Abstract: The construction of super-large-scale computing power is of great significance for enhancing China's technological innovation capacity, scenario-based application capacity, and future industrial competitiveness in the embodied intelligence industry. Its role is mainly reflected in strengthening model training and simulation verification capabilities, promoting industrial-chain collaboration and scenario data feedback, supporting edge inference and human-machine collaboration, and driving embodied intelligence from technological research and development toward application in real-world scenarios. At present, the empowerment of the embodied intelligence industry through super-large-scale computing power still faces practical constraints, including insufficient adaptation between computing power supply and demand, the need to further strengthen underlying technological support, inadequate depth of integration between computing power and application scenarios, and insufficient supply of governance rules. Looking ahead, efforts should be coordinated in core technology research, spatial layout of computing power, integration between computing power and application scenarios, and construction of a modern governance system, so as to build a computing power support system that is independently controllable, efficiently scheduled, scenario-driven, safe, and green, thereby promoting the continuous transformation of computing power resources into development capabilities for the embodied intelligence industry.