Group 1 - MiniMax was listed on the Hong Kong Stock Exchange on January 9, with an opening price of 235.4 HKD, representing a 42.7% increase from the issue price of 165 HKD, and the stock price rose over 50% during early trading [1] - As of the report, MiniMax's stock price reached 271.9 HKD, with a total market capitalization of 83.021 billion HKD [1] - MiniMax is a general artificial intelligence (AGI) company that has developed a series of multimodal general models, including MiniMax M2.1, Hailuo 2.3, Speech 2.6, and Music 2.0, and has launched various AI-native products globally [1] Group 2 - As of September 2025, MiniMax has over 200 countries and regions with more than 212 million individual users [1] - In the first nine months of 2025, the company's revenue grew by over 170% year-on-year, with over 70% of revenue coming from overseas markets [1]

云启资本管理合伙人陈昱（右三）在上市仪式现场。 本文为IPO早知道原创 作者｜ Stone Jin 微信公众号｜ipozaozhidao 据 IPO 早 知 道 消 息 ， MiniMax Group Inc. （ 以 下 简 称 "MiniMax" ） 于 2026 年 1 月 9 日 正 式 以"0100"为股票代码在港交所主板挂牌上市，成为史上IPO规模最大的AI大模型公司。 成 立 至 今 ， MiniMax 已 获 得 多 家 战 略 投 资 方 和 一 线 机 构 的 投 资 和 支 持 。 其 中 ， 云 启 资 本 作 为 MiniMax最早的投资人之一，从天使轮起连续投资6轮。 本文由公众号IPO早知道（ID：ipozaozhidao）原创撰写，如需转载请联系C叔↓↓↓ 作为闫俊杰（注：MiniMax创始人兼CEO）最早期认识的投资人朋友之一，云启资本管理合伙人陈 昱回忆称，初识闫俊杰时，GPT尚未引爆新一轮AGI浪潮。 "当时国内AI行业主流仍倾向于针对特定场景研发专有模型，而俊杰和我的第一次对话中就提到了基 础模型（Foundation Model）路径之于实现通用人工智能的重要意义。公 ...

Core Insights - The article discusses the breakthrough of Nested Learning by Google's DeepMind, which may address the long-standing issue of "catastrophic forgetting" in AI, potentially leading to advancements towards Artificial General Intelligence (AGI) [1][52] - Nested Learning is positioned as a successor to the Transformer architecture, suggesting a shift from passive training to active evolution in AI systems [1][2] Group 1: Nested Learning and AGI - Nested Learning is highlighted as a significant research focus for DeepMind, with predictions that it could lead to minimal AGI by 2028 with a 50% confidence level [7][9] - The concept of Nested Learning is described as a framework that allows AI to build associative memory, enabling continuous learning without the need for retraining [1][19] - Shane Legg, co-founder of DeepMind, emphasizes that there are no current blockers to achieving continual learning, indicating progress in this area [5][7] Group 2: Technical Aspects of Nested Learning - The HOPE architecture is introduced as a mechanism for implementing Nested Learning, which combines fast self-updating systems with slow, multi-timescale memory [6][8] - The article outlines the importance of memory architecture, attentional bias, retention mechanisms, and learning rules in designing effective AI models [20][21] - The Nested Learning framework is said to unify various existing attention mechanisms and optimizers, allowing for a more dynamic understanding of memory in AI [21][24] Group 3: Performance and Implications - The HOPE architecture has shown superior performance in tasks requiring long context and continual learning compared to existing models, indicating its potential effectiveness [33][47] - The article raises concerns about the implications of AI systems that can learn continuously, suggesting that they may develop preferences based on past experiences, which could lead to ethical considerations [52]

