Zhongcheng Hualong Launches HL200 Inference Chip and 10,240-Card AI Computing Cluster with 5.12 TFLOPS/W Energy Efficiency

Zhongcheng Hualong has officially launched its HL200 inference chip and supernode AI computing cluster solution. The HL200 delivers 4 PFLOPS of FP4 performance per card, 5.12 TFLOPS/W energy efficiency, and supports horizontal scaling up to 10,240 cards. As domestic AI chips move from single-card benchmarking toward cluster and ecosystem-level systemization, energy efficiency and elastic networking for inference compute are becoming key competitive factors for AI computing platforms...