Fei-Fei Li, recognized as the "godmother of AI," has successfully closed a $1 billion funding round for World Labs, her startup focused on enabling machines to understand the world in three dimensions. Notable investors include giants like NVIDIA, AMD, Autodesk, and the Andreessen Horowitz fund. Li, like other leaders in artificial intelligence, argues that the future of AI lies in world models rather than pursuing artificial general intelligence (AGI). Who is Fei-Fei Li and why is she relevant? Li was instrumental in the development of ImageNet, the image database that revolutionized deep learning and laid the groundwork for today's generative AI. Her work allowed computers to learn to identify objects in images, driving advancements reflected in voice assistants and generative text and image models. From Stanford University, where she leads the Human-Centered AI Institute, and from World Labs, founded in 2024, Li faces a new challenge: enabling machines to understand the physical world beyond text or flat images. The challenge to solve. Large language models like GPT or Claude have shown tremendous capacity for processing text, but the real world is three-dimensional, with physics and geometry, filled with interacting objects. "For AI to be truly useful, it must understand worlds, not just words," Li stated in her announcement. World Labs focuses on spatial intelligence, developing models capable of perceiving, generating, and interacting with 3D environments. The goal is for AI to reason about space: identifying the position of an object, anticipating movements, understanding spatial relationships, and predicting the consequences of actions in an environment. Advancements and next steps. Last November, World Labs launched Marble, its first commercial product: a model capable of generating editable 3D environments from text, images, videos, or panoramas. Users can create, modify, and export virtual worlds, which is particularly useful in video games, visual effects, virtual reality, and sectors with high demand for 3D content. With the new funding, World Labs will expand its focus to robotics, where spatial intelligence is essential. A robot that understands its environment can plan, adapt, and execute tasks in changing scenarios without constant reprogramming. Autodesk, which has invested $200 million, sees a clear opportunity: its design software is fundamental for architects, engineers, and 3D content creators. The collaboration will integrate World Labs' generative models into Autodesk's workflows, facilitating the creation and manipulation of virtual environments. According to Daron Green, chief scientist at Autodesk, the initial synergy will focus on entertainment and audiovisual production, allowing users to combine objects designed in Autodesk with AI-generated worlds. Competition in the sector. World Labs is not the only company betting on world models. Google DeepMind is developing the Genie model family, while Yann LeCun, former head of AI at Meta, has founded AMI Labs with a similar focus. Other startups like Decart and Odyssey are also exploring this field, though still in early stages. Each player has its own vision: LeCun advocates for the need for a completely new AI architecture to create world models, while Li aims to evolve current generative models and refine them progressively.