炒股就看金麒麟分析师研报，权威，专业，及时，全面，助您挖掘潜力主题机会！ 作者丨吴佳楠 编辑丨朱益民 国产GPU公司迎来IPO融资热潮。 1月2日，壁仞科技（6082.HK）上市首日盘中一度涨近120%，成为"港股GPU第一股"。 上个月，号称"国产GPU第一股"的摩尔线程（688795.SH）上市首日股价涨超4倍，12月11日收盘股价超 900元/股，远超发行价114.28元/股。 激情燃烧的高估值迎来"折扣"，背后是GPU创新叙事的商业化现实难题，还需突破。 另一家GPU公司——沐曦股份（688802.SH）2025年12月17日上市首日，投资者中一签浮盈最高逼近40 万元，刷新近十年A股上市首日单签盈利纪录。 然而，资本狂追热捧的背后却是故事另一面，GPU公司股票市值大幅回落。 摩尔线程、沐曦股份上市之日狂飙后，目前已经大幅回调近40%，分别自最高价回落约37%、35%。壁 仞科技上市首日最终收涨不超过80%，总市值不到千亿港元，只有摩尔、沐曦市值的三分之一左右。 短短一月内，"国产GPU四小龙"其中的摩尔线程、沐曦股份、壁仞科技都先后走向了公开市场融资，并 在资本市场创下数项新的纪录。 2025年12 ...

2025 年的冬天，南京的资本市场比寒潮中的气温更炽热。 12 月 17 日，国产 GPU 领军企业沐曦股份在科创板鸣锣，首日股价最高涨幅突破 850% ，总市 值飙升至 3300 亿元，创下 A 股注册制以来新股涨幅纪录。 这并非孤例。 就在同一个月， 12 月 5 日，同为南京投资布局的"国产 GPU 第一股"摩尔线程上市，首日涨幅 425% ，市值轻松突破 3000 亿元；回溯至 2020 年，寒武纪在南京 8 亿元战略投资的"雪中送 炭"下，不仅渡过客户流失危机，更成长为市值超 5600 亿元的 AI 芯片巨头。 看向隔壁安徽省。长期以来，合肥凭借京东方、长鑫存储、蔚来等押注式投资，稳居"最牛风投城 市"宝座。 图源"南京新闻" 现如今，短短一个月内，三大芯片巨头接连上演"上市即暴涨"的戏码，且都与南京资本深度绑定。 这不禁让人好奇： 南京什么时候悄悄攒下了这么强的投资功力？这种"耐心打法"，会不会改写城市 风投的游戏规则？ 芯片三巨头的"南京基因" 2020 年 9 月，当陈维良带着几位 AMD 前同事创立沐曦股份时，国产 GPU 赛道还是一片"无人 区"。 彼时，全球 GPU 市场被英伟达、 AM ...

郑纬民认为，算力自主有三个要求：一是芯片设计能力要自主；二是制造与供应链风险要可控；三是系 统与集群交付能力要强。算法自强则指的是，针对GPU能做大模型适配，能做大模型训练，能让模型优 化。 中新网12月22日电(记者 吴家驹)"真正决定主权AI成败的，在于是否有足够多的开发者，愿意长期在这 套栈上写代码。"近日，摩尔线程首届MUSA开发者大会上，中国工程院院士、清华大学计算机系教授 郑纬民如是表示。 郑纬民表示，过去很长一段时间，做芯片是全球化分工的，有人做架构设计，有人做设备制造，有人做 代工，有人做封装。如今，"主权AI"已成为每个国家必须回答的现实问题，是提升未来国家竞争力的关 键。 在他看来，"主权AI"其核心在于实现"算力自主、算法自强、生态自立"的完整体系。 他表示，在某种程度上，生态自立比算力自主和算法自强更重要。因为生态自立，意味着芯片要从"能 跑"到"愿意用"，而不是能跑一个软件就可以。"真正决定主权AI成败的，在于是否有足够多的开发者， 愿意长期在这套栈上写代码。" 同时，郑纬民表示，当前中国芯片行业面临着内卷与碎片化问题，例如：不同的厂家提供不同的接口， 要做不同的适配。开发者是生态建 ...

Core Viewpoint - The Moer Thread's first MUSA Developer Conference (MDC 2025) showcased significant advancements in GPU technology and emphasized the importance of ecosystem development in the GPU industry, positioning itself as a key player in the market with a valuation of 300 billion [1][26]. Group 1: Ecosystem and Developer Focus - The ecosystem is identified as the core moat and value in the GPU industry, with a strong emphasis on developer engagement as a critical component of ecosystem construction [2][27]. - The MUSA architecture has undergone comprehensive upgrades centered around developers, aiming to reduce development and migration costs while enhancing usability [2][27]. - Moer Thread's strategy appears to challenge NVIDIA's CUDA ecosystem, highlighting the necessity for domestic GPU manufacturers to build robust ecosystems beyond mere technological competition [2][28]. Group 2: MUSA Architecture and Full-Stack Technology - MUSA (Meta-computing Unified System Architecture) is the first domestic architecture to support AI computing, graphics acceleration, scientific computing, and more on a single chip, representing a full-stack technology system [8][32]. - The full-function GPU is designed to handle various tasks, supporting multiple precision levels, which enhances efficiency and compatibility for emerging applications [9][33]. - MUSA has evolved to its fifth generation since its launch in 2022, with significant performance improvements across various GPU engines [11][35]. Group 3: Upcoming Products and Innovations - The new full-function GPU architecture "Huagang" was introduced, featuring substantial upgrades in computing density, energy efficiency, and precision support [16][40]. - Two upcoming chips, "Huashan" and "Lushan," are set to focus on AI training and high-performance graphics rendering, respectively, indicating a strategic dual focus on AI and graphics [20][44][48]. - "Huashan" will support a wide range of precision calculations and enhance AI training capabilities, while "Lushan" aims to significantly improve graphics performance metrics [23][48].