Group 1: Company Overview - MiniMax (稀宇科技), a leading global artificial intelligence (AGI) company, was listed on the Hong Kong Stock Exchange on January 9, achieving a market capitalization of over 76.3 billion HKD, making it the largest IPO in the AI large model sector to date [1][3] - The company has developed a series of full-modal general large models, including MiniMaxM2.1, Hailuo2.3, Speech2.6, and Music2.0, positioning itself among the top four global companies in this field [3] - As of September 2025, MiniMax has over 212 million individual users across more than 200 countries and regions, with a revenue growth of over 170% year-on-year for the first nine months of 2025, and over 70% of its revenue coming from international markets [3] Group 2: Market Response and Financials - The global offering of approximately 33.58 million shares was priced at the upper limit of 165 HKD per share, raising approximately 5.54 billion HKD, with the public offering receiving an oversubscription of 1,837 times and the international offering 37 times [1] - The listing attracted 14 prominent cornerstone investors, including Aspex, Eastspring, MiraeAsset, ADIA, Alibaba, and E Fund, indicating strong market confidence and diverse investor interest [1] Group 3: Industry Development in Shanghai - Shanghai's artificial intelligence industry has shown significant growth, with 394 large-scale enterprises generating revenue of 435.49 billion CNY in the first three quarters of 2025, reflecting a year-on-year growth of 39.6% [4] - The city has implemented various measures to foster AI development, including the introduction of resource vouchers and a financial ecosystem to support startups, significantly lowering the barriers to entry for new companies [5] - In 2025, a national AI fund of 60 billion CNY was established in Shanghai, alongside a municipal AI pilot fund of 22.5 billion CNY, creating a comprehensive investment system to support early-stage AI ventures [5]

Core Insights - Beijing Zhiyu Huazhang Technology Co., Ltd. (智谱) officially listed on the Hong Kong Stock Exchange on January 8, marking it as the world's first publicly traded company focused on General Artificial Intelligence (AGI) foundational models [1] - The company's CEO, Zhang Peng, transformed a technical vision into a business valued at over HKD 50 billion within seven years, highlighting a significant milestone for China's AGI industry [1] Group 1: IPO and Market Performance - On its first trading day, Zhiyu's market capitalization exceeded HKD 57 billion, with shares closing at HKD 131.5, reflecting a maximum intraday increase of 13.17% [2] - The IPO was highly successful, with a subscription rate of 1,159.46 times for the public offering and 15.28 times for the international offering, indicating strong market recognition [2] - The total funds raised from the IPO surpassed HKD 4.3 billion, excluding the "green shoe" option [2] Group 2: Business Model and Revenue Strategy - Zhiyu's initial survival strategy focused on localized deployment to meet customized needs of government and enterprise clients, a sector often overlooked by larger companies [3] - The company aims to diversify its revenue structure by increasing the proportion of income from its Model as a Service (MaaS) platform while maintaining a solid base from localized deployments [3] - Zhiyu's software tools and model business have already generated an annual recurring revenue (ARR) exceeding RMB 100 million (approximately USD 14 million), with plans to increase API revenue to 50% of total income [3] Group 3: Research and Development Investment - The company has committed significant resources to R&D, with investments of RMB 84.4 million, RMB 528.9 million, and RMB 2.195 billion planned for 2022, 2023, and 2024 respectively, totaling around RMB 4.4 billion [2] - R&D personnel constitute 74% of the workforce, supporting rapid iterations of the GLM series models every 2-3 months [2] Group 4: Strategic Partnerships and Ecosystem Expansion - Zhiyu has established partnerships with local governments and enterprises, such as the collaboration with Huafa Group in Zhuhai to create a city-level GLM model space [5] - The company is also expanding its reach internationally, having launched a national-level MaaS platform and AI talent training lab in Malaysia [7][6] Group 5: AGI Vision and Future Directions - The ultimate goal of Zhiyu remains the realization of AGI, with ongoing investments in technology exploration and research [4] - The company has begun to explore the Agent field as part of its AGI strategy, viewing it as a critical component in the path to achieving AGI [8] - Future iterations of the AutoGLM series are expected to enhance capabilities in multimodal and embodied intelligence, with applications across various devices [8]

