Apple has announced the most significant change to its Artificial Intelligence (AI) group since the 2024 launch of Apple Intelligence, confirming that long-serving AI chief John Giannandrea is stepping down.

The move comes as the company faces increasing pressure and criticism for falling behind its major tech rivals in the AI race.

Giannandrea, who has held the position of Senior Vice President for Machine Learning and AI Strategy since joining in 2018 and reported directly to CEO Tim Cook, will transition to an advisor role before fully retiring next spring (2026).

The New AI Chief: Amar Subramanya

Apple has hired a high-profile industry veteran to take the reins, signaling a fresh focus on foundational research and talent acquisition. Amar Subramanya will join Apple as Vice President of AI. Subramanya brings a wealth of relevant experience, having most recently served as a Corporate Vice President of AI at Microsoft. Crucially, he spent 16 years at Google, where he was the head of engineering for Google’s Gemini Assistant before his brief stint at Microsoft. He also worked at Google’s DeepMind AI unit.

Subramanya will now report to software chief Craig Federighi, marking a key organizational change. Previously, Giannandrea reported directly to Tim Cook.

Apple confirmed that Subramanya will lead critical areas, including the development of Apple Foundation Models, machine learning research, and AI safety and evaluation. The remaining teams previously under Giannandrea will be shifted under COO Sabih Khan and services chief Eddy Cue for better organizational alignment.

The shake-up occurs during a period of intense scrutiny over Apple’s AI execution. The suite, intended to put Apple alongside leaders like OpenAI and Google, has received lukewarm reviews from users and critics since its launch. One of Apple Intelligence’s centerpiece features—a significantly improved and personalized Siri assistant—was delayed until 2026, signaling internal development struggles.

CEO Tim Cook acknowledged that Federighi “has been instrumental in driving our AI efforts, including overseeing our work to bring a more personalized Siri to users next year.”

Although Apple’s stock is up 16% in 2025, it has lagged behind many other big tech companies. Investors worry that Apple is spending “much less” on the necessary cloud infrastructure, data centers, and frontier models compared to rivals like Microsoft, Google, and Meta, as its strategy prefers to run AI on-device.

Adding competitive pressure is the emergence of new AI-driven hardware. Jony Ive, Apple’s legendary hardware designer, sold his startup io to OpenAI for $6.4 billion earlier this year. Ive and OpenAI CEO Sam Altman have since collaborated on AI-driven hardware prototypes, which they intend to reveal in two years or less, potentially challenging the iPhone’s dominance built since 2007.

To address its competitive gap, Apple has increased its spending on AI and struck a key deal with OpenAI to integrate ChatGPT into some of its products, including Siri. The hiring of Subramanya, a leader experienced in commercializing foundation models at the industry’s top competitors, is a decisive move to reset and accelerate Apple’s internal AI development strategy.

Inside Amar Subramanya’s Trailblazing Run from Google Gemini to Apple’s AI Hot Seat

Apple’s appointment of Amar Subramanya as its new Vice President of AI is not just a standard executive change; it is a high-stakes talent coup that brings a chief architect of two rival AI superpowers—Google and Microsoft—to the helm of a division criticized for falling behind.

Subramanya’s career trajectory is a direct map of the global AI race’s most competitive battlegrounds. His deep expertise in turning cutting-edge Machine Learning (ML) research into products that scale to billions of users is precisely what Apple, grappling with the delay of its next-generation Siri until 2026 and lukewarm reviews for Apple Intelligence, desperately needs.

Google: The Architect of Gemini

Subramanya spent 16 years at Google, rising through the ranks to a senior leadership position defined by the company’s highest-profile AI initiative. He served as the Head of Engineering for Google’s Gemini Assistant (the successor to the Google Assistant), where he gained critical, practical expertise in building and deploying advanced AI products.

The Gemini project represents Google’s definitive effort to integrate its foundational Large Language Models (LLMs) into a highly interactive, multimodal consumer application.

His tenure included time within Google’s DeepMind AI unit, one of the world’s most influential AI research groups. This exposure gives him invaluable insight into the theoretical and engineering requirements necessary to develop state-of-the-art foundation models—the very core technology Apple needs to master.

