2026年1月8日，成立六年的智谱AI（HK02513）正式登陆港交所，成为全球首家以通用大模型为核心业务的上市公司。发行价为每股116.20港元，开盘120 港元，首日市值为528.28亿港元，其公开发售部分获约1164倍超额认购，募资超43亿港元。 仅隔一天，1月9日，成立仅4年的MiniMax（HK00100）接力挂牌，发行价165港元，开盘即大涨，盘中涨幅一度超109%，市值突破1000亿港元。公开发售 部分获1837倍超额认购，募资约55.4亿港元，刷新近年港股AI新股热度纪录。 短短两天，两家中国大模型头部企业接连上市，累计募资近百亿港元，市值总和逼近1700亿港元。这场资本盛宴背后，两家头部公司迥异的商业路径与尚未 盈利的现实，将中国AI大模型的商业叙事、持续投入逻辑及其即将面临的公众与监管审视，完整推至台前。 商业模式上，二者已现分野。智谱AI以企业级和本地化部署业务为主要收入来源，并在稳固政企客户的基础上，加速向云端与模型即服务（MaaS）模式转 型，试图通过标准化产品降低边际成本、提升规模效应。MiniMax则更强调面向消费者市场的产品化路线，超七成收入来自海外，其C端产品已覆盖全球超 2 ...

智谱由清华大学技术成果转化而来，是一家典型的技术驱动型独角兽，也是"全球大模型第一股"。至 此，全球资本市场将首次迎来一家以自主研发通用人工智能（AGI）基座模型为核心业务的上市公司。 今天，中国AI领域迎来里程碑事件，被称为"中国AI六小虎"之一的"智谱"，在香港成功上市，募集资金 为43.48 亿港元（大约389.85亿元人民币）。 交易首日，智谱股价上涨10%，成功将全球科技领域的目光从硅谷和伦敦转向香港。当交易所的"锣 声"为中国AI敲响，智谱已不只是一家简单的初创公司快速崛起的"范本"，同时也代表了中国AI在全球 竞技场中已走进资本与耐力的"深水区"。 首先，盈利压力将指数级放大。资本市场将用放大镜审视其毛利率、研发投入转化效率及持续造血能 力。大模型训练与推理的巨额成本，将与其收入增长在财报上直接对冲。智谱需要证明，其商业模式的 单位经济模型是健康的、可规模化的。 其次，技术迭代的生死时速并未放缓。上市带来的资源，必须更高效地转化为下一代模型的竞争力。面 对全球巨头在AGI路径上的疯狂冲刺，以及国内百模大战的激烈内卷，智谱需要将其先发上市的优势， 迅速转化为技术代际的领先优势。 其三，全球竞争的维 ...

1月8日上午9时30分，港交所锣声敲响，智谱正式挂牌上市，开盘价报120港元，较116.2港元的发行价 上涨3.3%，截至收盘，智谱股价131.5港元，总市值578.9亿港元。作为"大模型第一股"，以实现AGI （通用人工智能）为目标的智谱引发资本市场高度关注，让投资者得以首次近距离观察国产大模型的商 业化路径，也让整个行业被间接推上资本市场的压力测试台。2022—2025年上半年，智谱营收和亏损额 齐涨，2025年上半年营收1.91亿元，同比增长325%；经调整净亏损17.52亿元，同比扩大70%。从技术 狂欢到资本检阅，智谱和同行们还有不少问题要回答。 总市值近580亿港元 半年前，OpenAI在一份官方报告点名智谱，将其列为在主权AI领域取得显著进展的重要对手之一，当 时智谱完成上市辅导备案刚两个月。半年过去，智谱连接推出GLM-4.5、GLM-4.6和GLM-4.7模型，并 在2026年1月8日成为港交所上市企业。 发行价116.2港元，开盘价120港元，上市当天智谱股价的整体趋势向上，最后报131.5港元，总市值 578.9亿港元。 对于智谱当日股价表现，比达分析师李锦清告诉北京商报记者，"虽然公司 ...

魔搭社区为参赛团队提供了100小时免费GPU算力服务和专家小组的技术支持，并有13万个开源模型、 5500+MCP服务和调试工具供参赛者使用。通过赛事机制和丰富的资源扶持，魔搭社区搭建了一个有效 的试验场，让独立开发者、创业团队、公益机构、学生、残障群体等多样化人群都能深度参与，其中年 龄最小的参赛选手仅16岁，他是来自南京金陵中学的高二学生王毅南，获孤独症儿童绘本LoRA训练挑 战赛二等奖。借助魔搭社区搭建的多方高效协作生态系统，AI创新应用也能更快实现在公益领域的成 长和落地，赛事评委、壹基金儿童发展部主任杨建表示："在科技不断进步的同时，魔搭社区作为有企 业担当和社会责任的平台，通过此次赛事切实推动了科技向善的实践。"此次赛事不仅展示了AI技术在 公益领域的创新应用，也体现了社会各界对弱势群体需求的关注与回应。 赛事共有四个赛道，分别是乡村课堂的AI助教挑战赛、孤独症儿童绘本LoRA训练挑战赛、老年生活有 点AI创新挑战赛、无障碍生活创新挑战赛，参赛选手致力于解决这四大公益场景中的实际问题。例 如，乡村助教赛题一等奖获奖作品智绘科普，由中国科学院大学他山学科交叉创新协会打造，参赛团队 表示："从遥不可及 ...