Core Viewpoint - MiniMax, a large model enterprise, successfully listed on the Hong Kong Stock Exchange, experiencing a significant opening surge of 42.67%, with a market capitalization exceeding 70 billion HKD, and further increasing to a 66% rise by the time of reporting [1]. Group 1: IPO Details - MiniMax's global offering consisted of 29.1976 million shares, with 17.4% allocated for public sale and 82.6% for international sale. The share price was set at 165 HKD, raising approximately 4.596 billion HKD in net proceeds [1]. - The Hong Kong public offering saw an overwhelming response with a subscription rate of 1,837.17 times, while the international offering had a subscription rate of 36.76 times. Over 460 institutions participated in the IPO, leading to an oversubscription of more than 70 times [1]. Group 2: Company Background - MiniMax was established in early 2022 and is recognized as a global artificial general intelligence (AGI) company, focusing on the development of multimodal models. It is noted as one of the few companies that have concentrated on this area since its inception and currently holds the highest international revenue among Chinese large model companies [1]. - The company has a young team consisting of 385 members, with an average age of 29 years [1]. Group 3: Leadership - The founder and CEO of MiniMax, Yan Junjie, was born in 1989 in Henan. He graduated from Southeast University in 2006 and obtained his PhD from the Chinese Academy of Sciences in 2015. He has a background in postdoctoral research at Tsinghua University and previously held the position of Vice President at SenseTime, starting as an intern [2].

Core Insights - MiniMax, a leading global artificial general intelligence (AGI) company, went public on the Hong Kong Stock Exchange, achieving a market capitalization exceeding HKD 76.3 billion shortly after its debut [1][2]. Fundraising and Market Response - The company issued approximately 33.58 million shares at the upper limit price of HKD 165 per share, raising a total of approximately HKD 5.54 billion. The public offering was oversubscribed by 1,837 times, while the international offering saw a 37 times oversubscription [3]. - Notable cornerstone investors included Aspex, Eastspring, Mirae Asset, ADIA, Alibaba, and E Fund, indicating strong backing from various investment types, including international long-term, leading technology, and strategic industry investors [3]. Company Performance and User Base - As of September 2025, MiniMax has over 212 million individual users across more than 200 countries and regions, showcasing its extensive global reach [3]. - The company reported a revenue growth of over 170% year-on-year for the first nine months of 2025, with over 70% of its revenue coming from international markets, reflecting a robust and diversified revenue structure [2][3]. Leadership Vision - The founder and CEO of MiniMax, Yan Junjie, emphasized that the true value of an AI company lies in its ability to continuously provide advanced intelligence for societal and economic development [4].

Group 1 - Xiaopeng Motors plans to achieve mass production of humanoid robots in 2026, alongside launching Robotaxi and flying cars [1][3] - The company showcased its latest generation humanoid robot, IRON, at a previous event, which sparked skepticism about its authenticity [3] - The 2026 global product launch introduced several new electric and range-extended models, including the P7+ and G7, featuring advanced battery technology and AI integration [3][5][6][7] Group 2 - Xiaomi awarded its self-developed chip "Xuanjie O1" the highest honor at the 2025 Xiaomi "Ten Million Technology Award" for its innovative design and performance [12] - The company aims to integrate self-developed chips, operating systems, and AI models in a single terminal by 2026, while also pushing forward its robotics business [12] - Xiaomi's new products include the Xiaomi 17 Pro series and various innovative technologies recognized at the award ceremony [13] Group 3 - Donut Lab introduced what it claims to be the world's first mass-producible solid-state battery at CES 2026, boasting high energy density and rapid charging capabilities [37][38] - The battery is reported to maintain performance across extreme temperatures and has a long cycle life, although independent verification of its claims is pending [38][43] - The company plans to scale production significantly in the coming year, aiming to capitalize on its technological advancements [43] Group 4 - Alibaba Cloud launched a multimodal interaction development kit, integrating various AI models to enhance hardware interaction capabilities [34] - The kit is designed to support a wide range of hardware platforms and aims to facilitate rapid deployment of AI solutions across different applications [34] Group 5 - The establishment of JD.com's "Chameleon Business Unit" aims to accelerate the commercialization of AI technologies and enhance market responsiveness [53][54] - The unit will focus on developing core AI products and integrating technological resources to improve efficiency [54] Group 6 - Zhizhu Technology, a company focused on AGI, officially listed on the Hong Kong Stock Exchange, emphasizing its commitment to exploring foundational models for general intelligence [56] - The CEO highlighted the importance of maintaining a focus on AGI development amidst the competitive landscape of AI models [56][57]

