Business

Apple’s new CEO built the iPhone. The AI model that runs on it belongs to Google

When the world's most valuable company promotes its hardware chief to CEO, the signal is unambiguous: Apple believes the future of AI runs through devices, not data centers. That thesis arrives at its
Victor Maslow

John Ternus steps into Tim Cook’s role as the inheritor of one of the most durable corporate franchises ever assembled — and one of its most discussed strategic gaps. Ternus, 50, spent more than two decades at Apple overseeing the engineering teams behind the iPhone, iPad, Mac, AirPods, and Vision Pro. He has made more of the company’s physical products than almost anyone still employed there. What he has not made is a competitive AI model.

For consumers, the succession surfaces a question the company has largely deflected for three years: whether Apple Intelligence — the suite of on-device AI features Apple launched with iOS 18 — is catching up to what Google, Microsoft, and Meta are shipping, or drifting further behind. Apple enlisted Google’s Gemini to power parts of its AI offering after its own Siri development hit development snags, a public acknowledgment that its reasoning and voice layer needed outside reinforcement.

Tim Cook spent fifteen years turning Apple into the world’s most valuable company by a wide margin — market capitalization above $3 trillion, services revenue exceeding $100 billion annually, and a supply chain discipline that competitors spent years trying to replicate. His framework was tight integration: Apple owned the chip, the operating system, the distribution channel, and the retail experience in a way that let it extract margin at every layer. Ternus built much of the physical substrate of that integration. Whether integration-first thinking works when the central competitive variable shifts to cloud AI infrastructure is the question his tenure opens.

Microsoft, Google, and Meta committed hundreds of billions of dollars in combined capital expenditure to build AI data centers at a scale Apple never matched. Apple chose instead to bet on on-device processing and smaller, locally-run models — a genuine architectural advantage for privacy and latency, but not the same thing as a large-language model trained on internet-scale data. Ternus has not announced any change in that capital allocation approach. Both he and Cook said publicly in late July that no major strategic reset is coming.

Apple’s 160,000 employees are watching a managed transition, not a rupture. Investors have held the stock near record highs throughout the succession process. But Ternus’s first public examination comes within days of formally taking office: a September event is expected to center on the foldable device his engineering teams have been building for years — referred to internally as the iPhone Ultra — alongside the next standard iPhone lineup.

Apple reports fiscal fourth-quarter earnings in late October, the first full quarter Ternus will own as CEO. By then, the foldable will have been on sale for six weeks, Apple Intelligence’s new features will have either shipped with Gemini integration or hit further delays, and the analyst community will have its first concrete data on whether device-native AI can sustain the premium pricing that Cook’s integration thesis made durable for fifteen years. The hardware is ready. The AI question has a new name on it.

Tags: , , , ,

Discussion

There are 0 comments.