Technology 5 min read By Callum Montgomery
AMD to buy World Labs in $8.2bn all-stock deal as Fei-Fei Li joins as chief scientist
AMD is acquiring World Labs, the physical AI startup founded by Fei-Fei Li, in an $8.2 billion all-stock deal. Li will become executive vice president and chief scientist, reporting to CEO Lisa Su, as AMD moves to challenge Nvidia in world models and autonomous machines.
AMD has agreed to acquire World Labs, the physical AI startup founded by Fei-Fei Li, in an all-stock deal valued at $8.2 billion, bringing one of the most prominent figures in artificial intelligence into the chipmaker as it seeks to challenge Nvidia in an emerging branch of the technology.
Li, widely referred to as the «godmother of AI» for her work two decades ago creating ImageNet, the vast visual database that helped make modern deep learning possible, will join AMD as executive vice president and chief scientist. She will report directly to AMD chief executive Lisa Su. The deal unites two of the most senior women in the AI industry and hands AMD a substantial position in world models, a technology that industry figures believe will be central to autonomous machines such as robots and self-driving cars.
World Labs was founded two years ago and had raised roughly $1 billion in funding. Its team will join AMD as part of the transaction, though the companies did not disclose how many employees are involved. AMD said the deal is expected to close by the end of the year, subject to regulatory approvals. Shares in the chipmaker were roughly flat in after-hours trading on Monday.
The acquisition addresses a strategic gap for AMD, which makes the GPU chips used to train and run AI models but has so far competed mainly on the hardware side of the market. World models represent a different approach from the large language models built by OpenAI, Anthropic and Google. Rather than training on text, they learn from data such as video that depicts the physical, three-dimensional world, with the aim of building software that understands how objects move and interact. That capability is seen as a prerequisite for extending AI beyond screens and into humanoid robots and other machines.
Su framed the deal as a way to combine World Labs’ expertise in AI and world models with AMD’s computing leadership. «Together, we’ll combine World Labs’ deep expertise in AI and world models with AMD’s compute leadership to power the future of AI and strengthen the open AI ecosystem,» she said in a post announcing the agreement.
Li said the two companies had already been working together. World Labs and AMD began a «deep technical partnership» last year focused on optimising model training and inference on AMD’s GPU chips. «As our teams worked together, we realized what a natural fit it would be to bring together our AI ecosystem of software and hardware, foundation models and applications,» she said. She described Su as a great friend and an early investor in World Labs.
Li also set out the case for tighter integration of hardware and software in AI development. «Without having a focused hardware effort, AI is hobbled in efficiency. And scale. And for our purposes, remains trapped in the digital world,» she said. In a separate post, she added: «We built the company guided by the foundational belief that language alone isn’t sufficient for model development. The universe isn’t made up of words, it’s made of real things.»
The deal intensifies competition in a field that is attracting significant investment. Nvidia, AMD’s arch-rival, has built its own world models and partnered with startups including Wayve, which develops the technology for self-driving cars. Yann LeCun, Meta’s former chief AI scientist, launched AMI Labs in March, joining a growing list of richly funded ventures betting that world models will become the next frontier of artificial intelligence.
For AMD, the purchase is also a bet that the next wave of AI demand will be shaped by applications that operate in physical environments rather than by chatbots and text-generation tools alone. By owning both the chips and the model-building expertise, the company hopes to design its products for those evolving uses and to compete more effectively with Nvidia, which has moved aggressively into the same territory.
9
