Big headline: Apple shakes up its AI leadership as John Giannandrea steps aside and a Google veteran steps in.
In an intentionally carefully worded press release on Monday, Apple announced that John Giannandrea, the company’s AI chief since 2018, will “step down” from Apple and remain with the company through spring as an adviser. He won’t be involved in day-to-day work after that period.
Taking his place is Amar Subramanya, a respected Microsoft executive who spent 16 years at Google and most recently led engineering for the Gemini Assistant. It’s a strategic hire, given Subramanya’s deep inside knowledge of Apple’s strongest competitors.
The move is being described as a shake-up, and it feels somewhat inevitable in hindsight. Apple Intelligence, the company’s response to the ChatGPT moment, has underperformed since its October 2024 debut. Reactions have ranged from lukewarm to genuinely concerning.
The early months were especially rough. A notification-summary feature intended to condense alerts into digestible snippets generated a string of embarrassing, incorrect headlines in late 2024 and early 2025. For instance, the BBC reported twice that Luigi Mangione, accused of killing UnitedHealthcare CEO Brian Thompson, had shot himself (he did not), and that a darts player, Luke Littler, won a championship before the final began (he hadn’t).
Then there was the promised overhaul of Siri, which ended up as a blemish for Apple.
A Bloomberg investigation published in May laid bare the scale of Apple’s AI struggles. One notable example: when Craig Federighi, Apple’s software chief, tested the revamped Siri on his own device just weeks before the planned April launch, he found that many advertised features didn’t work. The launch was delayed indefinitely, contributing to class-action lawsuits from iPhone 16 buyers who had been promised an AI-powered assistant.
By that point, Giannandrea had been pushed to the wings, according to Bloomberg. The outlet reported that Tim Cook had removed Siri from Giannandrea’s purview as far back as March, handing it to Vision Pro creator Mike Rockwell. Apple also removed its secretive robotics division from Giannandrea’s control.
Bloomberg’s reporting painted a picture of internal dysfunction: weak ties between AI and marketing, misaligned budgets, and a leadership crisis so deep that some staff joked about Giannandrea’s group as “AI/MLess.” The coverage also noted a brain drain, with AI researchers departing to competitors such as OpenAI, Google, and Meta.
There are even whispers that Apple now intends to lean on Google’s Gemini to power the next version of Siri—a startling turn given the intense rivalry between the two tech giants that stretches back more than a decade across phones, app ecosystems, cloud services, and beyond, now extending into AI.
Giannandrea arrived at Apple from Google, where he led Machine Intelligence and Search. In Apple’s orbit, he steered AI strategy, ML infrastructure, and Siri development.
Subramanya inherits those duties, reporting to Craig Federighi with a mandate to help Apple close the AI gap.
This moment lands at a curious point for Apple. Competitors are pouring billions into sprawling AI data centers, while Apple emphasizes on-device processing using its own silicon. This privacy-first stance minimizes data collection. When cloud processing is needed for more demanding tasks, Apple routes requests through Private Cloud Compute, aiming to process data briefly and erase it promptly.
The big question remains: will this on-device approach pay off, or will it leave Apple lagging behind the big cloud-based AI engines? The trade-offs are clear. On-device models are smaller and less powerful than the massive models running in rivals’ data centers, and Apple’s caution about data collection means researchers train on licensed or synthetic data rather than leveraging the vast real-world datasets that fuel competitors’ systems.
