AMD will acquire World Labs, the San Francisco AI lab founded by Stanford professor Fei-Fei Li, in an all-stock deal valued at $8.2 billion, the chipmaker said on Monday. The acquisition pulls Li and a research team focused on spatial intelligence and physical AI into a company racing to design hardware for the next generation of AI applications, with Li set to join AMD as executive vice president and chief scientist once the transaction closes. The deal is expected to close by the end of 2026, subject to regulatory approvals and other customary closing conditions.
- Deal value
- $8.2 billion all-stock transaction
- Acquirer
- AMD (NASDAQ: AMD), headquartered at 2485 Augustine Drive, Santa Clara, California
- Target
- World Labs, a San Francisco based AI lab founded in 2024
- Founder role
- Fei-Fei Li joins as executive vice president and chief scientist, reporting to AMD CEO Lisa Su
- Expected close
- By the end of 2026, subject to regulatory approval and customary closing conditions
- World Labs focus
- Spatial intelligence models that generate, reconstruct and simulate interactive 3D environments from text, image and video, plus technology for robotic learning and simulation
- Stock reaction
- AMD up 0.4% in after-hours trading on September 28, 2026 after a 3.6% drop in the regular session, according to MarketWatch
A $8.2 billion bet on spatial intelligence
AMD is paying $8.2 billion in stock for a lab most people outside AI research had not heard of two years ago. World Labs, founded in 2024 by Fei-Fei Li alongside Justin Johnson, Christoph Lassner and Ben Mildenhall, develops spatial-intelligence models that can generate, reconstruct and simulate interactive 3D environments from text, image and video inputs, and it also builds technology for robotic learning and simulation. The lab's work is aimed at AI that understands the physical world, a problem AMD says is reshaping what kinds of compute hardware customers need. According to MarketWatch, AMD's stock was up 0.4% in after-hours trading Monday after ending the regular session down 3.6%, a sign that Wall Street saw the price as full but the strategic logic as defensible. The all-stock structure means the final cost to AMD will rise or fall with its own share price before the deal closes, as laid out in the AMD press release announcing the acquisition.
Fei-Fei Li moves from Stanford to AMD
Li, who is often called the godmother of AI for her work on the ImageNet dataset and her long career at Stanford, will join AMD as executive vice president and chief scientist reporting to AMD chair and CEO Lisa Su. In a statement, Li said joining AMD would give the team 'the resources and engineering depth to accelerate our research and help define the infrastructure needed for the next era of AI.' Su, in turn, called Li 'exceptional research leadership and model expertise' and said the acquisition would help AMD 'develop the hardware, software and systems that will power the next generation of AI.' For Li, the move is unusual. She has spent her career in academia, first at Princeton, then at Stanford, where she directs the Stanford Human-Centered AI Institute, and she remains one of the most cited researchers in the field. Joining a chipmaker is a bet that the next wave of AI research will be tightly coupled with the silicon it runs on, and that sitting inside AMD is the best place to shape that.
AMD's pitch for open AI compute
AMD framed the deal as a move to strengthen what it calls the open AI ecosystem, meaning AI models whose code, training data and other information can be modified and used by any developer. The chipmaker said demand for computing infrastructure to support reasoning, robotics, simulation and physical AI is becoming 'more diverse,' and that the World Labs acquisition would give it more insight into how those workloads are evolving so it can design future chips around them. The pitch is also a competitive one. Nvidia still dominates the market for AI training and inference silicon, and open model efforts from Meta, Mistral and a long list of Chinese labs are widely seen as a way to challenge that position. Buying a frontier research lab is a way for AMD to deepen the technical ties between its hardware roadmap and the open model community that depends on it, and to give customers a credible reason to standardise on AMD silicon for the next round of physical AI work.
What the World Labs team brings to AMD
For AI researchers and engineers, the most concrete consequence of the deal is where they will work. The World Labs team will continue to focus on advancing AI model research inside AMD once the deal closes, with Li reporting directly to Su. The lab's work on spatial intelligence and physical AI also signals where AMD expects future AI demand to come from: robotics, simulation and embodied agents that have to reason about the physical world. By owning a frontier research team rather than just partnering with one, AMD gets an inside view of which model architectures and workloads will matter next, and a recruiting base for the AI engineers it will need to compete with Nvidia and the hyperscalers. The deal does not disclose how many World Labs employees will join AMD, but at an $8.2 billion valuation the headcount is almost certainly in the dozens of researchers and engineers, with a similar number of product and operations staff.
Primary source
Check the original source
AMD press release: AMD to Acquire World Labs to Advance the Future of AI Compute is the source to consult for the underlying data, statement, ruling or live context.