Core Viewpoint - The article discusses the public offering of Beijing Zhipu Huazhang Technology Co., Ltd. (referred to as "Zhipu") and its potential to become the "first global large model stock" ahead of its competitor MiniMax, which has yet to release its prospectus [2][22]. Financial Performance - Zhipu's revenue projections for 2022 to 2024 are 57.4 million, 125 million, and 312 million respectively [6]. - The primary revenue source is from providing private AI models, expected to generate 264 million in 2024, accounting for over 80% of total revenue [8]. - The company is currently operating at a loss, with losses of 144 million, 788 million, and 2.958 billion from 2022 to 2024, totaling 3.89 billion [15]. Business Model - Zhipu's business model includes two main deployment types: local and cloud-based services [7]. - Local deployment is characterized as a "one-time project," with high pricing based on model type, scale, and implementation costs [9][10]. - The top five clients contributed 1.42 billion in revenue in 2024, representing 45.5% of total revenue [11]. - Cloud deployment is based on token consumption and subscription duration, but currently accounts for less than 20% of revenue [12]. Technological Advancements - Zhipu has made significant advancements in technology, with its GLM series models undergoing upgrades every 3 to 6 months, leading the industry in iteration efficiency [17]. - The GLM model has achieved top performance in code generation, ranking alongside leading models from Anthropic and OpenAI [18]. - The company has gained recognition in the international academic and industrial sectors, with its ChatGLM model highlighted as a prominent representative of Chinese foundational models [19]. Client Base and Funding - As of September 30, 2025, Zhipu's GLM model serves 12,000 enterprise clients, over 80 million end-user devices, and more than 45 million developers, making it the largest independent general-purpose large model provider in China [20]. - Since its inception, Zhipu has completed eight rounds of financing, raising over 8.3 billion RMB, with investments from top-tier capital and notable industry players [21].

（来源：千龙网） 凭借领先的技术与商业模式，智谱已连续三年营收翻倍。招股书显示，2022年、2023年、2024年收入分 别为5740万、1.245亿、3.124亿，年复合增长率达到130%。2025年上半年收入为1.9亿。 收入高速增长，毛利率也持续保持在50%以上。招股书显示，2022年、2023年、2024年毛利率分别是 54.6%、64.6%、56.3%，2025年上半年毛利率为50%。 原创大模型技术持续领先，智谱凭借收入规模最大、高增长、高毛利率等领先性优势，自成立以来，受 到了国资、产业资本、VC/PE等各类众多知名投资机构认可。在IPO之前，智谱已完成了8轮融资，融资 规模超83亿元。 行业人士指出，一方面智谱为全球投资者提供了分享中国AI基础层技术红利的纯正标的，有望重塑中 国科技股在资本市场的估值逻辑；另一方面，这也预示着中国大模型产业将从早期的狂热投入，正式进 入以技术实力、营收能力与可持续商业模式为核心考量的新发展阶段。 中国最大的独立大模型厂商，北京智谱华章科技股份有限公司，已率先通过港交所聆讯并正式递交了招 股书，有望以"全球大模型第一股"身份在港交所挂牌上市。这标志着资本市场将首 ...

二、大模型应用在财富管理领域还是方兴未艾的阶段，还需要继续实践和推进。C端投资者接受度呈 现"冰火两重天"，63.8%的投资者尝试用大模型指导投资，但78%的投资者缺乏专业金融知识； 四、成本ROI顾虑突出，线下算力集群投资过亿，中小机构难以承受。 三、行业基础设施不完善，数据公开与私有化并存，接口标准化不足，性能与稳定性有待提升； 针对这些痛点，陈风建议采用"小步快跑"的落地策略，优先通过调用大模型+Agent的形式开展业务验 证，利用公共云算力将成本降低100倍，在控制风险的前提下快速推广投资策略优化、C端服务升级等 应用，逐步跑通商业逻辑。 专题：财经年会2026：预测与战略暨2025全球财富管理论坛 《财经》年会2026：预测与战略暨2025全球财富管理论坛于2025年12月18-20日在北京举行。阿里云智 能新金融行业副总经理、资深研发总监陈风介绍，阿里云"通义点金"大模型自2023年推出以来，已与中 金、易方达等180余家金融机构达成合作，30余家实现生产上线。 但他也坦言，当前大模型在财富管理领域的落地仍面临四大痛点： 一、技术成熟度存疑，不同模型在投资决策等核心场景表现差异显； 新浪声明：所有 ...

