Illustrated globe above the layers of a computer chip
Illustrated globe above the layers of a computer chip
+ AI News

AMD to buy Fei-Fei Li's World Labs for $8.2 billion

AMD agreed to buy World Labs in an all-stock deal. Fei-Fei Li will become its chief scientist as AMD seeks to shape chips and software around new AI workloads.

  • AI
  • Amd
  • World labs
  • Fei fei li
  • World models
  • Ai chips

AMD has agreed to buy Fei-Fei Li’s World Labs in an all-stock deal worth about $8.2 billion. Li will join AMD as executive vice president and chief scientist, reporting to CEO Lisa Su. The companies expect the deal to close by the end of 2026, subject to regulatory approval.

World Labs develops models that generate and work with 3D environments. Its first commercial product, Marble, lets users create interactive worlds from prompts. AMD says the acquisition will help it design AI hardware, software, and systems around the needs of new models and applications. The companies already had a partnership on model training and inference optimization, according to TechCrunch.

Why AMD wants the model team

World models can create simulated environments for applications such as robot training. TechCrunch reports that Nvidia already offers open-weight world models through Cosmos, while AMD’s public models have focused on text and video. Buying World Labs gives AMD a research team working on a different class of AI workload, along with Li in a role that reports directly to Su.

That could give AMD earlier information about the computing demands of spatial AI. It does not, by itself, show how Marble will run on AMD chips or what tools outside developers will receive. The announced plan says the World Labs team will continue its model research. It gives no performance results for the combined effort.

Sources

The AI Feed Desk

The AI Feed Desk

Editorial desk

The AI Feed Desk tracks AI provider updates, model releases, agent tooling, and enterprise adoption, turning fast-moving announcements into source-linked context for builders and operators.

Noticed a typo, incorrect information, or translation error?

Tell us so we can fix it.

Help Improve This Article

Related Articles

Logan stands beneath a red illuminated arch on a street lined with signs in Marvel's Wolverine

Sony brings Quick Spectral Super Resolution AI upscaling to the standard PS5

Sony's QSSR adds AI upscaling to the standard PS5. Its lighter neural network brings a new graphics option, with resolution targets that differ from PS5 Pro.

The AI Feed Desk

By The AI Feed Desk

AstaBrief name in green surrounded by outlined research documents on a dark green background

Ai2 open-sources AstaBrief, the fast report-generation model in Asta

Ai2 releases AstaBrief's weights and training data. Fast mode averages 51.1 seconds per report in Asta, with evaluation limits worth checking before adoption.

The AI Feed Desk

By The AI Feed Desk

Chart comparing training speed for 8, 32, 64 and 128 experts with about 3.2 billion active parameters per token

Ai2 releases Olmo-core 3, open training infrastructure for large MoE models

Ai2's open MoE training stack keeps experts on GPUs and reports higher throughput. Its benchmarks give builders a starting point, with limits on what they prove.

The AI Feed Desk

By The AI Feed Desk

A closed boundary separates a shopping assistant symbol from a checkout terminal; illustration

Amazon blocks Meta's Muse shopping agent from buying on its site

Amazon says Meta's Muse is an unauthorized AI agent, exposing the access and order-recovery rules that shopping agents need beyond product browsing.

The AI Feed Desk

By The AI Feed Desk

Scatter chart comparing decision index scores and median latency for Clef, Clef-flash, Jev and other decision models

Amazon and Cloudflare release open decision models as the Jev clones multiply

AWS releases Strands Decider 2B. Cloudflare adds Clef and Clef-flash, with open weights and hands-on fine-tuning. Task scores show where speed costs accuracy.

The AI Feed Desk

By The AI Feed Desk