About World Labs
World Labs was founded in 2024 by Fei-Fei Li, a Stanford University computer science professor whose computer vision research includes her role creating ImageNet, a massive labeled image dataset that was an influential dataset in the development of modern computer vision starting in the early 2010s, along with co-founders including Justin Johnson and Ben Mildenhall, pursuing research into "world models," AI systems designed to develop a genuine spatial and physical understanding of three-dimensional environments, including how objects and spaces relate to each other and how physical scenes would look and behave from different perspectives or under different conditions, capabilities that the company argues are foundational to giving AI systems genuine physical and spatial intelligence relevant to robotics, gaming, architecture, and various other applications requiring an AI system to understand and reason about three-dimensional physical space rather than only processing two-dimensional images or text.
The company's research draws on Li's decades of computer vision research experience, applied to the increasingly prominent AI research direction of building models that can generate, understand, and reason about three-dimensional spatial environments, a capability distinct from and complementary to the large language models that have dominated much recent AI progress, which primarily process and generate text rather than spatial, three-dimensional understanding of physical environments. World Labs has released early demonstrations and research showing its models' ability to generate navigable, explorable three-dimensional scenes from simple inputs such as a single image, letting users or AI agents move through and interact with a generated three-dimensional space in ways that suggest potential applications spanning virtual reality content creation, video game environment generation, robotics training simulations, and various other applications requiring rich three-dimensional spatial understanding and generation capability. The company has positioned its world model research as an important complementary research direction to the large language model-focused research that has dominated much of the recent AI industry's attention and investment, arguing that genuine physical and spatial intelligence represents a similarly important but comparatively underdeveloped frontier of AI capability. World Labs has raised very substantial venture funding from investors including Andreessen Horowitz, Radical Ventures, and Nvidia, reflecting investor interest in both the world model research direction and Fei-Fei Li's research background.
