AMD and World Labs announced today that AMD will acquire World Labs, a company that develops deep learning models aimed at understanding physical reality, in a transaction valued at $8.2 billion.
As part of the deal, World Labs founder Fei‑Fei Li will join AMD as executive vice president and chief scientist. The two companies have worked together before: they formed an inference optimization and training partnership last year and have maintained close ties since; Li was a guest at AMD’s CES presentation earlier this year.
Fei‑Fei Li, a Stanford computer science professor known for her pioneering work in computer vision — including the creation of the ImageNet dataset and the competitions it spawned — founded World Labs in 2024. The company’s stated mission is to build deep learning models with a stronger grounding in the physical world. Li has argued that true general intelligence requires an understanding of physics and the ability to reason about non‑textual data.
World Labs’ first product, Marble, is marketed as a tool for creating entertainment experiences and for generating simulated environments useful for robot training.
The term “world model” covers a wide range of systems, from language models trained to interpret visual inputs to models capable of producing and maintaining high‑fidelity simulations of reality. World Labs’ work sits in this space, and the acquisition is framed as a way to better integrate model research, systems design, and compute hardware.
Strategically, the deal is likely to bolster AMD’s position against long‑time rival Nvidia in building an ecosystem around AI‑specific chips. Nvidia already offers open‑weight world models such as Cosmos, while AMD’s public offerings to date have been focused on text and video models.
World models are considered important for deploying generative AI on robotic platforms — from autonomous vehicles to industrial robots and general‑purpose humanoids. Because real‑world training data for general‑purpose robots is limited, synthetic data produced by world models is expected to play a key role in realizing visions put forward by companies like Tesla and Figure.
The companies said the acquisition is expected to close before the end of the year, subject to regulatory approval.
