AMD to Acquire World Labs in $8.2 Billion Deal to Expand Physical AI Push

SAN FRANCISCO, Sept 29: Advanced Micro Devices (AMD) has agreed to acquire artificial intelligence research company World Labs for approximately $8.2 billion in an all stock transaction, expanding the semiconductor company’s ambitions beyond processors and deeper into emerging AI applications.

The agreement announced on September 28 will bring World Labs’ technology and research expertise into AMD as the chipmaker seeks to strengthen its capabilities in physical AI, spatial intelligence and robotics.

World Labs was founded by AI researcher Fei-Fei Li and focuses on what are commonly described as world models. These systems are designed to understand and represent three dimensional environments, allowing AI to reason about physical spaces rather than dealing only with text or conventional two-dimensional information.

The technology has potential applications in robotics, simulation, design and other fields where machines need to understand their surroundings.

AMD said the acquisition will help it develop hardware, software and systems for emerging AI models and applications. The deal reflects a broader change in the semiconductor industry, with chip companies increasingly seeking control over more parts of the AI technology stack.

Traditionally, AMD’s core business has centred on processors and graphics chips. The growth of artificial intelligence has expanded the importance of specialised computing hardware, while competition has increasingly moved into software and model development.

By acquiring World Labs, AMD will gain a team focused on a category of AI that could become important as systems move from digital environments into the physical world.

Physical AI refers broadly to systems that can understand, predict or interact with real-world environments. Robotics is one of its most visible applications.

A robot operating in a warehouse, factory or other environment needs to understand objects, distances, movement and changes in its surroundings. Conventional language models can process instructions, but physical systems require additional capabilities to interpret spatial information and predict how environments may change.

World Labs’ work is focused on this problem.

The company develops models capable of creating and reasoning about interactive three-dimensional environments. Such technology can be useful for training robots and other systems in simulated environments before they operate in the real world.

Simulation can reduce the need to test every scenario with physical machines. Developers can create virtual environments, expose AI systems to different conditions and examine how they respond.

This makes spatial AI relevant not only to robotics but also to industrial design, autonomous systems and scientific research.

AMD has already had a relationship with World Labs. Reuters reported that the startup had previously raised $1 billion in funding, including investment from AMD, and the companies had been working together on model training and inference optimisation using AMD GPUs.

The acquisition therefore builds on an existing technology relationship rather than representing a completely new partnership.

Fei-Fei Li will join AMD as executive vice president and chief scientist once the transaction closes. She will report to AMD Chief Executive Lisa Su, according to Reuters. The deal is expected to close by the end of 2026, subject to the required conditions and approvals.

Li is widely recognised for her contributions to computer vision and AI research. Her involvement gives AMD additional expertise as the company works to expand its AI technology portfolio.

For AMD, the transaction also represents a move toward integrating research talent with its hardware development efforts.

AI performance depends on more than processors alone. Models require software frameworks, training systems, data pipelines and specialised hardware to operate effectively. Companies developing AI chips are therefore increasingly looking for ways to influence the software and model layers that determine how their hardware is used.

The World Labs acquisition fits into this strategy.

AMD’s competition in AI computing has intensified as demand for accelerators and data-centre processors has increased. Nvidia remains a major player in the market, while other technology companies are developing their own chips and AI systems.

The race is now extending into new areas such as robotics and spatial computing.

World Labs’ technology could help AMD prepare for a market in which AI systems are expected to interact more directly with physical environments.

For example, industrial robots could potentially use spatial models to understand factory layouts, identify objects and anticipate movements. Designers could use simulated environments to test products or machines before physical prototypes are built.

Autonomous systems could also benefit from models that provide a richer representation of their surroundings.

The technology is still developing, and the commercial impact of spatial AI will depend on the ability of researchers and companies to turn experimental systems into reliable products.

AMD’s acquisition gives the company direct access to a team already working on these problems.

The agreement also illustrates how AI companies are becoming acquisition targets for major semiconductor manufacturers. Rather than relying entirely on internal research, chip companies can acquire specialised teams with expertise in emerging AI disciplines.

The deal comes after Nvidia’s own major acquisition of AI company Hugging Face, highlighting the increasing importance of software, models and developer ecosystems alongside computing hardware.

AMD has also been expanding its AI capabilities through other acquisitions and partnerships. Its strategy has increasingly involved building a broader platform around AI computing rather than limiting its role to supplying processors.

World Labs adds another component to that strategy.

The startup’s emphasis on spatial intelligence could become particularly relevant as AI development moves toward what many researchers describe as physical or embodied intelligence. These systems are designed not simply to generate text or images but to understand environments and perform actions within them.

Such capabilities require models that can process visual information, spatial relationships and changes over time.

The challenge is substantially different from generating a response to a written question. A physical AI system needs to account for the consequences of its actions and operate within real-world constraints.

That makes computing efficiency important. Training and running advanced spatial models can require significant processing power, particularly when models are used for simulations involving complex environments.

AMD’s existing GPU business provides one part of that infrastructure. The World Labs acquisition could give the company additional expertise in developing models that make use of that hardware.

The deal is also part of a wider shift in the AI industry toward specialised systems. Instead of one general purpose model being expected to handle every task, companies are developing systems optimised for coding, scientific research, cybersecurity, image generation, robotics and other specific applications.

Spatial intelligence represents another area in that expanding ecosystem.

The acquisition will allow AMD to combine its semiconductor capabilities with World Labs’ research in this field. The company has said that the combination can support development across hardware, software and AI systems.

For World Labs, joining AMD provides access to the resources of a large semiconductor company and the opportunity to integrate its technology more closely with computing hardware.

The transaction is expected to be completed by the end of the year, subject to regulatory and other closing conditions.

The $8.2 billion all stock value makes it one of AMD’s largest AI-related transactions and demonstrates the scale of investment now flowing into emerging AI technologies.

As AI development expands from language and image generation toward systems capable of interacting with the physical world, spatial understanding is becoming an increasingly important research area.

AMD’s decision to acquire World Labs places that technology within its broader AI strategy and gives the semiconductor company another route into the development of future robotics and physical AI systems.