AMD closes its $8.2 billion World Labs deal, and Fei-Fei Li becomes its chief scientist

Text
iHOLYSHITVISUAL / TECHCompany newsiHOLYSHIT / ILLUSTRATION
iHolyShit illustration

AMD announced on November 16 that it has completed its acquisition of Fei-Fei Li's World Labs. Li becomes executive vice president and chief scientist, reporting to CEO Lisa Su. The roughly $8.2 billion all-stock deal closed ahead of the end-of-2026 target.

A chipmaker buying an AI model company and making its founder the chief scientist is not something you see every day. On November 16, AMD, the US chip company based in Santa Clara, California, announced that it has completed its acquisition of World Labs, an AI company in San Francisco. World Labs co-founder and CEO Fei-Fei Li joins AMD as executive vice president and chief scientist, reporting directly to AMD chair and CEO Lisa Su.

Some background. On September 28, AMD agreed to buy World Labs in an all-stock deal valued at about $8.2 billion, and said it expected to close by the end of 2026. It closed about six weeks later, well ahead of that target. World Labs was founded in 2024 and builds so-called world models: AI that turns text, images and video into 3D spaces you can interact with, with uses in robot training, factory simulation and scientific research. In early September it released a new model called Atlas, which can take a few 2D pictures and predict what the same scene looks like from a different camera angle.

Li is one of the best-known names in AI, often called the godmother of AI. She created ImageNet, the large labeled image collection that helped show that more data makes computers much better at recognizing pictures. She later served as a vice president at Google and chief scientist of AI and machine learning at Google Cloud, and she is a founding co-director of Stanford's Institute for Human-Centered AI. World Labs' other leaders, Justin Johnson and Ben Mildenhall, will keep leading the team with her.

Why would a chip company want this? Su's explanation in the announcement is that building the infrastructure for the next era of superintelligence means optimizing computing platforms and models together. In plain terms, how good an AI chip is depends on the models it runs. A chipmaker with its own frontier model team can see earlier what future models will demand from chips, software and systems, and design products accordingly. The two companies already worked together: World Labs began optimizing model training and inference on AMD GPUs last year. AMD says the World Labs team will keep doing foundational model research while working with AMD engineers to apply what it learns to future AI infrastructure for customers and developers in what AMD calls the open AI ecosystem. AMD's main rival in AI chips is Nvidia.

For everyday users, nothing new ships right away; this is a bet on the future. Three things to watch: whether World Labs' world models start running on AMD chips, whether AMD launches hardware tuned for 3D and robotics AI, and whether Li's team keeps publishing models openly. The company has not given timelines for any of them.

Share this article

What do you think of this story?

Anonymous feedback, up to 5 reactions per article.

Reader comments

Comment without signing in, or sign in to display your member name.

Discuss the technology and the story. Hate, politics and war, threats, abuse, personal accusations, sexual content, ads, links and personal information are not allowed.

0/500

No comments yet. Be the first to share.

Sources