On September 28, AMD made a surprising acquisition that caught the attention of both the chip and AI communities: it acquired the San Francisco startup World Labs through a stock-only transaction, valued at about $8.2 billion, equivalent to approximately 55.1 billion Chinese yuan at the current exchange rate. What's more eye-catching is the person—Li Fei-Fei, co-founder of World Labs and known as the "Mother of AI," will join AMD as Executive Vice President and Chief Scientist, directly reporting to Dr. Lisa Su.

World Labs was founded by Li Fei-Fei and others in 2024. Its main focus is on world models: enabling AI to generate or reconstruct interactive 3D environments based on text, images, and videos. This capability is not just for show; it can be directly used for robot training, factory simulations, and scientific research—meaning it aims to create not just an image, but a digital world that can be entered and manipulated.

Li Fei-Fei is called the "Mother of AI" because she created the large-scale annotated image database ImageNet. This project proved with real evidence that larger data sets significantly improve the performance of computer vision. Later, she became the Chief Scientist of AI and Machine Learning at Google Cloud, and also served as Co-Director of the Human-Centered AI Institute at Stanford University. A person standing at the intersection of academia and industry, now incorporated into AMD's core R&D team, her significance speaks volumes.

Earlier this month, World Labs also unveiled a new model called Atlas, which can predict what a scene would look like from a different camera angle using only a few 2D images. World Labs gave a metaphor: this working method is similar to how large language models predict the next word in a sentence—except while large models guess the next character in a sequence of words, Atlas guesses the next visual perspective in a geometric space.

In the press release, Dr. Lisa Su clearly outlined the reasoning: acquiring World Labs is to allow AMD to better understand the direction of cutting-edge AI models, thereby working backward to develop the chips and computing systems truly needed to run these models. When a company rooted in selling chips brings the creator of world models into the vice president's office, it's clear that what it seeks is not just technology, but navigation for the next generation of AI computing landscape.