Apple, an "AI latecomer," joined hands with Google on the 12th (local time) to make up for its poor showing in the AI market. It decided to build its own AI system, including the voice assistant Siri, on Google's AI Gemini. The move was set in motion when Amar Subramanya, 46, a former Microsoft (MS) vice president in the AI division, joined Apple late last year.
He worked at Google for 16 years and took part in developing Gemini, and appears to have played a key role in sealing this alliance. Apple's new executive in charge of its AI business brought an alliance-driven AI strategy.
In early December last year, Apple was busy reshuffling its executive team overseeing AI development. John Giannandrea, then senior vice president of AI strategy, stepped down. Giannandrea also worked at Google from 2010 to 2018, but at Apple drew criticism for being late with final products that change the user experience, including Siri. That's why Apple earned the nickname AI latecomer. Apple entered Generative AI only two years after OpenAI's ChatGPT came out. With its own AI efforts lagging, Apple had, since late 2024, integrated OpenAI's ChatGPT into Siri and Apple Intelligence for complex queries, but repeatedly delayed the launch of the AI assistant "more personalized Siri" promised at the 2024 Worldwide Developers Conference (WWDC).
By recruiting Vice President of AI Subramanya, Apple put him in charge of core technology areas such as foundation models, Machine Learning research, and AI safety. It also set his reporting line to Craig Federighi, head of software. In other words, it aims to shorten the "AI development–test–deployment" process to "technology–product–launch." The goal is to bundle the operating system (OS), apps, services, and AI into one pillar to accelerate the AI transition. Conscious of criticism that Apple is late to AI, Subramanya, the new AI vice president, delivered results in less than two months of joining by forging an alliance with Google.
Vice President Subramanya is an AI researcher who studied AI systematically from school. He earned a bachelor's degree in electrical, electronics, and communications engineering from Bangalore University in India and a Ph.D. in computer engineering from the University of Washington in the United States. In his doctoral program, he studied semi-supervised learning (SSL) and graphical models. Semi-supervised learning is a method of training Machine Learning models using both partially labeled data and unlabeled data. He researched how to train AI systems efficiently when labeled data is scarce. In 2007, while in his Ph.D. program, he received an MS Research Graduate Fellowship. The fellowship provides funding and an opportunity for researchers to collaborate with MS on open research projects.
He began his career as a software engineer at IBM in 2001 and served as an MS intern and visiting researcher in 2005. He joined Google as a research scientist in 2009 and was promoted to principal engineer in 2017. In 2019, he was appointed vice president of engineering and worked closely with DeepMind, then Google's AI research unit. He also served as the engineering lead for Google's AI Gemini. Before moving to Apple, in Jul. 2025 he joined MS as vice president in the AI division. He worked on foundation models for MS's AI services such as Copilot.
What stands out in Vice President Subramanya's career are his stints at Google and MS. In particular, his role as engineering lead for Gemini likely worked to Apple's advantage in its alliance with Google. Foreign media reported, "Vice President Subramanya's expertise and network are expected to help integrate Google's technology into Apple's ecosystem as effectively as possible," and "The Apple-Google alliance was significantly influenced by Subramanya, who led the Gemini project, being appointed Apple's head of AI."
Given that MS achieved both technological advances in AI and commercial success with ChatGPT through its alliance with OpenAI, Subramanya's experience at MS could also be a positive factor.
Apple is expected to roll out a more personalized Siri through iOS 26.4, set to debut in March. However, it remains to be seen whether Google and Apple's alliance will last. After boosting the user experience by adopting an external model, Apple could seek to regain control once its own AI model reaches sufficient maturity.
Hong In-gi, a professor at Kyunghee University's College of Electronics and Information, said, "Apple has built its business model by creating a self-contained ecosystem, but in the AI world, which requires a lot of data, it appears to have judged there are limits and joined hands with Google," adding, "For device makers going forward, how well they optimize and apply AI will be the key battleground."