His work at Google was centered on scaling complex machine learning and natural language processing systems, establishing a reputation as a researcher-turned-builder capable of productizing theoretical AI advances.

Microsoft: The Target of the Talent War

Subramanya’s stopover at Microsoft was brief but highly significant, placing him at the heart of Redmond’s aggressive AI expansion. He joined Microsoft in mid-2025 as Corporate Vice President of AI, reporting directly to Mustafa Suleyman, the CEO of Microsoft AI and co-founder of DeepMind.

His hiring was part of an aggressive talent acquisition strategy by Microsoft, which reportedly poached over 24 engineers and researchers from Google/DeepMind in a concentrated effort to bolster its consumer AI products.

At Microsoft, Subramanya was focused on developing advanced foundation models to drive AI-powered products such as Microsoft Copilot, which currently relies heavily on OpenAI’s GPT models. His experience was expected to accelerate Microsoft’s capability to build proprietary models and reduce its external dependency.

Upon joining Microsoft, he publicly praised the culture as “refreshingly low ego yet bursting with ambition”—a comment widely interpreted as a contrast to the competitive internal dynamics at Google. This unique perspective on both companies’ corporate cultures makes him an ideal leader to address the internal execution and collaboration challenges that have plagued Apple’s AI team.

OpenAI has added a fresh layer to its fast-expanding web of strategic deals, announcing on Monday that it is taking an ownership stake in Thrive Holdings — a company launched in April by Thrive Capital, one of OpenAI’s biggest and most influential backers.

The move widens OpenAI’s ongoing push to diversify its revenue streams at a time when the company is still battling to turn its wildly popular ChatGPT product into a profitable business. By embedding itself directly inside operating businesses across the “real economy,” OpenAI is placing bets that its models can unlock new efficiencies, reshape traditional workflows, and ultimately return financial gains that could support its heavy compute needs.

Thrive Holdings focuses on buying, owning, and running companies that could benefit from new technologies like AI, with an early focus on accounting and IT services. As part of the deal, OpenAI will embed engineers, researchers, and product teams inside Thrive Holdings’ companies, aiming to speed up their AI adoption and cut operating costs.

OpenAI, currently valued at $500 billion, did not disclose how much it invested. But the structure of the partnership means the size of OpenAI’s stake can expand as Thrive Holdings’ companies grow. A person familiar with the agreement — who asked not to be named because the details are private — said the setup is designed to align long-term incentives. Another person familiar with the deal said it also serves as a way for OpenAI to be compensated for its services.

Joshua Kushner, CEO and founder of both Thrive Capital and Thrive Holdings, called the partnership a natural extension of their long-running relationship.

“We are excited to extend our partnership with OpenAI to embed their frontier models, products, and services into sectors we believe have tremendous potential to benefit from technological innovation and adoption,” he said in a statement.

Brad Lightcap, OpenAI’s COO, framed the partnership as a real-world testing ground for how AI can transform core business operations.

“This partnership with Thrive Holdings is about demonstrating what’s possible when frontier AI research and deployment are rapidly deployed across entire organizations to revolutionize how businesses work and engage with customers,” he said.

It’s another example of OpenAI’s increasingly circular dealmaking strategy: the company invests in partners, those partners expand their AI infrastructure or operations using OpenAI’s products, and OpenAI gains both strategic influence and a potential revenue backstop. In recent months, it has taken positions in infrastructure players like AMD and CoreWeave, which supply the compute backbone that its own models depend on.

The announcement also arrived alongside a second major update: OpenAI revealed that ChatGPT Enterprise will be deployed to “tens of thousands” of employees at Accenture. The consulting giant is already one of the most aggressive adopters of generative AI in the corporate world, with its own teams dedicated to helping clients integrate tools like ChatGPT into daily operations.

The dual announcements show an evolving playbook — one that goes far beyond model releases and consumer subscriptions for OpenAI. With escalating compute demands, rising operating costs, and investors expecting OpenAI to one day convert its global popularity into sustainable earnings, the company is working to deepen its roots inside businesses where AI can drive measurable, billable change.

The Thrive Holdings deal drops OpenAI more directly into sectors where AI adoption has barely scratched the surface. And if the partnership works as designed, the company won’t just sell AI — it’ll partly own the gains created by it.