Core Insights - The article discusses the evolution of the autonomous driving industry, highlighting the achievements of the company WeRide under the leadership of Han Xu, who emphasizes the importance of talent acquisition and technological advancements in the field of Robotaxi [1][2][5]. Group 1: Company Achievements - WeRide has become the first publicly listed Robotaxi company in both the US and Hong Kong, marking a significant milestone in its eight-year journey [2][8]. - The company has successfully transitioned from a phase of skepticism about autonomous driving to achieving operational milestones, including the removal of safety drivers from vehicles [17][18]. - WeRide has deployed its Robotaxi services in 11 countries, demonstrating its global reach and operational capabilities [15][18]. Group 2: Industry Insights - Han Xu asserts that only three companies in China can truly operate Level 4 (L4) autonomous vehicles, emphasizing the technical barriers that still exist between Level 2 (L2) and L4 technologies [6][19][22]. - The article highlights the distinction between companies claiming to have L4 capabilities and those that have demonstrated actual operational success with a fleet of autonomous vehicles [21][24]. - Han Xu predicts that if Tesla continues to rely solely on production vehicles without integrating advanced sensor technologies, it will struggle to achieve the same level of autonomy as competitors like Waymo [45]. Group 3: Talent Acquisition and AI Impact - The company is actively recruiting top talent with salaries ranging from 3 to 5 million, reflecting the increasing demand for skilled professionals in the AI and autonomous driving sectors [46][49]. - Han Xu describes AI as a significant amplifier of talent value, suggesting that exceptional individuals can command much higher salaries in the current market [46][48]. - The company aims to attract talent by offering competitive compensation packages, indicating its financial strength and commitment to innovation [50][51]. Group 4: Future Predictions - Han Xu forecasts that within three years, if Tesla does not adopt multi-modal sensor technology for its Robotaxi, it will not reach the operational standards set by Waymo [53]. - He also predicts the emergence of a "Superdriver" within eight years, a level of autonomous driving that surpasses the capabilities of the best human drivers [53].

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！ 2025年12月20-21日，第二届"深圳香蜜湖金融年会"在深圳市福田区隆重举行，本届年会以"识变局，开 新局——促进粤港澳大湾区科技-产业-金融良性循环"为主题。中国金融传媒集团特聘高级专家、中国 银行业协会原首席信息官、深圳香蜜湖国际金融科技研究院学术委员会委员高峰出席并发言。 专题：2025年深圳香蜜湖金融年会 高峰首先分析了在大模型带来系列挑战背景下智能金融的技术架构选择。他指出，当前主流AI架构主 要有两种：一是集中式平台化AI架构，通过API/SDK对外提供服务，即"+AI"模式，为目前多数金融机 构所采用；二是将大模型与多智能体技术深度嵌入业务的新一代智能平台，即"AI+"模式，互联网银行 及部分大型银行正朝此方向演进。此外，混合架构也普遍存在。无论哪种架构，都离不开四大关键能力 的支撑，即数字基础设施、数据资产管理体系、算法平台和模型管理。 针对大模型应用中备受关注的"幻觉"问题，高峰提出了金融级的综合治理思路。在数据侧，需确保"喂 给AI靠谱的料"，通过规范流程、整合异构数据、建立动态更新机制等手段，在安全前提下促进数 ...

