AMD acquires World Labs AI startup, upping the ante against Nvidia
AMD will acquire Fei-Fei Li's world-model startup World Labs for $8.2 billion, challenging Nvidia's robotics simulation lead.
AMD announced it will acquire World Labs by year-end pending regulatory approval in a transaction valued at $8.2 billion. World Labs was founded in 2024 by Fei-Fei Li with Justin Johnson, Christoph Lassner and Ben Mildenhall, raising $230 million including AMD investment, and builds world models that simulate physical environments, shipping tools like Marble (Gaussian-splat 3D space generation) and Atlas. The deal targets synthetic training data for robotics, where AMD lacks a competitor to Nvidia's Cosmos stack despite its open-source Micro-World model.
- AMD to acquire World Labs for $8.2 billion by year-end, pending approval
- World Labs founded 2024 by Fei-Fei Li, raised $230 million including AMD money
- Products include Marble 3D generation tool and Atlas world models
- Acquisition targets Nvidia's lead in world models and robot training data
Full article240 words · extracted from arstechnica.com · click to collapse
Chipmaker AMD and world models company World Labs announced that AMD will acquire the AI company by the end of the year, pending regulatory approval. The transaction is valued at $8.2 billion.
In 2024, computer vision scientist Fei-Fei Li founded World Labs with fellow researchers Justin Johnson, Christoph Lassner, and Ben Mildenhall, with $230 million in funding, some of which came from AMD.
The company began working on new world models, which are AI models that aim to provide a useful, predictive simulation of the physical world. Competing architectures exist (earlier this year, Ars interviewed World Labs co-founder Ben Mildenhall and others in the field about exactly that), but many involve training a model on vast amounts of video data.
That was the case for Marble, World Labs’ first publicly usable tool. Marble can generate small 3D spaces using Gaussian splats. They can then be exported as 3D assets for use in film production, game development, and other applications. The company has since introduced additional models and tools with different emphasis or use cases, such as Atlas.
One use case that has justified enormous investment in the space is generating synthetic data for training robots. That’s a field where Nvidia, AMD’s direct competitor, has a strong lead. AMD has produced its own AI models before, including Micro-World, an open-source world model, but it has not produced a stack that competes with Nvidia’s Cosmos and related models and applications.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arstechnica.com/ai/2026/09/amd-acquires-world-labs-ai-pioneer-fei-fei-lis-world-models-startup/