Core Viewpoint - The article highlights the significant advancements made by Moore Threads in the GPU sector, particularly through the launch of the MUSA architecture and its associated products, which aim to enhance the developer ecosystem and position domestic GPUs at a competitive level in the global market [1][4][19]. Group 1: MUSA Architecture and Innovations - MUSA stands for Meta-computing Unified System Architecture, representing a comprehensive framework that encompasses chip architecture, instruction sets, programming models, and software libraries [6][7]. - The latest GPU architecture, Huagang, boasts a 50% increase in density and a 10-fold improvement in efficiency, with three new chips focusing on AI training, graphics rendering, and intelligent SoC [8][10]. - The MUSA architecture has been iteratively developed over five years, culminating in the latest iteration that optimizes low-precision computing for AI applications [11][13]. Group 2: New Product Launches - Moore Threads introduced three new chips: Huashan, Lushan, and Yangtze, along with two hardware products, AIBOOK and AICube, and the KUAE 2.0 AI Foundry cluster [20][21]. - The Huashan chip targets AI training and high-performance computing, supporting full precision from FP4 to FP64 and significantly enhancing Transformer throughput [22][25][27]. - The Lushan chip focuses on graphics computing, achieving a 64-fold increase in AI performance and a 15-fold improvement in 3A game rendering performance [28][30][31]. - The Yangtze chip is designed for edge computing, providing 50 TOPS of heterogeneous AI computing power for various applications [32][34]. Group 3: Software Ecosystem and Developer Engagement - The MUSA software stack 5.0 was launched, offering a complete toolchain from compilers to AI frameworks, with plans to open-source key components to foster community engagement [15][16]. - Moore Threads aims to build a robust developer ecosystem through the establishment of the Moore Academy, targeting a community of 1 million developers by 2025 [59][61]. - The company emphasizes the importance of a comprehensive ecosystem that integrates software, hardware, and developer trust to create a sustainable competitive advantage in the GPU market [56][58].

转自：北京日报客户端 12月20日-21日 摩尔线程首届MUSA开发者大会期间 "生态"一词成为被频频提及的关键词 摩尔线程高级副总裁高林丽在接受采访时透露，"首个摩尔线程MUSA生态中心将落地海淀，希望未来 能有更多的生态应用和产业项目在此落地、发声、裂变，最终形成一个广阔而良好的生态。" 政策与生态协同，跑出"中国芯"加速度 据了解，摩尔线程自2020年在海淀区成立以来，专注于全功能GPU自主研发，以"一年一芯片"的迭代速 度推出四代产品，在AI智算、数字孪生等领域填补了国内技术空白，是海淀区硬科技企业创新发展的 典型代表。企业的目标是成为具备国际竞争力的GPU领军企业，为融合人工智能和数字孪生的数智世界 打造先进的加速计算平台。企业的愿景是为美好世界加速。 大会现场，摩尔线程宣布将建设MUSA生态中心，并同步启动MUSA开发者计划，旨在为从AI学习到科 研创新的各阶段开发者提供算力支持与技术赋能。据介绍，MUSA生态中心将成为摩尔线程与所有开发 者、合作伙伴及高校师生共同实现愿景的"主场"，承担四项关键使命：一是提供核心技术支撑；二是展 示成果、交流创新；三是建设人才培养基地；四是打造创新孵化和服务平台 ...

