AMD Buys Fei-Fei Li's World Labs for $8.2 Billion

AMD will acquire World Labs, the deep learning startup founded by Stanford professor Fei-Fei Li, for $8.2 billion, the two companies said today. The deal hands AMD one of the leading developers of AI models built to understand physical reality — and a direct answer to rival Nvidia's world-model portfolio.

The strategic logic runs on both sides. World Labs argued in a statement that AI development requires "close collaboration across model reseach, systems and compute." AMD says understanding frontier workloads like those created at World Labs will shape its chip-making roadmap.

Fei-Fei Li joins AMD as executive vice president and chief scientist. The relationship predates the acquisition: the two companies formed an inference optimization and training partnership last year, and Li appeared as a guest at AMD's CES presentation earlier this year.

Li is considered a pioneer of AI, particularly computer vision, for her role creating the ImageNet database and the challenges that followed. She founded World Labs in 2024 to build deep learning models with a more robust understanding of reality. Her argument: true general intelligence requires a grounding in physics and the ability to reason about data beyond text.

In a post announcing the deal, Li framed the acquisition as the result of a desire to scale World Labs' technical breakthroughs beyond the lab. "Now that we have tangible proof of the possibilities, we want to do everything we can to accelerate the future," Li wrote. "To do this requires scaling our efforts, widening our reach, and getting closer to the hardware."

"World model" remains a loose term. It covers everything from language models trained to understand visual inputs to models capable of generating and sustaining high-fidelity simulations of reality. World Labs' first product, Marble, targets entertainment experiences — but also the creation of simulated environments for training robots.

That robotics angle matters for AMD's competitive position. The acquisition is likely to help the company compete with long-standing rival Nvidia in building an ecosystem for AI-specific chips. Nvidia already offers a suite of open-weight world models, such as Cosmos. AMD, by contrast, has only offered text- and video-based models to the public.

World models carry weight in the effort to deploy generative AI on robotic platforms, from autonomous vehicles to industrial robots and general-purpose humanoids. Useful data to train general-purpose robots remains scarce, and synthetic data from world models is seen as key to realizing the visions put forward by companies like Tesla and Figure.

The acquisition is expected to close before the end of the year, subject to regulatory approval. For AMD, the bet is that owning both the models and the silicon beneath them closes a gap Nvidia opened years ago.