中经记者 曲忠芳 北京报道 智谱方面称，亏损主要是由于对研发作出重大投资。在报告期内，智谱研发投入逐年增加，2022年至 2024年分别是8440万元、5.29亿元、21.95亿元，占收入的百分比分别是147%、425%、703%。2025年上 半年研发支出为15.95亿元，占总收入的百分比高达835%，而2024年上半年这一百分比数字为1913%。 三年半的时间里累计研发投入超过44亿元。 12月19日，港交所官网显示，北京智谱华章科技股份有限公司（以下简称"智谱"）已通过聆讯并公布了 IPO（首次公开募股）招股书等资料，不仅成为业界所称的"大模型六小龙"中率先步入IPO流程的企 业，而且也抢在OpenAI、Anthropic等海外大模型明星企业前面，有望成为"全球大模型第一股"。随着 488页招股书的披露，公众终于得以一窥智谱大模型生意的"成色"。 需要指出的是，就在12月21日一早，"大模型六小龙"中另一家MiniMax也在港交所披露了聆讯资料集。 《中国经营报》记者注意到，不同于此前多数企业在递表时会先公示一个"申请版本"的招股书，智谱与 近日通过聆讯的壁仞科技、天数智芯等，均是直接披露聆讯后资料集， ...

12月19日，北京智谱华章科技股份有限公司（下称"智谱"）正式披露聆讯后的招股书。 智谱抢在了MiniMax母公司上海稀宇科技有限公司（下称"MiniMax"）之前公开招股书，亦被不少市场人士视为更有希望冲击"全球大模型第一股"。 其中，MiniMax被报道通过了港交所的聆讯，但迄今尚未公布招股书。 由于"第一股"的名头一般是按照上市日期确定，若此后智谱和MiniMax同日登陆港交所，或许也难言谁是第一。 随着此次招股书的披露，外界对于智谱的经营状况也有了更深层次的了解。 智谱的云端部署类型业务是基于token消耗量计费的合约，以及定价还与订阅时长、模型类型与规模及纳入的计算资源量有关，不过该部分业务目前占比仍不 足2成。 为了支撑上述两种交付模式，智谱构建了一个全栈式的产品矩阵，模型矩阵包括端侧小模型、经济型模型和千亿参数的旗舰大模型等各种参数规模，可以针对 性地解决特定客户需求。 同时该矩阵提供对话、通用智能体、代码生成、图像理解、文生图/视频、语音交互等各类功能，实现对各类模型应用场景的充分覆盖。 2022年至2024年，智谱的收入分别为5740万、1.25亿元、3.12亿元。 具体来看，智谱主打Maa ...

在AGI（通用人工智能）的全球竞赛中，中国大模型独角兽企业的资本化进程终于迈出了实质性的一 步。 12月19日晚间，北京智谱华章科技股份有限公司（简称智谱）披露招股书（申请版本，下同），宣布赴 港冲刺"全球大模型第一股"。 这家成立于2019年的公司，由清华大学技术成果转化而来。招股书显示，智谱的营收在过去三年实现年 均复合增长率超过130%。不过，在营收高歌猛进的同时，作为一家典型技术驱动型企业的智谱也面临 着巨额研发投入带来的亏损挑战。 MaaS支撑下收入高速增长，毛利率高达50% 从财务数据看，近几年，智谱的收入增长呈现出明显的加速态势。 2022年至2024年，智谱收入分别为5740万元、1.25亿元和3.12亿元，年均复合增长率超130%。收入增速 在2025年上半年进一步放大，2025年上半年，智谱实现营收1.91亿元，较2024年同期的4490万元同比增 长325.39%。 毛利率层面，2022年至2024年，智谱的毛利率分别是54.6%、64.6%、56.3%，2025年上半年毛利率为 50.0%。 收入的扩张，与其商业模式高度相关。 招股书显示，智谱的商业模式以模型即服务（MaaS）平台为核 ...

