Advanced Micro Devices (AMD) announced Monday it will acquire World Labs, a San Francisco-based AI startup cofounded by Stanford professor Fei-Fei Li, in an all-stock deal valued at $8.2 billion.
The transaction marks AMD’s second-largest acquisition on record, following its $50 billion purchase of Xilinx in 2022. Under the agreement, Li will join AMD as executive vice president and chief scientist, reporting directly to CEO Lisa Su.
World Labs specializes in developing “world models”—AI systems capable of generating and reconstructing interactive 3D environments from text, images, or video. The technology is designed to simulate real-world scenarios for applications such as robotics training, factory optimization, and scientific research. Earlier this month, the company unveiled Atlas, an AI model that predicts scene transformations from limited 2D inputs, comparing its function to how large language models generate subsequent words in a sentence.
AMD stated the acquisition will provide deeper insight into next-generation AI model development, enabling the chipmaker to design computing systems tailored to emerging workloads. The company plans to keep World Labs operationally separate until the deal closes later in 2024.
Li, widely regarded as a pioneer in AI, is credited with creating ImageNet, a foundational dataset that advanced computer vision by demonstrating the impact of large-scale labeled data. She previously served as chief scientist for AI and machine learning at Google Cloud and cofounded Stanford’s Institute for Human-Centered Artificial Intelligence.
The deal follows a broader trend in the tech industry, where major companies are acquiring AI research firms to secure talent and intellectual property. Recent examples include Nvidia’s $33 billion investments in Groq and Hugging Face, OpenAI’s $6.4 billion acquisition of io, and Meta’s $14 billion investment in Scale AI. AMD had previously invested in World Labs prior to the full acquisition.
Industry analysts note that world models represent a potential leap toward physically grounded AI, where systems can better understand and interact with the physical world. AMD’s move aligns with its strategy to compete with Nvidia, the current leader in AI chip sales, by expanding its capabilities in AI infrastructure beyond hardware.
The transaction is expected to close in late 2024, pending regulatory approvals.