Google’s response to the AI competition — elevate the person responsible for its data center technology
Google Appoints Amin Vahdat as Chief Technologist for AI Infrastructure
Google has taken a significant step in the ongoing competition to lead AI infrastructure by naming Amin Vahdat as its first-ever chief technologist for AI infrastructure. This new role, which reports directly to CEO Sundar Pichai, highlights the increasing importance of AI systems within the company. The move comes as Google plans to invest up to $93 billion in capital expenditures by the end of 2025, with parent company Alphabet anticipating even greater spending in the future.
Amin Vahdat’s Extensive Background
Vahdat is no stranger to the field. With a doctorate from UC Berkeley and early experience as a research intern at Xerox PARC in the 1990s, he has spent the last 15 years quietly shaping Google’s AI infrastructure. Before joining Google in 2010 as an engineering fellow and vice president, he held academic positions at Duke University and later at UC San Diego, where he was a professor and the SAIC Chair. Over his career, Vahdat has published nearly 400 research papers, consistently focusing on optimizing large-scale computing systems.
Driving Innovation at Google
Vahdat is already a prominent figure within Google. Less than a year ago, he introduced the company’s seventh-generation TPU, known as Ironwood, at Google Cloud Next. As VP and GM of ML, Systems, and Cloud AI, he revealed that each pod contains over 9,000 chips, delivering 42.5 exaflops of computing power—more than 24 times the performance of the world’s leading supercomputer at that time. “In just eight years, demand for AI computing has grown by a factor of 100 million,” he told the audience.
Building Google’s AI Backbone
Behind the scenes, Vahdat has overseen crucial projects that keep Google at the forefront of AI, such as developing custom TPU chips for AI training and inference, which provide a competitive advantage over companies like OpenAI. He has also led the creation of the Jupiter network, Google’s ultra-fast internal system that connects its servers and enables the transfer of vast amounts of data. According to Vahdat, Jupiter now supports up to 13 petabits per second—enough bandwidth for every person on Earth to join a video call at once.
Additionally, Vahdat has played a key role in advancing Borg, Google’s cluster management software that coordinates operations across its global data centers. He also supervised the development of Axion, Google’s first custom Arm-based CPUs for data centers, which were introduced last year and continue to evolve.
A Central Figure in Google’s AI Ambitions
Simply put, Vahdat is a cornerstone of Google’s AI initiatives.
Strategic Leadership and Retention
In an industry where top AI experts are highly sought after and generously compensated, Google’s decision to promote Vahdat to the executive level is also a strategic move to retain his expertise. After investing 15 years in developing his leadership, Google is ensuring that Vahdat remains integral to its AI strategy.