Core Insights - The stock price of Moer Thread (688795.SH), the first domestic GPU company listed on the Sci-Tech Innovation Board, surged to 941.08 CNY per share, achieving a market capitalization of 403.1 billion CNY within just five trading days since its listing on December 5, marking the fastest market cap growth for a new stock in 2023 [1][2] Group 1: Stock Performance - Moer Thread's stock price increased by over 700% within five trading days, starting from an opening price of 650 CNY per share on its debut, which was a 468.78% increase from the issue price [2] - On its first trading day, the stock reached a peak of 688 CNY per share, closing at 600.5 CNY, resulting in a single-day gain of 425.26% and a market cap exceeding 282.3 billion CNY [2] - Investors who held onto their shares since the IPO have seen significant profits, with a potential gain of 413,400 CNY for a single lot of 500 shares at the closing price on December 11 [2] Group 2: Market Demand and Policy Support - The rapid growth of Moer Thread's valuation is attributed to three main factors: scarcity of comparable stocks, policy support, and strong industry demand [4] - Moer Thread is recognized as the only domestic company capable of competing with NVIDIA in the GPU market, having developed five chips and four generations of GPU architecture within five years [4] - The company benefits from favorable policies, including the rapid IPO process enabled by the "1+6" reform policy for the Sci-Tech Innovation Board, which supports unprofitable hard-tech companies and prioritizes strategic emerging industries [5] Group 3: Industry Context - The demand for domestic GPU solutions is critical due to increasing technology competition between China and the U.S., with a national strategy emphasizing the need for domestic computing power [5] - Currently, domestic GPUs hold less than 1% market share in China, indicating significant potential for growth and replacement [5] - Moer Thread's listing serves as a benchmark for the domestic GPU sector, potentially catalyzing further investment and development in the industry [6] Group 4: Future Prospects - Following Moer Thread's success, other companies like Muxi Co., which focuses on high-performance GPU chips, are preparing to enter the market, indicating a growing interest in the domestic GPU sector [6] - Muxi Co. has already completed its subscription process and is expected to be the second domestic GPU company listed on the A-share market, with a strong market response anticipated [6][7]

当下，摩尔线程已成为资本市场最亮眼的"明星股"之一。此前，12月5日，摩尔线程上市首日股价涨幅 便高达468%，换手率超85%；开盘十分钟内，股价一度冲高至688元。 国金证券认为，AI时代驱动GPU需求快速扩张，国产替代进入加速窗口期。据弗若斯特沙利文预测， 2024年全球GPU市场规模超万亿元，预计2025–2029年CAGR有望达24.5%。中国市场成长更快，市场规 模将从2024年的1,425亿元跃升至2029年的13,368亿元，2025–2029年CAGR高达53.7%，国产GPU厂商有 望深度受益。摩尔线程作为国内稀缺的全功能GPU厂商，在硬件、生态与集群交付方面持续突破，有望 成为国产GPU加速替代的重要力量。25年上半年公司营收超95%由AI智算业务贡献，同时境内营收占比 超99%。 公开资料显示，摩尔线程成立于2020年，核心研发团队占比超75%，创始人张建中曾任英伟达全球副总 裁。公司自主研发的MUSA架构支持全计算精度，是国内少数具备图形渲染与AI计算能力的GPU企业， 其产品已进入多个省级智算中心采购清单。 （文章来源：21世纪经济报道） "国产GPU第一股"摩尔线程的股价继续狂飙。 ...

哈喽大家好，今天小无带大家聊聊摩尔线程上市的狂欢大戏。千亿市值风光背后藏着巨额亏损，这波资本热潮到底藏着多少隐忧？ 这意味着什么？中一签的投资者最高能赚近27万元，妥妥的年内最"造富"新股啊！当时我寻思着，这是要开启算力股的新时代了？ 结果没想到，高光没撑多久，股价摸到688元的高位就开始往回撤。 发行价才114.28元/股，开盘价直接冲到650元，对应市值瞬间突破3055亿元。 | 全栈式软件 | 人工智能 | 科学计算 | 数字变生 | 游戏娱乐 | 机器人与自动驾驶 | | --- | --- | --- | --- | --- | --- | | 应用于不同场景 | 833 | 英 | G | 6:0) | P | | | MUSA软件找 | 编程语言和API按口 | | 开发者工具链 | | | | ష్ట | 编译器平台 运行环境 | | | 内核驱动 | | MUSA 统一系统架构 | | | | | | | | 硬件层 | 全功量GPUZ片 | | | | | | a p | AI计算引擎 | 科学计算引擎 | 图形计算引擎 | 智能多媒体计算引擎 | | | | 加 | 图 | 12 | 头 ...

新华财经北京12月10日电（胡晨曦） 今日，市场人气热股摩尔线程再度演绎强势拉升行情。 摩尔线程盘中涨幅一度飙升近27%，最高触及797.97元，较首日最高价高出近16个百分点；截至收盘， 其股价收涨近17%，报于735元，大幅反超首日录得的688元最高价。 此外，另一国产GPU龙头沐曦股份已经开启申购，再度引发机构与散户"打新"热潮。 摩尔线程股价再创新高 新一代GPU架构即将发布 12月10日，摩尔线程收盘上涨16.98%，报735元／股，总市值约3455亿元，创历史新高。目前，摩尔线 程股价已超越源杰科技，位居A股市场第三位，仅次于寒武纪、贵州茅台。根据科创板交易规则，摩尔 线程上市后前5个交易日不设涨跌幅限制，且上市首日即可作为融资融券标的。据摩尔线程此前公告， 其上市初期流通股数量仅2938.24万股，占总股本的6.25%，可能存在流动性不足风险，投资者需理性研 判投资价值。 消息面上，摩尔线程12月9日宣布，摩尔线程MUSA开发者大会定档12月19日-20日，其创始人、董事长 兼CEO张建中将首次系统阐述以MUSA为核心的全栈发展战略与未来愿景，并重磅发布新一代GPU架 构、推出涵盖产品体系、核心 ...