As reported by Zhidx, Madison, the daughter of NVIDIA's founder Jensen Huang and the leader of the Physical AI Business division, recently made a public appearance. During a conversation with Xie Chen, the CEO of Lightwheel Intelligence, she shared her profound perspectives on robotics and AI technologies.

Madison pointed out that the robotics sector is currently grappling with a data gap problem. While researchers working on large language models have access to a wealth of internet data, the acquisition of robotics data predominantly relies on expensive and inefficient manual collection methods. To address this challenge, simulation and synthetic data play a pivotal role. She emphasized that NVIDIA requires mechanisms akin to synthetic data factories, a need that aligns perfectly with Lightwheel Intelligence's business model.

Madison further underscored that, in the future, the vast majority of robotics training data will be derived from synthetic sources. The current bottleneck in the robotics field lies in the simulation of wires and cables. In this regard, NVIDIA has already teamed up with Lightwheel Intelligence to achieve high-precision simulation of the physical properties of cables.

Moreover, Madison shared her vision for the future of manufacturing. She believes that all equipment leaving the factory should be accompanied by a digital twin. Physical factories will operate in tandem with AI factories, with digital twins serving as the cornerstone for their efficient and autonomous functioning.