World Labs raises $1B with $200M from Autodesk for 3D AI models
Summary
World Labs raised $1B, including $200M from Autodesk. They'll partner to integrate AI models for 3D environments with Autodesk's tools, aiming to boost creative design, starting with entertainment.
Autodesk joins one billion dollar round
World Labs secured $200 million from software giant Autodesk as part of a larger $1 billion funding round. The startup, founded by AI pioneer Fei-Fei Li, focuses on "spatial intelligence" and the creation of physical AI systems. Other major backers in this round include Nvidia, AMD, Fidelity, and Emerson Collective.
World Labs emerged from stealth in 2024 with an initial $230 million and a $1 billion valuation. While the company has not confirmed its new post-money valuation, recent industry reports suggest the startup aimed for a $5 billion figure during this latest capital raise. The massive influx of cash positions the company to compete directly with established AI labs like OpenAI and Google DeepMind.
The deal establishes a strategic partnership between World Labs and Autodesk to explore how generative 3D environments can integrate with professional design tools. The companies will focus first on media and entertainment use cases before expanding into other sectors. This collaboration signals a shift in the AI industry from generating 2D images and text to building interactive, three-dimensional worlds.
Spatial AI targets the built world
The partnership centers on "world models," which are AI systems capable of generating and reasoning about immersive 3D environments. Unlike Large Language Models (LLMs) that predict the next word in a sequence, these models understand geometry, physics, and dynamics. World Labs intends to build systems that can simulate the physical world with high fidelity.
Autodesk currently dominates the market for 3D computer-aided design (CAD) software used in architecture, engineering, and manufacturing. By investing in World Labs, Autodesk aims to automate complex design workflows that currently require manual labor. The two companies will collaborate at the research level to ensure their respective models can "consume" each other's data effectively.
The current investor lineup reflects a broad interest in the hardware and software requirements of spatial AI. Notable participants in the $1 billion round include:
- Nvidia: Providing the compute power necessary for 3D simulation.
- AMD: Supporting the hardware infrastructure for large-scale world models.
- Autodesk: Offering a direct path to commercial design and engineering markets.
- Fidelity: Providing late-stage venture capital for scaling operations.
- Emerson Collective: Investing in the long-term societal impact of physical AI.
Entertainment leads the market strategy
World Labs and Autodesk plan to prioritize media and entertainment as their first commercial frontier. This strategy aligns with other major players like Runway and Google, who view gaming and film production as the most immediate markets for world models. Autodesk already maintains deep relationships with major production houses and provides tools for character animation and visual effects.
Daron Green, Autodesk’s chief scientist, noted that the partnership is in its early stages but will involve deep technical collaboration. Users might eventually start a project in World Labs by prompting a 3D sketch of an office layout. They could then move that sketch into Autodesk software to refine specific architectural details, such as desk designs or structural supports.
The workflow also functions in reverse, allowing users to take an object designed in Autodesk and place it into a context created by World Labs. This interoperability allows for rapid prototyping in virtual environments that behave like the real world. Despite the close collaboration, Green confirmed that the current agreement does not include a formal data-sharing component.
Neural CAD meets world models
Autodesk is simultaneously developing its own internal AI capabilities, which it calls "neural CAD." This technology uses generative AI trained on geometric data to reason about entire mechanical systems. Neural CAD moves beyond simple image generation to create functional 3D models that understand how different components interact.
The integration of World Labs’ technology could extend these capabilities from individual objects to entire digital representations of the physical world. World Labs’ first product, Marble, already allows users to create and download editable 3D environments. Released in November 2024, Marble serves as a proof of concept for the startup’s commercial ambitions.
Fei-Fei Li, often called the "Godmother of AI" for her work on ImageNet, argues that AI must understand "worlds, not just words" to be truly useful. Her vision for World Labs involves reconciling semantic understanding with spatial and physical reality. This approach differentiates the company from competitors that focus primarily on generative video or text-to-image synthesis.
The competitive landscape for simulation
World Labs enters a crowded field where tech giants are racing to master 3D simulation. Google DeepMind has teased its own world models for gaming, while Runway recently launched Gen-3 Alpha to improve temporal consistency in AI video. The primary challenge for these companies is ensuring that generated environments follow the laws of physics without glitching or "hallucinating" impossible structures.
Autodesk’s involvement gives World Labs a significant advantage in technical validation. Because Autodesk software is used to build real skyscrapers and airplanes, its endorsement suggests that World Labs' models have the potential for high precision. Most generative AI tools currently lack the geometric accuracy required for professional engineering and construction.
The startup will use the $1 billion in new capital to hire more researchers and secure the massive amounts of GPU compute needed to train spatial models. As the company scales, it faces the task of proving that its "world models" can offer more than just aesthetic improvements for gamers. The ultimate goal is a system that can assist in the design and management of the physical infrastructure that defines modern life.
Related Articles

Snyk CEO Peter McKay steps down, seeks AI-focused successor
Snyk CEO Peter McKay steps down, saying the company needs an AI-focused leader for its next phase. He'll stay until a successor is found.

Reliance Jio to invest $110 billion in AI datacenters over seven years
Reliance Jio plans to invest $110 billion in AI datacenters over seven years, aiming to make AI services as affordable as it made mobile data in India.
Stay in the loop
Get the best AI-curated news delivered to your inbox. No spam, unsubscribe anytime.

