Who is Prof Fei-Fei Li, the ‘Godmother of AI’ whose startup is valued at over $1 billion?

Li was born in China and migrated to the United States when she was just 15. The Stanford professor came from a humble background. As a teenager, she helped her parents run a dry-cleaning business in Parsippany, New Jersey, to make ends meet. She worked at the shop for seven years, continuing through the middle of her graduate studies, until she turned 18.

“We were not financially well off at all. My parents were doing cashier jobs, and I was working in Chinese restaurants,” she told Bloomberg in a Q&A. “My family and I decided to run a little dry-cleaner shop to make some money to survive.”

Li attended Princeton University for her undergraduate studies and later pursued a Ph.D. at Caltech in California. She has also worked at Google, where she served as Chief of AI at Google Cloud until October 2018, when she left following leaked emails related to the controversial Project Maven. She is currently the co-director of Stanford’s Human-Centered AI Institute, which aims to advance AI research, education, policy, and practice to benefit humanity, and she is also the co-founder and CEO of World Labs.

It is her pioneering work on ImageNet, a project started in 2006 that classified millions of digital images and became a training ground for modern AI vision systems, for which she is widely recognised. In fact, it was the ImageNet project that kickstarted the deep learning boom.

“Pre-ImageNet, people did not believe in data,” Li said in an interview at the Computer History Museum in 2024. “Everyone was working on completely different paradigms in AI with a tiny bit of data.”

At World Labs, Li is working on building world models – basically AI systems that leverage spatial intelligence, which Li describes as “the ability for AI to understand, perceive, reason, and interact with the world. It comes from a continuation of visual intelligence.”

Her memoir/technology book, The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI, was released in 2023.

Earlier this month, Li’s startup launched its first commercial world model product, called Marble, which is available via freemium and paid tiers. The Marble model turns text prompts, photos, videos, 3D layouts, or panoramas into editable, downloadable 3D environments, a major step toward what the AI leader calls “spatial intelligence.”

Li sees “spatial intelligence” as the next chapter in artificial intelligence, building on the leap made by large language models (LLMs). She believes it will drive creativity, spatially-aware robotics, and scientific research. While LLMs, trained on language, are powerful, they are limited. True human-like intelligence also requires an understanding of physical space, which can be achieved through “world models” that simulate consistent 3D environments.

Li isn’t the only one backing the idea of so-called “world models.” Many world-renowned AI researchers believe that world models will propel the AI revolution into its next phase, including Meta’s former Chief AI Scientist Yann LeCun, who recently announced he would step down to launch his own world model startup.