John Ternus is now Apple's CEO, and his entire career has been spent making physical objects. He led the Apple Silicon program - the custom chips inside every modern Mac, iPad, and iPhone - and served as VP of Hardware Engineering since 2020. He has never run a services business.
That matters because Apple's last decade was largely Tim Cook's pivot toward services. The App Store, iCloud, Apple TV+, and Apple One subscriptions became the company's highest-margin lines. But on the AI side, Cook's tenure ended with Siri still lagging well behind ChatGPT and Google Gemini, and Apple Intelligence - the suite of AI features Apple announced in 2024 - receiving underwhelming reviews.
What a Hardware-First CEO Actually Changes
The case for optimism here is specific: Apple Silicon chips are genuinely good at running AI models locally, meaning on your device without sending data to a server. The M-series chips include dedicated neural engine cores that can run mid-size language models entirely offline. If Ternus pushes harder on chip performance, that directly improves on-device AI - which is Apple's stated competitive angle against Google and Microsoft.
For people using tools like Claude for Desktop on a Mac, faster Apple Silicon chips mean better local inference (the process of generating AI responses), lower latency, and stronger privacy since data doesn't leave your machine.
The concern is the other half of the job. Ternus has no public track record running software organizations or AI research teams. Apple's AI software has consistently trailed competitors regardless of how good the underlying hardware was. A faster chip doesn't fix a weak language model, and Ternus has never shipped one.
Apple is also deep in negotiations around integrating third-party AI models - including Claude and ChatGPT - directly into iOS and macOS. Those are software and partnership deals, not hardware problems. How Ternus handles that side of the role is the real unknown heading into his first year.