Cambricon(SH:688256) Hua Xia Shi Bao·2025-07-02 06:55Core Viewpoint - The domestic GPU sector is experiencing a significant moment with the listing of AI chip companies, particularly with the IPO attempts of Moore Threads and Muxi Integration on the STAR Market, amid rising AI computing demand and U.S. chip restrictions [1][2]. Company Overview - Moore Threads is referred to as the "Chinese version of NVIDIA" due to its similar business focus on GPU development, design, and sales, including AI computing products and graphics acceleration [2]. - The company was founded during a time of increasing demand for semiconductor localization in China, coinciding with the establishment of the STAR Market's registration system, which has facilitated the listing of semiconductor companies [2]. Financial Performance - Moore Threads plans to raise 8 billion yuan through its IPO, but it has reported cumulative losses exceeding 5 billion yuan from 2022 to 2024 [1][7]. - The company's revenue projections for 2022, 2023, and 2024 are 46.09 million yuan, 124 million yuan, and 438 million yuan, respectively, while net losses are projected at -1.84 billion yuan, -1.67 billion yuan, and -1.49 billion yuan for the same years [7]. - Research and development expenses are significantly high, with R&D costs reaching 1.116 billion yuan, 1.334 billion yuan, and 1.359 billion yuan from 2022 to 2024, representing an R&D-to-revenue ratio of 2422.51%, 1076.31%, and 309.88% respectively [7]. Market Context - The AI chip market has seen a cooling of investment enthusiasm, with a shift towards rational investment focusing on product commercialization capabilities [4][8]. - The global GPU market is projected to reach 3.611974 trillion yuan by 2029, with China's GPU market expected to grow to 1.363578 trillion yuan, increasing its global market share from 30.8% in 2024 to 37.8% in 2029 [8]. Competitive Landscape - Moore Threads faces significant competition from established players like NVIDIA and other domestic GPU companies, particularly in training scenarios where their products currently cannot compete with NVIDIA or Huawei [8]. - The company is working on building its own ecosystem, MUSA, which aims to be compatible with NVIDIA's CUDA ecosystem, although it acknowledges existing gaps in technology and market presence compared to NVIDIA [6][8].