Core Viewpoint - The article discusses the competitive landscape of China's AI industry, highlighting the strategic choices of major players like Alibaba, ByteDance, and Tencent as they navigate a battle for capital efficiency, infrastructure dominance, and traffic entry points. Group 1: Alibaba's Strategy - Alibaba is adopting a "full-stack" approach similar to Google's, with a significant capital expenditure increase of 80% year-on-year, reaching 32 billion RMB in the September quarter [2] - The company's cloud revenue grew by 29% year-on-year, with AI-related revenue achieving triple-digit growth for the ninth consecutive quarter, and is expected to accelerate to 38% growth in the December quarter [3] - Alibaba aims to establish a "full-stack" barrier in the AI market, positioning itself as a dominant player through heavy asset investment [4] Group 2: ByteDance's Approach - ByteDance leverages its massive traffic advantage, with a daily token consumption of 30 trillion, approaching Google's 43 trillion, significantly outpacing competitors like Baidu and DeepSeek [7] - The company's app "Doubao" leads in domestic AI application activity, while its overseas education app Gauth saw a 394% year-on-year revenue increase [8] - ByteDance's strategy creates substantial inference demand, allowing it to encroach on traditional cloud giants in the Model as a Service (MaaS) sector, capturing 49.2% of the public cloud market share for large models [11] Group 3: Tencent's Strategy - Tencent maintains a conservative approach, reducing capital expenditures while focusing on seamlessly integrating AI capabilities into its extensive social and payment ecosystem [12][14] - The company has integrated its AI assistant "Yuanbao" into WeChat Pay, enhancing operational efficiency for small and medium-sized businesses [14] - Tencent's strategy emphasizes high implementation certainty despite lower capital expenditure figures compared to competitors [14] Group 4: Competitive Dynamics - The competition between China and the U.S. in AI has entered a "dynamic alternation" phase, with Chinese models rapidly iterating and catching up within 3-6 months after significant advancements in U.S. models [4][17] - Chinese companies exhibit resilience through unique "Chinese speed" and open-source ecosystems, with 80% of AI startups utilizing open-source models [17] - Cost control is a competitive advantage for Chinese models, as seen with Kuaishou's "Kling" video generation model, which offers significantly lower prices than global counterparts [17] Group 5: Valuation Insights - Goldman Sachs analysts assert that the Chinese AI sector is not in a bubble, with projected P/E ratios for Tencent and Alibaba at 21x and 23x for 2026, respectively, lower than those of major U.S. tech companies [18]

Core Insights - The report by Goldman Sachs highlights the intense competition in China's AI sector, focusing on the strategic choices of major players like Alibaba, ByteDance, and Tencent, and suggests a new normal of "dynamic alternation" in the US-China AI competition [1][2] Group 1: Alibaba's Strategy - Alibaba is adopting a "full-stack" approach similar to Google's, with a significant increase in capital expenditure, which surged by 80% year-on-year to reach 32 billion RMB in the September quarter [3][4] - The company's cloud revenue grew by 29% year-on-year, with AI-related revenue achieving triple-digit growth for the ninth consecutive quarter, and is expected to accelerate to 38% growth in the December quarter [4][6] Group 2: ByteDance's Approach - ByteDance is leveraging its massive traffic advantage, with a daily token consumption of 30 trillion, approaching Google's 43 trillion, and significantly surpassing competitors like Baidu [9][13] - The company's application "Doubao" leads in domestic AI application activity, while its overseas education app Gauth saw a 394% year-on-year increase in monthly revenue [9][13] Group 3: Tencent's Strategy - Tencent is maintaining a conservative approach, reducing capital expenditure while focusing on seamlessly integrating AI capabilities into its extensive social and payment ecosystem [14][15] - The company has integrated its AI assistant "Yuanbao" into WeChat Pay, enhancing operational efficiency for small and medium-sized businesses [15] Group 4: US-China AI Competition - The report outlines a "dynamic catch-up" cycle in the US-China AI competition, where Chinese models typically follow significant advancements in US models within 3-6 months [16][17] - Chinese companies are noted for their resilience and aggressive cost control, with many leveraging open-source models to enhance their capabilities [17] Group 5: Valuation Insights - Goldman Sachs indicates that the current state of the Chinese AI sector does not reflect a bubble, with projected P/E ratios for Tencent and Alibaba at 21x and 23x respectively, lower than those of major US tech companies [18]

Core Insights - The third Marine Intelligent Computing Conference was held in Guiyang from November 5 to 7, focusing on the integration of high-performance computing and artificial intelligence in marine science [1][3] - The conference aimed to create a global and open platform for technology and academic exchange, gathering experts and leaders in the field [1][3] Group 1: Importance of Marine Intelligent Computing - The ocean is a critical strategic space for national development, covering 71% of the Earth's surface, and marine intelligent computing is essential for understanding and managing marine resources [3] - High-performance computing and AI are transforming marine research paradigms, shifting from "observation-driven" to "computation-driven" and "intelligence-driven" approaches, enhancing capabilities in disaster warning, resource development, and environmental management [3][5] Group 2: Innovations and Applications - The integration of high-performance computing and AI with marine science has led to groundbreaking innovations, such as AI-driven storm surge disaster risk warning systems and high-resolution marine forecasting models [5][6] - The conference featured discussions on marine big data, intelligent forecasting, and new domestic software and hardware technologies, promoting key technological innovations and applications [5][6] Group 3: Expert Contributions and Research - Experts presented cutting-edge research, including advancements in storm surge disaster studies, marine satellite technology, and four-dimensional marine monitoring techniques [6] - The development of large models for marine forecasting is highlighted, showcasing their importance in various applications, including marine environment analysis and ice prediction [6] Group 4: Future Directions and Infrastructure - The demand for computational power in the marine sector is increasing, prompting the need for a robust ecological architecture for new computational bases to support marine model development [8] - The concept of Model as a Service (MaaS) is introduced, allowing users to access AI models as standardized cloud services, significantly lowering the barriers to AI adoption in marine applications [6][8]