AIgiants such as Google DeepMind, Meta, and NVIDIA are shifting their R&D focus toward ‘world models’ in an effort to gain a competitive edge in the race toward machine ‘superintelligence.’ World models learn from video and robotics data to comprehend the physical world, offering broad application potential. According to a senior executive at NVIDIA, the potential market size could reach up to $100 trillion, encompassing sectors like autonomous driving, robotics, and manufacturing.