On August 21, during the 2026 World Robot Conference, Zhang Jianzhong, the visionary founder of Moore Threads, emphasized that the embodied artificial intelligence (AI) sector is experiencing a pivotal 'ChatGPT moment,' akin to the transformative phase witnessed by large language models. He highlighted that the computing power infrastructure has emerged as a cornerstone for industry advancement. With language model parameters soaring past the trillion mark, propelled by the principles of scaling laws, embodied AI models are making significant strides, evolving from models with billions and hundreds of billions of parameters to those boasting trillion-parameter scales. As a result, the scale of intelligent computing clusters is set to grow exponentially, expanding from clusters of thousands of cards to tens of thousands, and ultimately progressing towards the monumental scale of 100,000 cards.