World Labs, the spatial intelligence startup founded by AI pioneer Fei-Fei Li, is joining AMD, the company announced on its blog on Monday. As part of the deal, Fei-Fei Li will become an Executive Vice President and Chief Scientist at AMD, working directly with CEO Dr. Lisa Su, while the World Labs team forms the core of a new frontier research organization inside the chipmaker.
For the latest breaking AI news and analysis, visit AI Buzz Wire.
The move consolidates one of the most closely watched independent AI research labs into the second-largest player in AI accelerators, and it signals AMD's intent to compete not just on GPUs but on the next frontier of AI: models that understand and reason about the physical, three-dimensional world.
From Partnership to Combination
The acquisition is the culmination of a relationship that began as a deep technical partnership last year, when World Labs started working with AMD on model training and inference optimization on AMD GPUs. According to the announcement, as the teams worked together it became clear that combining World Labs' AI ecosystem of software, foundation models and applications with AMD's hardware would be "a natural fit."
"To accelerate into this future requires scaling our efforts, scaling our reach, and getting closer to the hardware," the company wrote in the announcement, framing the deal as a way to build "an end-to-end open AI ecosystem spanning hardware, software, platform" and models.
The transaction is expected to close by the end of 2026, subject to regulatory approvals and other customary closing conditions.
What World Labs Brings to AMD
Founded in 2024, World Labs has been the flagship startup of the spatial intelligence movement, building world models that can perceive, generate and reason about 3D environments. The company's research team is led day-to-day by Justin Johnson and Ben Mildenhall, who will continue running the World Labs team under Fei-Fei Li as it joins AMD.
The startup has drawn attention for its Atlas world model family, which targets simulation, robotics and creative applications that require a grounded understanding of physical space. Fei-Fei Li, best known as the co-creator of the ImageNet dataset that catalyzed the deep learning era, has argued that spatial intelligence is the natural next step beyond language models.
Why AMD Is Buying Into World Models
For AMD, the deal is a statement of strategic ambition. Nvidia's dominance in AI computing rests not only on its chips but on the tightly integrated stack of software, models and applications built around them. By acquiring a frontier research organization focused on world models, AMD gains both the talent and the model portfolio to argue that the future of AI hardware decisions will be made at the level of complete ecosystems, not individual GPUs.
An open ecosystem is the explicit goal. World Labs' announcement emphasizes an "end-to-end open AI ecosystem," a positioning that contrasts with the more closed approach of its larger rival and that could appeal to enterprises and governments wary of single-vendor lock-in.
The deal also lands amid a broader industry debate over open versus closed AI development, with open-weight models from Chinese and Western labs gaining ground on proprietary systems. AMD, which has leaned into supporting open model ecosystems on its Instinct accelerators, is doubling down on that positioning at the research frontier.
A Fast Rise From Lab to Acquisition
The pace of the journey is striking. World Labs was founded only in 2024, and in two years it has gone from stealth research effort to a company whose breakthroughs its own announcement describes as having "given us a clear vision for AI's potential to solve problems in the spatial and physical world." The team's work on large world models, tools for generating and interacting with 3D environments, and simulation research made it one of the most closely watched independent labs in the field.
Fei-Fei Li's own standing made the company a bellwether. As the co-creator of ImageNet, the dataset and benchmark effort widely credited with igniting the modern deep learning boom, she has repeatedly shaped which research directions the industry treats as central. Her argument that spatial intelligence, not just language, is the next frontier has drawn both adherents and skeptics; AMD is now betting decisively on the adherents' side.
Open Questions
Several practical questions remain. The companies have not disclosed the financial terms of the transaction, nor what will happen to World Labs' existing commercial products and API customers once the deal closes. Regulatory review could also shape the final structure, given the deal's expected close at the end of 2026 is explicitly subject to approvals. And it remains to be seen how much independence the World Labs research organization will retain inside a public chipmaker with its own product cadence.
What It Means for the Spatial AI Race
The absorption of World Labs marks a turning point for the spatial intelligence field, which until now has been defined by independent players. With the field's most prominent founder now inside a major semiconductor company, competitors will face a rival with direct control over the hardware-software co-design that world models demand.
For Fei-Fei Li, the move returns her to a large-platform vantage point she last held at Google, but this time with her own research organization, her own leadership team, and a chipmaker's resources behind her thesis that the next generation of AI will need to understand the physical world, not just describe it.
Stay Ahead of AI
For continuous coverage of the AI industry's most important developments, bookmark AI Buzz Wire and never miss a breaking story.
Read more AI news
