AMD to acquire World Labs, led by Fei-Fei Li, for $8.2 billion
The deal will bring World Labs' world-model technology and its founder into AMD, aiming to sharpen the chipmaker’s AI roadmap and compete with Nvidia.
1 source · cross-referenced
- AMD agreed to acquire World Labs, a developer of deep learning models for understanding physical reality, for $8.2 billion.
- World Labs founder Fei-Fei Li will join AMD as executive vice president and chief scientist upon closing.
- The acquisition is expected to close before the end of 2026, pending regulatory approval.
- AMD and World Labs previously partnered on inference optimization and training workflows.
AMD announced it will acquire World Labs, a developer of deep learning models intended to understand physical reality, in an $8.2 billion all-cash transaction. World Labs stated that advancing AI requires "close collaboration across model research, systems and compute," while AMD said frontier workloads like those at World Labs will inform its chip roadmap. Upon closing, World Labs founder Fei-Fei Li will join AMD as executive vice president and chief scientist.
AMD and World Labs formed a partnership last year focused on inference optimization and training workflows, and ties have remained close. Li, a Stanford computer science professor and AI pioneer known for creating the ImageNet database and its associated competitions, founded World Labs in 2024 to develop deep learning models with stronger grounding in physics and the ability to reason beyond text.
World Labs describes its first product, Marble, as a tool for creating entertainment experiences and simulated environments for robot training. The company positions Marble as a "world model," a term that broadly encompasses models trained to understand visual inputs or capable of generating high-fidelity simulations of reality.
The acquisition is expected to close before the end of 2026, subject to regulatory approval. AMD aims to use the deal to bolster its position against Nvidia in the AI chip ecosystem, particularly in areas like robotics where synthetic data from world models is seen as critical for training general-purpose robots.
Nvidia already offers open-weight world models such as Cosmos, while AMD has publicly released only text- and video-based models. Industry observers view world models as vital for deploying generative AI on robotic platforms, including autonomous vehicles, industrial robots, and humanoid robots.
- Oct 14, 2026 · TechCrunch — AI
Scimplify raises $25M to expand AI-driven specialty chemicals and critical materials production
Trust79 - Oct 13, 2026 · TechCrunch — AI
Character.AI files for Chapter 11 bankruptcy, Disney to acquire assets for $20 million
Trust79 - Oct 13, 2026 · The Verge — AI
Character.AI files for Chapter 11 bankruptcy and sells assets to Disney
Trust79