Ternus Takes Over Apple Tomorrow — and Bloomberg Says AI Is Job No. 1

Tomorrow is the first day of the rest of Apple’s life. On September 1, 2026, John Ternus — the hardware engineering chief who oversaw every Apple silicon transition from the M1 through the M6 — officially becomes Apple’s chief executive officer, and Tim Cook moves upstairs to the executive chairman seat he has held in waiting since the April announcement. The handover was choreographed months in advance, but the timing still lands with unusual dramatic force: within days of taking the job, Ternus will stand on the Steve Jobs Theater stage for Apple’s September 9 “Surprise and shine” event, his first major product launch as CEO, widely expected to introduce the iPhone 18 Pro line and Apple’s first foldable iPhone.

But according to a Bloomberg analysis published as the handover week began, the launch event is not the real story. AI catch-up is Job No. 1.

The inheritance

Ternus inherits the most profitable consumer hardware franchise in history — and the most awkward AI position of any major technology company. Apple was late to large language models. It fumbled the original Siri rewrite so badly that the more capable assistant was delayed roughly eighteen months. It advertised features that did not exist, and paid for it: in May 2026, Apple agreed to a $250 million settlement of a U.S. class action over the delayed Siri capabilities, with eligible iPhone owners in line for up to $95 each and a final claims process now underway. Its partnership with OpenAI, once billed as the centerpiece of Apple Intelligence, has since soured into litigation. And its answer to the model gap — a deal to power the new Siri with Google’s Gemini — was announced back in January and only began shipping in earnest at WWDC in June.

That is the backdrop for Bloomberg’s framing. The outlet’s Mark Gurman reports that when Ternus takes the reins on September 1, he “won’t have much time” before the company begins confronting a stack of inherited problems: an AI effort that still trails rivals in capability and perception, rising component costs squeezing margins, and internal questions about staff retention as rivals poach Apple engineers with frontier-lab compensation packages.

Why the AI deficit is structural, not cosmetic

It is tempting to read Apple’s AI struggles as a string of bad launches. The deeper issue is architectural. For a decade, Apple’s differentiation has been vertically integrated silicon plus a privacy-first software stance — on-device processing first, cloud second. The generative AI era inverted that calculus: frontier models are trained and served from data centers at a scale Apple never built, and the brightest consumer AI surfaces (chatbots, agents, coding tools) live in the cloud, on other companies’ infrastructure.

Apple’s answer has been a hybrid: keep on-device models for everyday intelligence, lean on partners for heavy reasoning. The Gemini-powered Siri announced at WWDC 2026 is the clearest expression of that strategy — users are expected to be able to choose among outside models, with Google’s Gemini as the flagship option, while Apple keeps the on-device layer private. Whether that satisfies users who have waited years for a genuinely smarter assistant is the open question Ternus must answer.

There is one genuine bright spot, and it is squarely in the new CEO’s home territory: Apple silicon. The M6 became the world’s first 2nm smartphone-class silicon, and the quad-die M5 Ultra brought 512GB of unified memory at 1.2TB/s to the Mac Studio — enough to run large open-weight models locally. Nvidia now reportedly views Apple as its principal rival in local AI processing, a remarkable reversal for a company that spent 2024 and 2025 as the AI industry’s favorite punchline. OpenAI has bought tens of thousands of Macs for reinforcement-learning workloads; Anthropic rents Mac capacity on AWS. The hardware is no longer the problem. The product layer above it is.

The margin squeeze AI is causing — and not yet curing

The cruel irony of Apple’s AI position is that the boom is already hurting the company financially before it helps. Memory prices have roughly quadrupled amid AI data center demand, and Tim Cook spent his final quarters as CEO warning that price hikes across the lineup are “unavoidable.” Unlike hyperscalers, Apple cannot offset component inflation with AI revenue: it sells boxes, not tokens. Every iPhone 18 must carry the memory bill that somebody else’s training run caused.

That squeeze defines Ternus’s first year. He must either raise prices into a soft consumer market — with the iPhone Fold reportedly starting near $1,999, the ceiling is already being tested — or compress margins in the quarter when investors will judge him most harshly.

What to watch

Three signals in the next two weeks will tell us how Ternus’s Apple will fight the AI war:

1. September 9 event substance. If the Siri/Gemini integration is presented as a first-class, deeply-demoed feature of the iPhone 18 Pro — not a footnote — it signals AI has moved to the center of Apple’s product narrative under the new CEO.
2. The foldable as an AI device. A $2,000 foldable needs a justification beyond the hinge. If Apple positions the larger screen as an agentic workspace — a place where a capable assistant actually does things — the narrative shifts from catch-up to category creation.
3. Talent moves. Watch whether Ternus restructures the AI organization or makes a marquee hire in his first 100 days. Gurman’s reporting has repeatedly flagged retention as a live issue; a decisive org chart move would be the clearest proof of the “Jobs-era decisiveness” Apple’s board reportedly hired him to bring.

The stakes

Apple turns 50 this year with roughly $400 billion in annual revenue and the industry’s most valuable brand — and the least credible AI story of any trillion-dollar company. Ternus’s mandate is not to win the model race outright; nobody expects Apple to out-train Google. It is to make AI feel like Apple again: private, instant, and obviously useful, running on the best consumer silicon on earth. He has the silicon. As of tomorrow, he has the keys. The September 9 event will be the first public test of whether the two can finally be fused.

The transition itself is the easy part. It was announced in April, planned for months, and Cook stays on as executive chairman. What begins tomorrow is harder: the moment Apple’s AI decade has to start earning its keep.