AMD buys Fei-Fei Li's World Labs for $8.2B in all-stock bet on physical AI

#What happened

Advanced Micro Devices agreed to acquire World Labs in an all-stock transaction valued at approximately $8.2 billion, the two companies said Monday (Sept 28, 2026), per TechCrunch, Reuters, and Barron's.

San Francisco-based World Labs was founded in 2024 by Stanford professor Fei-Fei Li, an AI pioneer known for building the ImageNet database. Li will join AMD as executive vice president and chief scientist, reporting directly to CEO Lisa Su.

The deal is expected to close before the end of 2026, subject to regulatory approval.

World Labs develops "world models" — deep learning models that generate, reconstruct, and simulate interactive 3D environments from text, images, and video. Its first commercial product, Marble, builds persistent 3D environments, and the startup recently launched the Atlas model, which predicts how a scene looks from a new camera angle from a few 2D images.

AMD and World Labs had formed an inference-optimization and model-training partnership last year, and AMD was already an investor. World Labs raised $1 billion in February from investors including AMD, Nvidia, Autodesk, and Fidelity.

In a Bloomberg Television interview, AMD CEO Lisa Su said the deal gives AMD "access to the world-class talent Fei-Fei has brought together" combined with AMD's hardware, software, and systems capabilities.

Li wrote in her announcement post: "Now that we have tangible proof of the possibilities, we want to do everything we can to accelerate the future. To do this requires scaling our efforts, widening our reach, and getting closer to the hardware."

#Why it matters

The acquisition is AMD's move to compete with long-standing rival Nvidia in an AI ecosystem beyond chips: Nvidia already offers open-weight world models like Cosmos, while AMD has only released text- and video-based open models, per TechCrunch.

Stifel analyst Ruben Roy rated the move a Buy with a $635 price target, and AMD shares rose 1.7% on Tuesday, outperforming a down Nasdaq, per Barron's.

The deal also signals the industry's bet that "physical AI" — models that understand and simulate the physical world for robotics, autonomous vehicles, and industrial robots — will drive future hardware demand, with synthetic data from world models seen as key to training general-purpose robots.

#Sources

• REPORT: TechCrunch, Sep 28
• REPORT: LA Post/Reuters, Sep 28
• REPORT: The Edge Markets, Sep 29
• REPORT: Barron's

Verified Sep 30, 2026