WorldLab, a stealth startup founded by Fei-Fei Li, a prominent Stanford AI professor, has raised two funding rounds, two months apart, according to multiple reports. The latest round was led by NEA and valued the company at more than $1 billion, according to people familiar with the investment who spoke to TechCrunch. This follows a $100 million round reported by the Financial Times in July.
That's a significant increase from WorldLab's $200 million valuation in its first round of funding in April, the people said. Reuters reported in May that investors in the first round included Andreessen Horowitz and Radical Ventures, the Canadian firm where Li is a science partner. Li and NEA did not respond to requests for comment.
WorldLab, which was founded in April and reportedly became a unicorn four months after launch, shows that investors continue to bet big on AI startups founded by prominent AI scientists, even if the startups' businesses are unproven.
But in this case, what she's working on is incredibly difficult, and potentially essential in the AI-driven world that Silicon Valley is frantically building: World Labs is working to build AI models that can accurately estimate the three-dimensional physics of real-world objects and environments, allowing it to create detailed digital replicas without the need for extensive data collection.
Widely known as the “Godmother of AI,” Li gave a TED Talk earlier this year in which she discussed how machines can be trained to develop human-like “spatial intelligence.”
“There's very little 3D data in the world,” said one investor familiar with WorldLab's approach. “Self-driving car companies collect that data by driving thousands of miles to create 3D data and then using it to train their machines. For all other applications, like serving coffee, 3D data doesn't exist. Collecting that data is expensive because the range of locations you have to collect it from is huge.”
Once available, World Lab's models could be used in gaming and robotics applications, she said.
Li is best known for her work on ImageNet, the dataset that revolutionized computer vision. She is currently on partial leave from her role as co-director of Stanford University's Institute for Human-Centered AI until December 2025.