转自：新华财经 新华财经北京12月20日电（记者沈寅飞）近日，大模型厂商北京智谱华章科技股份有限公司（以下简 称"智谱"）已率先通过港交所聆讯并正式递交了招股书。 据介绍，成立于2019年的智谱，由清华大学技术成果转化而来，定位为专注于基础模型研发的独立厂 商。凭借原创的GLM（General Language Model）预训练架构，构建了覆盖语言、代码、多模态及智能 体的全栈模型矩阵，模型适配了40余款国产芯片。 截至今年6月，该公司研发人员占比74%，核心科研团队和学术顾问团队已发表500篇顶尖高影响力论 文，累计引用次数超过58000余次。除了汇聚了当前国内人工智能顶级人才外，公司高度重视研发投 入。招股书显示，2022年、2023年、2024年公司研发投入分别为8440万、5.289亿、21.954亿，2025年上 半年研发投入为15.947亿，累计研发投入约44亿。研发投入支撑了其技术快速迭代，GLM系列模型每3- 6个月完成一次基座迭代。 目前，智谱收入主要来自于大模型收入。公司采用MaaS模式，即通过API调用向开发者和企业输出智能 能力。招股书显示，智谱大模型已赋能了全球12000家企业客户 ...

Core Insights - ByteDance has reported significant financial growth, with profits exceeding $40 billion in the first three quarters of this year, and is projected to reach $50 billion for the year, averaging a daily profit of approximately $9.64 million [3][5][11] - The company is increasing employee salaries across the board, with a 1.5 times increase in salary adjustment investment for the current performance evaluation cycle [5][14] - A new salary and job level system will be implemented in January 2026, aimed at providing greater salary increase potential for employees [9][10][18] Financial Performance - ByteDance's revenue is expected to reach $186 billion this year, marking a 20% increase from the previous year, with a projected net profit margin of 26.9% [5][11] - The company's valuation has risen significantly, with reports indicating a valuation of $330 billion in September, later increasing to $480 billion following stock buybacks and investment interest [5][11] Salary Adjustments - The salary structure will see an increase in cash compensation, with a shift from a four-year vesting schedule for stock options to a three-year schedule [6][14] - Performance bonuses will increase by 35% compared to the previous cycle, with adjustments in the distribution of cash and stock options [6][14][16] Job Level System Changes - The job level system will transition from a numerical format to a new L1-L10 format, consolidating certain levels to streamline the hierarchy [9][18] - The new system aims to provide employees with more opportunities for salary increases and better align with industry standards [10][13] Strategic Rationale - The salary increases and job level changes are part of ByteDance's strategy to attract, motivate, and retain talent in a competitive market, particularly in the context of the growing AI sector [11][12]

Core Viewpoint - Beijing Zhiyu Huazhang Technology Co., Ltd. (Zhiyu) has submitted its prospectus to the Hong Kong Stock Exchange, marking the first complete financial disclosure among the "AI Six Tigers" [1][2]. Financial Performance - Zhiyu's revenue for 2022, 2023, and 2024 is projected to be RMB 57.4 million, RMB 124.5 million, and RMB 312.4 million, respectively, with a compound annual growth rate (CAGR) of 130% [2]. - The company reported net losses of RMB 1.43 billion, RMB 7.88 billion, and RMB 29.58 billion for 2022, 2023, and 2024, respectively, with a cumulative loss exceeding RMB 6.2 billion [2][4]. - Gross margins have remained above 50%, with figures of 54.6%, 64.6%, 56.3%, and 50.0% for the years 2022, 2023, 2024, and the first half of 2025 [2][4]. Business Model - Zhiyu operates a dual-driven business model focusing on localized deployment for B-end and G-end clients, which accounted for 84.5% of revenue in 2024, while cloud deployment contributed 15.5% [5]. - The company aims to increase the revenue share from its API business to 50% [7]. Market Position - According to Frost & Sullivan, Zhiyu ranks first among independent general model developers in China and second overall, with a market share of 6.6% in 2024 [2]. Client Base and Growth - As of the first three quarters of 2025, Zhiyu has served over 12,000 institutional clients, showing significant growth [3]. - The company anticipates a revenue increase of over 60% year-on-year for Q3 2025, despite expected widening net losses due to substantial R&D expenditures [3]. Strategic Initiatives - Zhiyu has initiated the "Z Plan" to foster an ecosystem for large model entrepreneurs, providing resources and technical support to early-stage startups [8][9]. - The company has also established a venture capital fund, "Z Fund," with a management scale of RMB 1.5 billion, investing in several startups [9]. International Expansion - Zhiyu is actively expanding its overseas business, particularly in Southeast Asia, where it has begun generating revenue from localized deployment services [10]. - The company is involved in building international alliances for AI infrastructure, positioning itself as "China's OpenAI" [8][10]. Funding and Valuation - Since its inception, Zhiyu has completed eight rounds of financing, raising over RMB 8.3 billion, with a post-financing valuation of approximately RMB 24.38 billion [11][12].

