According to Fortune, Fei-Fei Li immigrated to the U.S. at 15 with little English or money and ended up running her parents’ dry cleaning business while studying at Princeton. Her creation of ImageNet in 2009—a massive dataset of 15 million labeled images across 22,000 categories—fundamentally changed AI research and led to the deep learning breakthrough of AlexNet in 2012. Today she’s a Stanford professor who co-founded World Labs, a startup that reached a valuation just over $1 billion within four months of launching. She currently advises global leaders including President Biden and serves on the UN’s scientific advisory board while reluctantly accepting her “godmother of AI” title.
From dry cleaner to datasets
Here’s the thing about Fei-Fei Li’s story—it’s not your typical Silicon Valley founder narrative. While other tech leaders were tinkering with computers in their garages, she was literally running a dry cleaning business. And she kept running it remotely during the first half of her PhD at Caltech. That experience, she says, taught her the resilience that would become essential when everyone told her massive image datasets were a waste of time.
The ImageNet breakthrough
Back in 2007, when Li proposed building what would become ImageNet, her mentor warned she’d “taken this idea way too far.” The entire AI field was convinced that algorithms—not data—were the bottleneck. But Li noticed something obvious that everyone else had missed: humans learn from massive amounts of experience, while computers were trying to learn from datasets with just a few hundred images. So she dragged along impatient graduate students and built what nobody thought was possible. The result? The dataset that would accidentally revolutionize AI.
Spatial intelligence revolution
Now Li is tackling what she calls “spatial intelligence” through her startup World Labs. Basically, she wants AI to understand and interact with the physical world visually, the way humans do, rather than just through language. Their first product, Marble, lets users create downloadable 3D worlds through prompts. It’s a fascinating pivot from her earlier work, and she’s been explaining the vision in talks like her TED presentation on spatial intelligence. The billion-dollar valuation after just four months suggests investors are buying into her next big bet.
Reluctant godmother
Li initially cringed at the “godmother of AI” nickname, but she’s come to accept it with a dose of feminist pragmatism. As she pointed out at Fortune’s Most Powerful Women conference, “In the entire history of science and technology, so many men are called founding fathers or godfathers. If women are so readily rejecting that title, where is our voice?” Meanwhile, she’s balancing her CEO role with advising global leaders, having joined the UN’s scientific advisory board while maintaining her position as a Stanford professor. Not bad for someone who started out answering phones at a dry cleaner.
