Fei-Fei Li’s World Labs Raises $1 Billion From Investors Including NVIDIA and AMD
World Labs was co-founded in 2024 by Stanford University professor and computer-vision pioneer Fei-Fei Li. The startup focuses on large world models, or LWMs, and “spatial intelligence.” These models are designed not only to generate text or images but also to understand, construct and manipulate three-dimensional environments. Potential applications span design, robotics and scientific simulation, making the technology a key bridge between generative AI and the physical world.
World Labs announced a new $1 billion funding round on February 18, 2026. Investors included NVIDIA, AMD, Autodesk, Andreessen Horowitz, Fidelity, Emerson Collective and Sea. Autodesk contributed $200 million and is partnering with the company to integrate world models into 3D design workflows.
All Coverage
2 original reportsThe Backstory
The history behind this eventFei-Fei Li’s World Labs Unveils Atlas World Model
World Labs, co-founded by computer-vision pioneer Fei-Fei Li, emerged in September 2024 with $230 million in financing to pursue “spatial intelligence”: AI systems that can understand three-dimensional environments, object interactions and change over time. Unlike conventional image or video generators, world models are designed to generate, reconstruct and simulate environments, potentially supporting filmmaking, game development, design and robot training. Atlas marks a push toward a general-purpose model spanning those tasks.
World Labs unveiled Atlas on Sept. 1, 2026, describing it as a multimodal autoregressive diffusion transformer pretrained from scratch to work natively with text, images, video and 3D data. The model can generate as much as one minute of 1440p video along specified camera paths, reconstruct scenes from sparse views, and turn footage from three to five ordinary phones into “bullet time” sequences. It can also produce point clouds and 3D Gaussian splats and support real-to-simulation workflows for robotics. Atlas is initially available in early access to selected partners, with pricing and broad availability not disclosed.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →