国产GPU第二股来了，营收狂飙13倍3年却亏损30亿

Core Viewpoint - Muxi Co., Ltd. is set to be listed on the Shanghai Stock Exchange's Sci-Tech Innovation Board on December 17, 2025, marking its entry as a non-profitable company into the Sci-Tech Growth Tier [1] Group 1: Company Overview - Muxi Co., Ltd. aims to become the second domestic GPU company to go public after Moore Threads, which has already gained significant market attention [1] - The company was founded in 2020 by former AMD employees and focuses on AI chips, particularly in the inference segment [3][4] - Muxi's first inference chip, Xisi N100, was delivered in January 2022 and began mass production in April 2023 [5] Group 2: Market Context - The AI chip market in China is experiencing rapid growth due to U.S. export controls and the need for domestic alternatives, with Muxi entering a competitive landscape dominated by NVIDIA and AMD [4][11] - Muxi's strategy involves developing its own core GPU IP to enhance flexibility and independence in a market where NVIDIA and AMD hold over 70% of the share [11][12] Group 3: Product Development - Muxi has launched several products, including the Xisi N series for inference and the Xiyun C series for training and inference, with plans for further development of the Xiyun C600 and C700 series [9][12] - The company has also developed a high-speed interconnect technology, MetaXLink, to improve performance and compatibility with existing ecosystems [15] Group 4: Financial Performance - Muxi's revenue has shown exponential growth, with reported revenues of 42.64 million yuan in 2022, 530.21 million yuan in 2023, and 743 million yuan in 2024, reflecting a compound annual growth rate of 4074.52% [21] - Despite revenue growth, Muxi has not yet achieved profitability, with cumulative losses exceeding 3 billion yuan over the past three years [23] Group 5: Future Outlook - Muxi anticipates reaching a breakeven point by 2026, driven by increasing sales of its training and inference products [24] - The company is positioned to compete effectively in the AI chip market, but its long-term success will depend on its ability to build a robust technology moat and ecosystem [28]