AMD has announced it has signed a deal to acquire the spatial computing startup World Labs for $8.2 billion.
World Labs was founded in 2024 by Fei-Fei Li, an AI pioneer who brought about rapid change in the field through her pioneering work in computer vision and deep learning, alongside co-founders Ben Mildenhall and Justin Johnson.
The firm produces world models, AI models that can accurately recreate real-world environments, in the pursuit of spatial intelligence or the ability to generate, perceive and interact with 3D environments in real time.
The transaction is expected to close by the end of the year, subject to regulatory approval, at which point Li will join AMD as its new executive vice president and chief scientist.
“Building the compute platforms for the next generation of AI requires a deep understanding of how models are evolving,” said Dr. Lisa Su, chair and chief executive of AMD.
“Fei-Fei and the World Labs team bring exceptional research leadership and model expertise. 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.”
World Labs that multimodal large language models can only provide a limited view of the world, and that breakthroughs across multiple scientific disciplines require AI that can accurately simulate the real world. This, it says, is where its spatial intelligence models come in.
It raised $230 million in its first year, following this with a $1 billion funding round that closed in February led by firms including AMD, Autodesk, and Nvidia. In July, it acquired the robotics company SceniX, stating that world models and spatial intelligence will help train generally useful robots.
In a post on Substack, Li said that World Labs is already seeing appetite for its model Atlas, which can produce 3D models of a scene using a single image, from sectors including robotics, design, construction and entertainment.
She added that World Labs established a technical partnership with AMD last year, training and inferencing using the firm’s GPUs, and that the deal is a natural extension of this software and hardware partnership.
“Together we’re committed to keep pushing the frontier of AI to solve real problems, and doing so openly with the community and the entire ecosystem,” Li wrote.


