On April 27th, reports emerged indicating that during its performance briefing, Moore Threads-U revealed that the development of its next-generation 'Huagang' architecture is proceeding smoothly, with the entire team fully engaged in the endeavor. The specific launch date will be set as promptly as possible, contingent upon the outcomes of the final testing and validation phases. In comparison to the S5000, the 'Huagang' architecture boasts a 50% surge in computational density and an impressive tenfold enhancement in energy efficiency. This innovative architecture supports comprehensive full-precision computing, spanning from FP4 to FP64, and has adeptly integrated AI computing acceleration with physical simulation capabilities onto a single chip. This integration empowers the company's GPU products to adeptly handle the intricate demands of composite scenarios, including large-scale model training, embodied intelligence, and digital twins.