John Ternus on Apple's Invisible AI Strategy
Exactly two months ago, on September 1, John Ternus officially took the chair Tim Cook had occupied for eleven years as Apple CEO. Rather than a marketing executive or a finance outsider, Ternus is what the tech world calls a "product engineer" — a man who spent twenty-five years at Apple, starting as a software designer in 2001, working on the iPod team, and eventually arriving at the Director of Product Engineering, where he oversaw development of the MacBook Air, the iPad Pro and the iPhone lines.
Now, with generative AI redefining the industry every week, Ternus has given one of his first major interviews to Steven Levy at Wired, and in it he laid out what he wants to be Apple's philosophy for the next decade: invisible AI.
The quote that sums everything up
The most cited statement from the interview is not new — Ternus had said something similar in other contexts — but it gains new weight from the source and the moment. "We're absolutely going to leverage AI to do that," Ternus said about building incredible products that make users' lives better. "We don't want to ship AI for AI's sake. When we do it right, people don't have to think about whether it was AI or traditional software that made it happen. What matters is: did we build something they really love?"
The statement, seemingly modest, encapsulates a deliberate strategic bet against the industry's dominant current. Since 2023, every major technology CEO — from Sam Altman at OpenAI to Satya Nadella at Microsoft and Jensen Huang at Nvidia — has positioned generative AI as the biggest transformation since the internet. The result has been a race for autonomous agents, for ever-larger frontier models, and for marketing that slaps "AI" on everything from coffee makers to text editors.
Apple, until the iOS 27.3 launch in December with integration to Alibaba's Qwen, remained notably quiet.
Apple's bet: silent integration vs. the agent race
Ternus's strategy is, essentially, an argument that the most powerful AI is the kind nobody notices. It's not innovation in itself — the concept of invisible technology dates back to Apple's earliest days, when Steve Jobs proposed that the best computer would be one you didn't even need to know was a computer — but it's a risky bet at a time when the market punishes slowness. Apple's stock has fluctuated over recent months, reflecting investor uncertainty about whether the company is genuinely falling behind.
Ternus openly acknowledged that perception during the interview. When Levy suggested that "history may judge him on how he handles AI," Ternus did not minimize it. He admitted that "everyone has an opinion" about Apple's position in the AI ecosystem, especially after the delayed Siri update, finally launched this summer. But he did not shrug. "I really like where we're at," he replied.
The crucial difference between Apple's approach and those of direct competitors is structural. Companies like OpenAI and Google build models and distribute APIs — the value is in the model itself. Apple, on the other hand, builds hardware, operating system, assistant and apps all under one roof. This enables an integration that pure software companies cannot reproduce: AI not as a product but as a fundamental layer of experience.
When the user asks Siri to open a photo in Photos, they're not calling an API — they're using the Apple chip's Neural Engine, with data never leaving the device. AI is invisible because the entire device is the AI.
Why Ternus is the right choice — and dangerous
There are two types of CEOs Apple has had: product visionaries (Jobs, Wozniak) and execution masters (Cook, now executive chairman, and Ternus, in his place). Ternus is from the second camp. He doesn't have Jobs's mania for redesigning a product from scratch; he prefers to refine an existing one until it's perfect. This may sound like the opposite of what a technology company needs at a time of technological rupture — but it's exactly the right attitude for an invisible AI strategy.
Generative AI, in its current form, is noisy. Large models produce hallucinations, require constant broadband and cost hundreds of millions of dollars per month in inference. Apple can wait — and it is waiting. When Ternus mentioned that Apple Intelligence is coming to China in December (with Alibaba's Qwen model, after the ban on using OpenAI models in that market), it was clear that the company is building a modular AI infrastructure: some features run on-device, others in the cloud, some even through partnerships with external suppliers when needed. The idea isn't to have the best model; it's to have the best final product.
This strategy has a dangerous side. If Apple waits too long to launch features that competitors already delivered, the average user may simply prefer the competitor's product — and never look back. Samsung, Google and Microsoft have already integrated conversational agents into their phones and PCs. The iPhone, even though it's the world's most popular device, is losing ground in the perception of innovation. Memory chips, as Ternus told TIME, are a crisis — the global shortage of memory is driving up production costs and may delay launches.
The challenge Ternus faces
The question is: in two or three years, when autonomous agents — those that execute entire tasks without human intervention — become commonplace, will Apple's bet on invisible AI have been wise or belated? If agents replace entire apps, as Ternus suggests when he says "the creation of products is a deeply human and personal thing," will Apple be preserving what's essential or simply delaying the inevitable?
Ternus does not answer that question in the interview. Instead, he talks about products, user experience and security. But the very choice not to answer is itself an answer. Apple doesn't want to be the first to deliver the next great AI revolution — it wants to be the last to deliver something no one needs to learn how to use.
Sources: WIRED, 9to5Mac, AIWatcher
✓ Independent sources cross-checked and verified before publishing
