AMD to acquire World Labs for 8.2 billion dollars

Semiconductor manufacturer AMD has announced an agreement to acquire World Labs, the spatial intelligence startup founded by AI pioneer Fei-Fei Li, for 8.2 billion dollars.

2 min read

AMD to acquire World Labs for 8.2 billion dollars

AMD has reached an agreement to acquire World Labs, a prominent developer of deep learning models designed to understand physical reality, in an 8.2 billion dollar transaction. The deal brings together advanced spatial intelligence research and silicon engineering, with the acquisition expected to close before the end of the year pending regulatory approval.

Key facts

Acquirer
AMD
Acquired
World Labs
Acquisition Price
$8.2 billion
Founded
Beginning of 2024
Co-founders
Ben Mildenhall and Justin Johnson
New Role
Executive Vice President and Chief Scientist

Integration and leadership

Following the completion of the deal, World Labs founder Fei-Fei Li will join AMD as executive vice president and chief scientist. Li, a Stanford computer science professor widely recognized as a pioneer in computer vision for her work on ImageNet, founded World Labs in 2024 to develop models that possess a robust understanding of the physical world.

Fei-Fei Li speaking onstage holding a microphone.
World Labs founder Fei-Fei Li will join AMD as executive vice president and chief scientist following the acquisition., Source ↗

In a public statement regarding the acquisition, Li emphasized the necessity of hardware proximity for scaling technological breakthroughs beyond the laboratory setting. World Labs previously established an inference optimization and training partnership with AMD, laying the groundwork for closer organizational ties.

Strategic context for hardware

For AMD, the acquisition provides critical insights into frontier workloads that will directly influence its future chip-making roadmap. The company faces stiff competition from rival Nvidia in establishing a comprehensive ecosystem for specialized AI hardware.

World Labs focuses on creating world models, which are increasingly viewed as essential for deploying generative AI onto robotic platforms such as autonomous vehicles and industrial humanoids. Because high-quality real-world data remains scarce for general-purpose robot training, synthetic data generated by advanced world models is considered vital for overcoming current data bottlenecks.

AMD leadership noted that understanding these advanced workloads will help shape competitive silicon offerings, positioning the combined entity to better serve developers working at the intersection of compute infrastructure and physical reasoning.

Source: Original announcement

Worth a read?

Report

Comments