Ian Buck

Ian Buck graduated summa cum laude in computer science from Princeton in 1999 and worked at NVIDIA as a systems engineer before starting a PhD at Stanford under Pat Hanrahan. As a graduate student he wired 32 GeForce cards together to drive the game Quake across eight projectors, then turned to a harder question: whether the parallel circuits built to shade pixels could run ordinary scientific code. His answer was Brook, a programming language that treated the graphics chip as a stream processor, described in the 2004 paper "Brook for GPUs." His thesis was titled "Stream Computing on Graphics Hardware."

NVIDIA hired him in 2004. Early on he managed the team behind the CUDA toolkit, compiler, libraries and driver, and on 8 November 2006 NVIDIA released CUDA, which let programmers write C for the GPU and which Jensen Huang insisted ship on every consumer GeForce card. The bet paid off years later, when neural network researchers found that CUDA made GPUs the cheapest way to train their models.

Buck became the company's link to the governments and cloud companies that bought those chips at scale. As general manager of accelerated computing he told a House oversight subcommittee in February 2018 that "A.I. is the biggest economic and technological revolution to take place in our lifetime," and asked Congress for more research funding, faster supercomputers and open government data. As vice president of hyperscale and high-performance computing he manages the hardware and software NVIDIA sells to Microsoft, Google, Amazon, Meta and OpenAI. At GTC in March 2026 he explained to reporters why NVIDIA had shelved its Rubin CPX chip in favor of racks of Groq 3 language processing units, from the company whose assets NVIDIA had bought in December 2025. Fortune put him on its 2025 list of the most powerful rising executives, and in September 2026 he was still speaking for NVIDIA in that role.