World Labs’ AI System Can Generate a 3D World from an Image

World Labs, the AI startup co-founded by AI pioneer Fei-Fei Li, has provided a first look at what its building: an AI system that can create entire interactive 3D scenes from just a single image. Basically, imagine uploading a picture and then being able to step into that world and move around it like a scene from your favorite video game. That’s what World Labs is making possible.

Why it matters: Unlike AI image/video generators that produce static content, World Labs’ system creates persistent 3D environments that maintain consistency as users explore them from different angles. This is a big deal, so let’s unpack it a bit.

Zoom in: Firstly, once a world is generated, it persists—no sudden changes if you look away and then back again. And, beyond just visual fidelity, World Labs’ tech brings the physics of real life into the virtual space. Users can also apply real-time camera effects like depth of field adjustments and dolly zooms, similar to traditional filmmaking techniques. The technology also enables interactive features such as dynamic lighting changes and scene modifications, all rendered directly in a web browser.

The big picture: The application of 3D generative AI extends to video games, movies, architecture, and even virtual education. As co-founder Justin Johnson noted, creating interactive, immersive worlds today takes years of development time and huge budgets. With AI-generated world models, they hope to make this process faster and more accessible, democratizing creativity in the process.

What they’re saying: Early creators who have tested the technology are finding practical applications in their workflows. Digital artist Eric Solorio has been using the system to stage characters within generated environments and orchestrate precise camera movements, demonstrating its potential for visual storytelling and content creation.

The numbers: The technology arrives at a time when creative professionals are seeking more control over AI-generated content. Earlier this year, World Labs secured $230 million in funding from investors including Andreessen Horowitz, Intel Capital, AMD Ventures, Ashton Kutcher, and Eric Schmidt. The company, now valued at over $1 billion, aims to release its first product in 2025.

The bottomline: The company has shared some examples of what we can expect on their site. You can interact with scenes directly in the browser, experiencing effects like depth of field or experimenting with the lighting. There are, of course, limitations: for now, you can only move around within a set area, and the occasional rendering glitch reminds you that this is still a work in progress.

The company is hoping to make the tech generally available in 2025. If you want to follow along or be among the first to try it out, you can sign up at worldlabs.ai/waitlist.