By ByteBot
Apple announced today that John Giannandrea, Senior Vice President of Machine Learning and AI Strategy, is stepping down and retiring in spring 2026 after seven years at the company. His replacement is Amar Subramanya, who spent 16 years at Google leading engineering for Gemini Assistant before joining Microsoft as Corporate VP of AI. Consequently, the timing is no coincidence. Giannandrea’s departure comes after Apple delayed its promised AI-powered Siri features by 18 months to spring 2026, admitted it couldn’t deliver what it announced at WWDC 2024, and faced class-action lawsuits for marketing iPhone 16 with features that don’t exist.
This isn’t a retirement. It’s an admission of failure.
The Siri Disaster That Forced This Change
Apple promised a smarter, AI-powered Siri at WWDC 2024. The features sounded compelling: personal context, on-screen awareness, deeper app integration. Moreover, Apple marketed these capabilities heavily with the iPhone 16 launch. Then reality hit.
In March 2025, Apple admitted it could not release the promised version of Siri as planned. The features that were supposed to ship in iOS 18.4 got pushed to iOS 26.4, an 18-month delay. Furthermore, Apple’s explanation revealed fundamental problems: the V1 architecture wasn’t working well enough, so they had to rebuild everything on V2 architecture.
The fallout was immediate. Customers filed multiple class-action lawsuits for false advertising. Indeed, Apple had sold millions of iPhone 16 units on the promise of AI capabilities that don’t exist and won’t exist until 2026. Even worse, Apple had to disable its AI-powered news summaries after the system twisted headlines from The New York Times and BBC to display inaccurate facts. When your AI can’t handle news summaries without fabricating stories, you have a fundamental problem.
Hiring the Guy Who Built the Winning Assistant
Amar Subramanya’s hiring is the ultimate irony. Apple is replacing Giannandrea with the executive who led engineering for Google’s Gemini Assistant, the very product that outperforms Siri. Subramanya then moved to Microsoft as Corporate VP of AI, working on Copilot and other AI initiatives. Now he’s at Apple, tasked with fixing what Giannandrea couldn’t in seven years.
Additionally, Apple’s organizational changes are equally telling. Giannandrea reported directly to CEO Tim Cook. In contrast, Subramanya reports to Craig Federighi, Apple’s engineering chief. That’s a demotion for the AI leadership role. The remainder of Giannandrea’s organization is being split between Sabih Khan and Eddy Cue. Apple is dismantling the structure Giannandrea built.
This move signals desperation more than strategy. Nevertheless, Apple can’t develop competitive AI internally, so they’re raiding Google and Microsoft for talent. It’s like hiring the opposing team’s star player after losing every game for seven years.
How Far Behind Apple Actually Is
While Apple struggled with Siri, competitors poured $255 billion into AI and lapped them multiple times. Microsoft invested $80 billion in AI for 2025 alone, nearly doubling its $41.2 billion spending from 2023. Similarly, Google increased AI investment to $75 billion, a 2.3x jump from $32.3 billion in 2023. Microsoft’s $13 billion investment in OpenAI integrated ChatGPT across Azure, Office 365, and enterprise tools. Meanwhile, Google’s Gemini is outperforming GPT-4 in several benchmarks.
Apple, by comparison, was described as taking a “deliberate jog while others sprint.” The company dismissed generative AI during its mainstream adoption phase between 2022 and 2023, when ChatGPT launched and redefined user expectations. Consequently, Apple fell behind and spent three years playing catch-up. The gap isn’t closing. It’s widening.
Apple’s privacy-first, on-device-first strategy sounded principled, but it became a liability. However, privacy constraints proved more limiting than Apple’s integrated approach enabled. Meanwhile, Microsoft and Google built cloud-heavy AI systems that deliver capabilities Apple can’t match. As a result, Apple is now pursuing partnerships with Google Gemini for advanced Siri features because Apple’s own AI isn’t competitive.
What Went Wrong Under Giannandrea
Giannandrea joined Apple in 2018 after eight years at Google leading Machine Intelligence, Research, and Search. His mandate was clear: fix Siri and establish Apple as an AI leader. Nevertheless, seven years later, Siri barely improved. Bloomberg’s Mark Gurman reported that Tim Cook lost faith in Giannandrea’s ability to execute on product development. In March 2025, Apple ousted Giannandrea as head of Siri development, moving the responsibility to Mike Rockwell.
The problems compounded throughout 2025. Furthermore, an exodus of Apple AI team members began as engineers left for Google, OpenAI, and Microsoft. Robby Walker, the executive overseeing the delayed Siri project, also departed after the organizational restructuring. Giannandrea’s influence diminished as Apple split his organization across multiple executives.
This raises the critical question: Was Giannandrea the problem, or is Apple’s approach fundamentally flawed? Indeed, Giannandrea succeeded at Google. He built world-class AI systems that still power Google Search and Assistant. At Apple, he faced privacy constraints, organizational resistance, and a company culture that dismissed generative AI when it mattered most. Therefore, if Apple’s culture and constraints are the real blockers, Subramanya will fail for the same reasons.
The Real Test Comes in 2026
Apple’s bet is that new leadership can salvage its AI strategy. However, Subramanya inherits a delayed Siri launch, frustrated developers, lawsuits from customers, and a competitive gap measured in years and hundreds of billions of dollars. His timeline is brutal: deliver the promised Siri features in spring 2026 while rebuilding Apple’s AI foundation for the future.
The official Apple announcement was careful to thank Giannandrea for his contributions and position Subramanya as a visionary hire. Nevertheless, the timing tells the real story. This is Apple acknowledging that seven years of AI strategy under Giannandrea failed, and they’re starting over with external talent from the companies that beat them.
Developers in the Apple ecosystem remain stuck with inferior AI tools while Microsoft, Google, and OpenAI race ahead. Ultimately, whether Subramanya can close the gap or whether Apple’s approach dooms any leader to the same fate will become clear when those delayed Siri features finally ship in 2026. If they ship.