（文/陈济深 编辑/张广凯） 12月19日，北京智谱华章科技股份有限公司（智谱）正式向港交所递交招股书，打响了中国大模型企业上市的第一 枪。 这不仅意味着智谱有望成为"全球大模型第一股"，也标志着中国大模型赛道正式迈入资本化新阶段。 作为全球AI新势力中的重要玩家，招股书也第一次揭示了这家明星独角兽商业化的"冰火两重天"。 过去三年，公司收入分别为0.57亿元、1.25亿元、3.12亿元。2025年上半年持续加速，短短6个月就入账1.9亿元，同比 增长325%。 收入增长的背后是智谱惊人的烧钱速度：过去三年，智谱经调整净亏损分别为0.97亿、6.21亿以及24.66亿元。 不过值得注意的是，智谱在B端业务占据主导的情况下，面对传统B端大厂和互联网巨头的竞争，长期维持住了超过 50%的毛利率。 随着2025年字节火山引擎、阿里云、百度等互联网巨头纷纷下场加码AI大战，智谱的这份招股书提供了一个观察中国 大模型企业在全球AI淘汰赛下真实状态的绝佳切片。 巨头夹缝中的"独立样本" 智谱的核心团队成员，基本都来自清华大学计算机系知识工程实验室，长期从事大规模预训练模型、语义大数据分 析、智能问答等领域的研究。 有这样 ...

Core Viewpoint - The application of large models extends beyond chatbots and programming assistants, and their true value will be realized across various industries in the future [8]. Group 1: Types of Companies Needing Private Models - Three types of companies require industry-specific or private models: traditional large enterprises, small and medium-sized enterprises with vast amounts of data, and disruptive new companies [8][34]. - Traditional large enterprises often possess valuable industry-specific data [34]. - Small and medium-sized enterprises specializing in niche areas can leverage their data as a source for large models [35]. - Disruptive companies in sectors like finance, pharmaceuticals, and e-commerce are most likely to benefit from developing their own private models [35]. Group 2: Implementation Criteria - Companies that only handle daily office tasks or primarily text data do not need to develop private models and can utilize existing large model APIs [4][37]. - If a company has sufficient text data, it can implement a Retrieval-Augmented Generation (RAG) model combined with a large model API instead of building its own [38]. - Companies with vast multimodal data or stringent privacy requirements, such as those in oil exploration or pharmaceuticals, should consider developing a private model [38]. Group 3: Market Predictions - The large language model market is predicted to be divided into three segments: domain-specific LLMs, general-purpose LLMs, and private LLMs [39][41]. - By 2033, domain-specific models are expected to capture approximately 40% of the market share, while general-purpose and private models are projected to each hold around 30% [47]. Group 4: Training and Optimization - The key to successfully deploying large models for business is post-training or agentization, which differentiates models from standard APIs [42]. - Companies should focus on maximizing computational efficiency and developing effective fine-tuning templates to create their industry-specific models [43][44]. - The company has developed a fine-tuning SDK to facilitate the creation of private models, allowing users to focus on model and algorithm innovation [17][45]. Group 5: Real-World Applications - A world-renowned automotive company has utilized this technology to create a multimodal automated decision support system [53]. - A leading e-commerce company's autonomous driving business has significantly improved with the help of this technology [53]. - Another world-class automotive company has developed an intelligent cockpit model with assistance from this technology [53].