According to CoinMeta, Zhang Jianzhong, the founder of Moore Threads, stated at the 2026 World Robot Conference that the embodied intelligence industry is facing a "chatgpt moment" similar to that of large models, with computing power being the key to development. Language models have already surpassed the trillion-parameter threshold, and driven by the scaling law, embodied intelligence models are also evolving from billion to tens of billions of parameters, and then to hundreds of billions and trillions of parameters. Consequently, intelligent computing clusters are moving from thousands of cards to tens of thousands of cards, and evolving towards a scale of hundreds of thousands of cards has become an inevitable trend.