Core Insights - The article discusses the rise of AI companies, particularly Zhiyuan and MiniMax, which have successfully gone public in Hong Kong, marking a significant moment for the AI industry in China [7][11][21]. Investment Landscape - In 2021, AI was not a popular investment theme, overshadowed by carbon neutrality and the metaverse, with major players like OpenAI still relatively unknown in China [2][6]. - Early investors, such as Qiming Venture Partners, recognized the potential of large models and invested in Zhiyuan and MiniMax despite the prevailing skepticism in the market [2][6][36]. IPO Success - Zhiyuan and MiniMax both went public in January 2026, with Zhiyuan's market capitalization exceeding HKD 57 billion and MiniMax projected to exceed HKD 50 billion [7][11]. - The IPOs generated significant interest, with Zhiyuan's public offering being oversubscribed by 1,159.46 times, leading to an increase in the public offering ratio from 5% to 20% [7][12]. - MiniMax attracted over HKD 2.5 trillion in subscription funds, indicating high demand for shares [7][12]. Market Dynamics - The article highlights a resurgence in the Hong Kong stock market, with 119 new listings in 2025, the highest in five years, driven by a growing interest in technology companies [12][15]. - The Hong Kong Stock Exchange has become more attractive for tech companies, with new regulations allowing unprofitable firms to list, which has facilitated the entry of AI companies [15][21]. Future Outlook - The rapid evolution of the AI sector is creating a pressing need for funding, with companies like Zhiyuan and MiniMax aiming for global expansion [18][37]. - The article suggests that the successful IPOs of these companies may lead to increased international visibility and valuation, benefiting the broader Chinese tech landscape [41][44].

今天，"人均95后"的大模型公司MiniMax（稀宇科技）正式港股上市。 自去年12月17日起，沐曦股份、英矽智能、壁仞科技、天数智芯先后登陆科创板或港交所。短短一个月 内，五家上海AI产业链相关企业叩开资本市场大门。 ■上海目前拥有集成电路企业超过1200家，汇聚了全国约40%的集成电路产业人才、近50%的产业创新 资源 本报记者 李晔 查睿 ■上海拥有全球范围内首屈一指的营商环境。尤其是上海拿出算力底座场景，为各路GPU公司提供公开 公平的技术比武大平台，"这既能帮助创业者找到自身的短板和快速迭代补足的能力，又能提供最接近 市场的场景环境，更好地让产品说话" 另据证监会官网IPO辅导公示系统显示，燧原科技近期已完成上市辅导，将申报科创板。这意味 着，"上海GPU四小龙"（沐曦股份、壁仞科技、天数智芯、燧原科技）的上市版图合拢在即。此外，去 年A股涨幅榜冠军也来自上海——去年7月起，由智元机器人总裁彭志辉出任董事长的科创板上市企业 上纬新材，一改平淡低迷，股价全年涨幅高达1820%。 上海AI产业链相关企业近期密集引爆资本市场，是幸运之神意外眷顾，还是意料之中？ 绩优硬核 上月17日登陆科创板首日，沐曦股价 ...