Google just handed over the keys to its AI engine room to one of its longest-serving technical visionaries. Amin Vahadat, a computer scientist whose fingerprints are on the company's most powerful computing systems, has been elevated to a new role as chief technologist for AI infrastructure, reporting directly to CEO Sundar Pichai.
The move, first revealed by Semafor and later confirmed by TechCrunch, shows how seriously Google is betting on the invisible machinery behind artificial intelligence. As Alphabet prepares to spend up to $93 billion in capital spending by the end of 2025, with even more spending expected next year, the company is making clear that its AI plumbing is just as important as the flashy model on top.
Who is Amin Vahdat?
If AI is the driver of modern technology competition, then infrastructure is its driver. And Amin Vahadat has been fine-tuning that engine for nearly 15 years. Since joining Google in 2010, he has become one of the company's most influential people, shaping the systems that keep large data centers running and AI models scaling.
Vahdat currently oversees everything from Google's custom Tensor Processing Units (TPUs), the chips that power Google's AI training and inference, to the Jupiter Network, its internal high-speed data transfer system that is the envy of the industry. In a post last year, he revealed that Jupiter is currently moving data at a staggering 13 petabits per second, which is enough for every person on Earth to make a video call at the same time, he noted.
At Google Cloud Next earlier this year, Vahdat made headlines by announcing the company's 7th generation TPU, codenamed Ironwood. He said the new chip had a whopping 9,000 processors per pod and could deliver up to 42.5 exaflops of computing power, roughly 24 times the performance of the world's fastest supercomputers at the time. “The demand for AI computing has grown 100 million times in just eight years,” Vahdat told the audience, revealing that Google's hardware ambitions are growing as fast as the AI boom itself.
Behind these headline-grabbing numbers lies the kind of work that rarely appears at glitzy launch events. Vahdat has long led teams that build custom data center technologies for Google, including Borg, the company's legendary cluster management software that silently allocates resources across thousands of machines. He also helped lead the development of Axion, Google's first custom Arm-based CPU designed for data centers, introduced last year.
For those outside of the cloud computing world, Vahdat may not be a household name. But within Google and the academic community, he is something of a legend. A graduate of the University of California, Berkeley with a PhD in computer science, he began his career as a research intern at Xerox PARC in the early '90s. Xerox PARC is the same innovation laboratory that gave the world graphical user interfaces and Ethernet.
Before joining Google, Vahdat spent several years in academia, serving as an associate professor at Duke University and then an SAIC professor at the University of California, San Diego. His research portfolio of nearly 400 papers revolves around one question: how can computers work together efficiently at scale?
That tenacity defined his career at Google. From designing distributed systems that can handle trillions of user requests every day to ensuring Google's cloud infrastructure can handle the growth of AI, Vahdat is a quiet behind-the-scenes force behind one of the world's most complex technology operations.
Keep key players close
Google's decision to promote Vahdat to an executive position is as much about strategy as it is about talent retention. In an industry where top-notch AI engineers are in constant demand and compensation packages routinely soar into the stratosphere, ensuring Vahdat stays on is a smart move.
This new role solidifies his position as Google's key architect of the future. Reporting directly to Pichai gives Vahdat a front-row seat on decisions that shape the company's AI roadmap, from the evolution of data centers to how Google aligns with rivals like OpenAI, Microsoft and Amazon.
Google's public AI story often revolves around product and generative model innovations like Gemini, but it's the invisible infrastructure that determines how far and how fast these systems go. And with Amin Vahdat officially taking charge of the mission, the biggest upgrade in Google's history could be happening behind the scenes, one chip, one server rack, and one petabit at a time.
– end
