World Labs
Builds a world model platform enabling AI to understand, simulate, and navigate real 3D environments.
Updated April 2026
Overview
- Website
- worldlabs.ai
- Founded
- 2024
- Headquarters
- San Francisco, CA
- Segment
- Image & Visual Generation
Product overview
World Labs develops large world models (LWMs) that enable AI systems to perceive, reason about, and interact with the 3D physical world, building on research from founders including Fei-Fei Li. Their technology aims to power embodied AI, robotics, simulation, and spatial computing applications. Distinct from LLM-centric labs by focusing on spatial intelligence as the core foundation for general-purpose AI that can act in the physical world.
Revenue model
Early-stage startup ($230M raised at $1B+ valuation); no public products or pricing disclosed as of early 2026; revenue expected from enterprise licensing of world model capabilities for robotics, simulation, and spatial AI applications.
Moat
World Labs' key competitive moat is its proprietary Large World Models (LWMs) and foundational technology in spatial intelligence, pioneered by an elite founding team of AI luminaries including Fei-Fei Li, Ben Mildenhall, Justin Johnson, and Christoph Lassner, who bring unmatched expertise in computer vision, Gaussian Splatting, and 3D rendering. This is reinforced by substantial strategic investments like Autodesk's $200 million, enabling deep collaboration on multimodal 3D world generation for applications in gaming, VFX, robotics, and digital twins, creating high barriers via specialized tech and ecosystem integration that outpace generalist competitors.