watching Google build commerce protocols and OpenAI sign retail partnerships. Apple hasn't said a word about any of this.

That's exactly what Apple does before it changes something.