Apple’s artificial intelligence strategy is heading for its biggest shift in years after the company confirmed that John Giannandrea will retire in spring 2026.
The announcement has raised questions about Apple’s long-term AI direction, as the iPhone maker faces rising pressure from rivals, mixed reviews of Apple Intelligence, and investor concerns that the company has moved too slowly in advanced model development.
Giannandrea, who joined Apple in 2018 after a long career at Google, has been one of the most senior figures shaping the company’s push into AI.
BREAKING: Apple’s AI boss is stepping down
John Giannandrea, Apple’s Senior VP of Machine Learning and AI Strategy, will leave in spring 2026 after a transition period.
He oversaw Apple’s foundation models, ML research, and previously Siri and robotics, reporting directly to… pic.twitter.com/hWvxeJxgpq
— Perplexity Finance (@PPLXfinance) December 2, 2025
He reported directly to CEO Tim Cook and was responsible for machine learning research, Siri development, and the foundation model strategy that supports Apple Intelligence. Apple said he will remain as an adviser until his retirement date.
Rising Anxiety Over Apple’s AI Pace
Apple’s transition comes at a sensitive time. Analysts, developers and long-time Apple customers have said that the company is lagging behind Microsoft, Google and Meta, all of which are accelerating spending on data centres, custom chips, and frontier models.
Apple Intelligence launched in 2024 with the promise of returning the company to the front of the AI race. Reviews were mixed. Users described the features as limited and inconsistent, with the most anticipated improvement, a stronger and more conversational Siri, pushed back to 2026. That delay signalled engineering strain inside a division already under scrutiny.
Reports earlier this year suggested that Tim Cook had voiced concerns about the pace of Giannandrea’s development pipeline. Those concerns appear sharper now that Apple is installing a new leader with a different background and a reputation for rapid scale execution.
Who Is Amar Subramanya?
Reuters reports that Amar Subramanya, a veteran researcher who has spent decades shaping large-scale AI products at the biggest tech companies, is taking on Giannandrea’s role.
Subramanya recently worked at Microsoft as a corporate vice president of AI. He also spent 16 years at Google, where he played key engineering roles in search, voice technologies and Google’s assistant platforms.
🚨 BREAKING: Apple taps Google Gemini architect Amar Subramanya to replace John Giannandrea as AI chief. Subramanya has recently joined Microsoft as corporate VP. He will be reporting to Craig Federighi, Apple’s SVP of software unlike Giannandrea who reported to Tim Cook.
This… pic.twitter.com/OkAJPppRNX
— Sahil Mohan Gupta (@DigitallyBones1) December 2, 2025
One of his most prominent positions was head of engineering for the Gemini assistant, giving him direct experience with the type of conversational systems Apple has struggled to match.
He will report to Craig Federighi, Apple’s senior vice president of software engineering. According to Apple, Subramanya will lead foundation models, core AI research and AI safety. Several teams previously reporting to Giannandrea will move to other senior executives, including COO Sabih Khan and services chief Eddy Cue.
Industry watchers say the split indicates a major reorganisation aimed at increasing accountability and speeding up operations within Apple’s AI group.
Why Subramanya’s Appointment Matters
Subramanya’s hiring comes as Apple faces its largest gap against competitors in nearly two decades. Microsoft continues to push advanced features across Windows and its productivity apps.
Google is combining DeepMind research with Gemini to develop a family of models for mobile and desktop applications. Meta is training larger models at a pace few expected. All three companies are investing billions in AI infrastructure.
Apple, by contrast, is taking a different path. The company prefers to run its AI models directly on devices rather than rely heavily on the cloud. Apple hardware is built around efficiency, privacy and tight integration with its chips. That strategy reduces ongoing infrastructure costs but slows development of more computationally heavy models.
This approach helped Apple promote its AI as private and secure, which appeals strongly to many users. However, it places enormous pressure on Apple’s chips to perform at a level that rivals often achieve across entire data centre clusters.
Subramanya’s expertise in scalable model engineering may help Apple rethink which systems belong on the device and which need to run in the cloud.
Recent Moves Add to the Tension
Apple promised investors earlier this year that it would significantly increase its AI spending. The company also struck a partnership with OpenAI to integrate ChatGPT into parts of the Apple ecosystem, including Siri. Users will be able to trigger ChatGPT for queries that require a larger model than Apple currently supports on the device.
The shift has revived debate over Apple’s long-term ambitions for AI hardware. Among the most unexpected developments is the rapid progress of new devices being built by OpenAI and Jony Ive, Apple’s former design chief. Ive’s startup recently sold to OpenAI for $6.4 billion.
Ive and OpenAI CEO Sam Altman have said their prototypes are complete and could be revealed within 2 years.
Their collaboration adds competitive pressure to Apple’s hardware business, as new devices centred on generative AI could disrupt the market Apple has dominated for nearly two decades.
The next year will determine whether Apple can redefine its AI ambitions or fall further behind an industry that is accelerating faster than at any point since the launch of the iPhone.
Originally published on IBTimes UK
