AMD will pay $8.2 billion for World Labs, the spatial AI startup founded in 2024 by Stanford professor Fei-Fei Li, and will bring the researcher widely known as the "Godmother of AI" onboard as Chief Scientist reporting directly to CEO Lisa Su.

The deal, announced this week, is expected to close by the end of 2026, subject to standard closing conditions. It gives AMD not just a research figurehead but a working software asset: World Labs' Marble family of foundation world models, built for 3D simulation and spatial reasoning.

The acquisition lands while AMD's AI-driven hardware business is running hot. The company posted a 50% revenue increase in its most recent quarter and is on track to record more than $40 billion in revenue for the year, driven by demand for Instinct GPUs and Epyc CPUs. Its market capitalization sits just under $1 trillion, making it the 15th most valuable company in the world.

From ImageNet to Instinct

Li's credentials run deep in the field. She began building ImageNet in 2007 — a dataset of more than 14 million images used for training and validating visual machine learning algorithms — and the annual ImageNet contest became the proving ground for computer vision. In 2012, Geoff Hinton and his University of Toronto students flattened the competition with AlexNet, a GPU-based neural network whose win helped ignite the deep learning cycle that Nvidia CEO Jensen Huang later termed the "Cambrian Explosion of AI."

Born in China in 1976, Li immigrated to New Jersey at 12. She earned a physics degree from Princeton and a doctorate in electrical engineering from Caltech, then returned to Princeton as an assistant professor from 2007 to 2009, the period in which she founded ImageNet. She joined Stanford in 2013 as co-director of the Stanford Institute for Human-Centered Artificial Intelligence and co-director of the Stanford Vision and Learning Lab. A sabbatical in 2017–2018 took her to Google Cloud as Chief Scientist of AI/ML, and back at Stanford in 2023 she established RAISE-Health, a program for responsible AI in healthcare.

World Labs, founded in 2024 to push computer vision, deep learning and robotics, raised more than $1.2 billion in funding before the sale.

Why the deal makes sense for AMD

The acquisition formalizes a relationship that already existed. In 2025, World Labs signed a technical partnership with AMD aimed at optimizing its model training and inference workloads on AMD GPUs.

"As our teams worked together, we realized it would be a natural fit to bring together our AI ecosystem of software and hardware, foundation models, and applications," Li wrote in a September 28 Substack post.

Her rationale for the sale echoes her rationale for founding the company in the first place: that language models alone cannot carry AI forward.

"We built the company guided by the foundational belief that language alone isn't sufficient for model development," Li wrote. "The universe isn't made up of words; it's made of real things. So many of the most important problems for AI to solve, from science to entertainment to robotics, require models to reason over the structure and behavior of the physical world."

"Just like I knew it was time to do a startup to jumpstart the work of creating foundation models beyond LLMs and achieve the most impact in AI, it's again very clear that the right path for myself and World Labs is to continue our mission as a part of AMD," she wrote.

Su framed the deal in terms of AMD's hardware roadmap. "Fei-Fei and the World Labs team bring exceptional research leadership and model expertise," Su stated. "Together, we can use that insight to develop the hardware, software and systems that will power the next generation of AI and strengthen the open AI ecosystem."

The competitive read

The move comes as AMD's chief rival moves on the software side as well: Nvidia is in the process of buying Hugging Face. For AMD, the acquisition addresses a persistent gap versus Nvidia — software and model ecosystems — by bringing foundation model development in-house under a researcher with a track record of reshaping the field. If the bet pays off, AMD's Instinct platform gains spatial and physical AI workloads, not just LLM training and inference, as a demand driver beyond 2026.