Li Feifei in Dialogue with Su Zifeng: Robots Will Become the "New Engine of Computing Power"
At the CoRL 2026 Conference on Robot Learning, World Labs founder Li Feifei held a dialogue with AMD Chair and CEO Su Zifeng. Li Feifei judged that robots will drive computing power into its next stage, and Su Zifeng predicted that in the coming years the market size of physical AI, especially robots, may reach hundreds of billions of dollars. AMD has already formed a robotics research team, but its layout is still at an early stage. The two also called for providing more computing resources to the public sector, including universities and nonprofit institutions.
Body · AI translation
Li Feifei in Dialogue with Su Zifeng: Robots Will Become the "New Engine of Computing Power"
Phoenix Tech Download the App for Exclusive Early Access
November 12, 2026 14:28:50 From Beijing
Updated November 12, 2026 14:38:34
Phoenix Tech News (Editor/Song Shumei) On November 12, at the CoRL 2026 Conference on Robot Learning, World Labs founder and CEO Li Feifei and AMD Chair and CEO Su Zifeng held a dialogue. The two discussed topics including robot development, physical AI, computing power demand, and model roadmaps. Li Feifei judged that robots will drive computing capability into its next stage of development; Su Zifeng predicted that in the coming years the market size of physical AI, especially robots, may reach hundreds of billions of dollars.
Su Zifeng said that robots need to understand physical laws, models, and the surrounding environment, which requires coordinated optimization of hardware, software, and systems. AMD hopes to provide a basis for the research and development of future chips and computing systems by gaining an in-depth understanding of model and robot workloads. At present, AMD has already formed a robotics research team, but the related arrangements are still at an early stage.
Regarding the future technology roadmap for robots, Li Feifei believes that robots driven by vision-language models and large language models will play an increasingly important role, but system-level challenges have not yet been fully solved. Su Zifeng also pointed out that there is no single model that can cover all applications; some scenarios are suitable for cutting-edge large models, while other scenarios may require more targeted specialized models.
Speaking of computing power demand, Li Feifei said that the incremental computing brought by robots will come not only from model training but also from edge devices and real-world environments. She believes that the relationship between AI and computing has undergone a fundamental change: "AI is computing, and computing is AI," and models and hardware need to evolve together.
The two also called for providing more computing resources to the public sector, including universities and non-profit institutions. Li Feifei believes that public research institutions not only undertake the tasks of basic innovation and talent cultivation but also need to participate in AI evaluation, benchmarking, and safety research. Su Zifeng responded that allocating part of computing power to long-term research and opening it to universities is a responsibility that industry enterprises should bear.
In September this year, AMD announced a plan to acquire World Labs in an all-stock transaction valued at approximately $8.2 billion, with the deal expected to be completed before the end of 2026. After the closing, Li Feifei will serve as AMD Executive Vice President and Chief Scientist, reporting to Su Zifeng.