美股IPO·2025-09-17 01:18Core Viewpoint - The article highlights the significant advancements in domestic AI chip technology in China, particularly focusing on Alibaba's AI chips showcased in a national news broadcast, and the growing importance of these technologies in supporting the digital economy's high-quality development [1][4][6]. Group 1: Project Overview - A total of 1,747 devices and 22,832 computing power cards have been signed for projects, with an aggregate computing power of 3,479P [3]. - Specific contributions include Alibaba Cloud with 1,024 devices and 16,384 PingTouGe computing power cards, totaling 1,945P of computing power [3]. - Other contributors include the Chinese Academy of Sciences with 512 devices and 4,096 MuXi computing power cards (984P), Beijing Jingyi with 83 devices and 1,328 BiRan computing power cards (450P), and Zhonghao Xinying with 128 devices (200P) [3]. Group 2: Technical Comparisons - A comparison table of key parameters for various computing power cards was provided, including Alibaba's PingTouGe PPU, NVIDIA's A800 and H20, Huawei's Ascend 910B, and BiRan's 104P [3]. - The PingTouGe PPU features HBM2e memory with a capacity of 96GB, inter-chip bandwidth of 700GB/s, and a power consumption of 400W, surpassing the A800 and approaching the H20 [3]. - The BiRan 104P computing power card has 32GB HBM2e memory, inter-chip bandwidth of 256GB/s, and a power consumption of 300W [3]. Group 3: Industry Implications - The gradual implementation of these projects is expected to enhance the role of domestic computing power in key sectors, providing strong support for the high-quality development of China's digital economy [6]. - The public comparison of different brands' computing power card specifications is anticipated to foster healthy competition and technological exchange within the industry, driving continuous upgrades in domestic AI chip technology [6]. - The progress of the China Unicom Sanjiangyuan Green Power Intelligent Computing Center project underscores China's capabilities in the green power computing sector and reflects the robust development of the domestic AI chip industry [7].