Apple bets on john ternus to restore jobs-era decisiveness amid AI lag

At a glance:

• john ternus takes over as apple's chief operating officer, expected to centralise decision‑making and speed up AI development
• apple's AI models lag behind openai's gpt‑4o and meta's llama 4 scout; siri overhaul delayed three times, now slated for iOS 27 in 2026
• product pipeline includes a foldable iPhone, AI smart glasses, a homepad, and the m5 silicon generation, while vision pro sales have plunged 95%

Apple’s leadership shift and the promise of jobs‑era decisiveness

Bloomberg’s analysis, published the morning after apple announced the leadership change, frames the appointment of john ternus as a bet on “jobs‑era decisiveness.” tim cook’s consensus‑driven style generated $416 billion in annual revenue and a $4 trillion market cap, but analysts argue it has become too slow for the rapid choices required in artificial intelligence, product strategy, and a supply chain under geopolitical pressure. ternus, who previously led hardware engineering, is expected to act as a singular decision‑maker, centralising authority that cook traditionally spread across multiple senior executives.

The reorganisation already shows tangible changes. in early april, ternus overhauled the hardware engineering organisation around a new AI platform intended to accelerate product development. johny srouji, now chief hardware officer, split the combined hardware group into five divisions:

• hardware engineering – tom marieb
• silicon – sri santhanam
• advanced technologies – zongjian chen
• platform architecture – tim millet
• project management – donny nordhues this structure concentrates technical leadership while giving ternus a flatter reporting chain than cook maintained. Bloomberg also notes that ternus “opposed” both the vision pro headset and the cancelled autonomous‑car project “to varying degrees,” suggesting sharper instincts about what not to ship.

The AI deficit and siri’s delayed overhaul

Apple’s AI strategy is the most urgent challenge ternus inherits. the company’s server‑based model is rated behind openai’s year‑old gpt‑4o, and human raters preferred meta’s llama 4 scout over apple’s cloud model in image‑analysis tests. apple’s on‑device models run at approximately 150 billion parameters, while the custom gemini model licensed from google—announced in january—costs roughly $1 billion per year and is a 1.2‑trillion‑parameter mixture‑of‑experts architecture, eight times larger than anything apple has built internally.

the siri overhaul, originally targeted for ios 18 in 2024, has been postponed three times: to spring 2025, then spring 2026, and finally partially to ios 27 in september 2026. apple switched from a first‑generation architecture to a deeper end‑to‑end rebuild after finding the original version could not meet quality standards, forcing engineers to start over. the gemini‑powered version is expected to reach 1.5 billion daily siri users through ios 26.4, but the delays have already pushed the homepad smart‑home hub—from spring to autumn—because it depends on the upgraded assistant.

ternus himself described the AI challenge as patient: “i think apple intelligence is going to continue to grow, and it’ll just make things you do better and easier. if we’re doing it right, people won’t even really notice or think about it.” however, consumers remain unconvinced that AI features justify new hardware purchases, and the gap between apple and its rivals continues to widen. gene munster, longtime apple analyst now at deepwater asset management, said ternus has “an opportunity to supercharge aapl’s multiple by changing the narrative, which is the biggest opportunity in big tech,” and expects “big hires under ternus from ai‑focused firms like anthropic and openai.”

The ambitious product roadmap for 2026‑27

apple’s hardware pipeline for the next two years is the most ambitious the company has attempted in years, and it falls squarely within ternus’s domain of expertise. a foldable iphone is expected alongside the iphone 18 pro in september, featuring a book‑style design and an internal display roughly the size of an ipad mini, priced higher than current models. apple is testing at least four frame designs for ai smart glasses, targeting mass production in late 2026 or early 2027 with projected shipments of three to five million units in the launch year. the glasses will not be standalone; they rely on a connected iphone for processing, meaning their success is tied to the delayed siri overhaul.

the m5 generation of apple silicon is rolling out across the mac lineup. the vision pro, which shipped roughly 390 000 units in its launch year, saw sales plunge 95 % to an estimated 80 000‑90 000 units in 2025. apple has cut marketing spend on the headset by up to 95 % and is shifting engineering resources toward the smart‑glasses projects. the $2 billion acquisition of q.ai, an israeli silent‑speech ai startup, signals the kind of sensor‑driven, ambient ai that lightweight wearables could eventually enable.

Supply‑chain pressure and regulatory headwinds

the tariff environment has improved but remains volatile. the supreme court ruled in february that trump’s iiepa‑based reciprocal tariffs were unconstitutional, potentially triggering more than $175 billion in refunds to importers. trump immediately imposed a 10 % blanket tariff under section 122 with no product exemptions. apple imports more than $100 billion worth of goods from china annually and is accelerating plans to shift all us‑bound iphone production to india by the end of this year, which requires doubling its current indian manufacturing capacity.

in the eu, the european commission fined apple €500 million for dma anti‑steering violations in april 2025 and gave the company 60 days to comply. the coalition for app fairness accused apple of “persistent non‑compliance” six months later. interoperability requirements are forcing apple to open apis for nfc, default browser and app settings, and messaging, while apple intelligence itself remains unavailable in the eu due to the same dma constraints that have complicated its china rollout. repeat violations can carry fines of up to 10 % of global revenue.

in china, apple holds an 18.9 % market share behind huawei’s 20 %, but grew faster than any competitor in the first quarter with 33 % year‑over‑year gains driven by iphone 17 pricing and government subsidies. huawei’s domestic supply chain insulates it from the global dram shortage that is constraining apple’s production, while chinese competitors are integrating domestic ai capabilities that apple cannot match due to regulatory restrictions on apple intelligence.

Wall street’s cautious optimism

no major firm downgraded apple or cut price targets after the announcement. wedbush, jpmorgan, bank of america, melius, and evercore all reiterated their ratings, with targets ranging from $325 to $350. bofa said “the timing of the leadership transition suggests near‑term results are extremely resilient.” evercore said the appointment “makes sense given apple’s history of leadership rooted in the core hardware business.” morgan stanley warned that any ai strategy shift under ternus “is likely to be long‑term,” suggesting apple will avoid the aggressive ai spending seen at competitors.

the stock fell just over 1 % on the first trading day, a reaction analysts attributed to timing surprise rather than fundamental concern. raymond james offered the most measured assessment: the transition is “incrementally positive for product innovation” but “introduces heightened execution risk at a critical juncture.”

the closest historical parallel is satya nadella’s appointment at microsoft in 2014. nadella inherited a company that was not failing but was stagnating strategically, and led a platform transition to cloud that tripled its market capitalisation in five years. ternus inherits a company at or near peak financial performance but facing an equally fundamental question about its position in ai. the difference is that nadella was a cloud and services leader taking microsoft into cloud, whereas ternus is a hardware leader being asked to solve a software and ai problem. whether that mismatch is a vulnerability or an asset, whether the person who builds the devices is best positioned to determine what intelligence runs on them, is the bet apple’s board has made. the next 12 months, starting with wwdc on 8 june, will reveal whether it was the right one.