Summary of Key Points from Conference Call Records Industry Overview 1. PCB (Printed Circuit Board) - The PCB sector has shown weak performance since Q4 of last year, primarily due to divergent market views on orthogonal backplane solutions, with some believing they may be replaced by copper cables/CPO or delayed until 2028. However, the orthogonal backplane is currently progressing steadily and is expected to enter mass production in H2 2027. Leading companies are experiencing stock price stagnation due to these divergences, highlighting their cost-effectiveness. Recommended companies include Dongshan Precision, Shenghong Technology, and Huidian Co. [1] - The CoWoP (Chip on Wafer on PCB) solution has stronger certainty, can reduce costs, improve efficiency, and bypass the shortage of substrate capacity. The value per square meter of PCB may increase several times, potentially reaching tenfold, with product launches expected by the end of 2027 and full implementation in 2028. Recommended companies in this direction include Pengding Holdings, Shennan Circuit, and Xinsong Technology. [1] 2. Storage - Contract prices remain in an upward cycle despite fluctuations in spot prices. Module companies are expected to see explosive Q1 performance, with Jiangbolong and Demingli realizing low-priced inventory. Recommended design companies include Zhaoyi Innovation (with a profit expectation of 6 billion) and Puran Co., Beijing Junzheng, and Hengshuo Co. [2] - Demand for memory modules is driven by AI servers and general servers, with recommendations for Lanke Technology (long-term profit of 10 billion) and Jucheng Co. (long-term profit of 1.5 billion). [2] 3. Communication - The recent decline in optical modules is related to the pullback of US tech stocks and speculation around CPO concepts. However, industry sources (such as Coherent and Xuchuang) indicate that CPO's potential to replace optical modules in ScaleOut scenarios is low, suggesting that short-term speculation may be excessive. [2] - North American cloud service providers have exceeded capital expenditure guidance for 2026 (620 billion, up 65% year-on-year), indicating potential accelerated demand for optical modules in 2027. Key upcoming catalysts include Nvidia's quarterly report (February 26), GTC conference (March), and OFC exhibition (NPO product showcase). Recommended companies include Zhongji Xuchuang, Xinyisheng, and Dongshan Precision. [2] - For copper connections as a Plan B alternative to orthogonal backplanes, companies to watch include Luxshare Precision, Wokai Nuclear Materials, and Huiju Technology (with potential for stock doubling). [2] - The price of scattered fiber has surged in the short term (from 25 to 50 yuan), but the low willingness of operators to raise prices raises doubts about long-term sustainability. [2] 4. Computing - Domestic computing resources are in short supply, with the recent downtime of Qianwen highlighting the scarcity of AI foundational resources. The demand for CPUs is expected to rise due to increased usage of agents compared to chatbots. Recommended companies include Haiguang Information (benefiting from both CPU and GPU), Cambrian (leading domestic AI chip manufacturer), and Tianshu Zhixin (expected to accelerate integration with leading players). [2] - Cloud infrastructure resources are expected to benefit from price increases, with recommendations for Kingsoft Cloud, Wangsu Technology, and Fourth Paradigm. [2] - In the AI application sector, the recent drop in overseas software and restructuring of SaaS business models may lead to a narrative reversal with the launch of native agent products in Q3 2026. Companies to watch include Alibaba for 2C entry reconstruction and third-party AI agents like TaxFriend, Zhongkong Technology, and Dingjie Smart. [2] 5. Media - Tencent has faced a decline due to market concerns over potential tax increases on internet platforms, although there is no space for increased game value-added tax. The company remains recommended despite rumors of Q4 earnings downgrades, maintaining a PE ratio of 15 times, which still offers value. [3] - The download situation for the Yuanbao app remains stable, and Tencent's AI capabilities may be closing the gap with larger competitors. [3] - In gaming, companies with upcoming catalysts such as Giant Network and Perfect World are recommended for short-term focus, while Century Huatong and Kaiying Network are suggested for medium to long-term attention due to expected catalysts. [3] - Tencent's establishment of a separate AI comic app is beneficial for the production side, which is entering a period of profitability. Recommendations include Kuaishou, Huanrui, and Rongxin. [3] 6. Pharmaceuticals - Attention is drawn to the update of the essential drug catalog, which may accelerate progress. [4] - The probability of inclusion in the essential drug catalog is high for unique products, with several specific products from companies like Jichuan Pharmaceutical and Panlong Pharmaceutical being highlighted. [4] - Emphasis on the global competitiveness of the innovative drug industry chain, with a focus on new-generation ADCs, IOs, small nucleic acids, and CGT. Recommended companies include Kanghong, Yingen, Yunding, and Chengdu Xian Dao. [4] - The brain-computer interface theme is noted, with a potential showcase of non-invasive products during the Spring Festival and a semi-invasive product approval for Borui Kang in March. [4] - Recommendations include Meihua Medical, Dongwei Semiconductor, and Sanbo Brain Science. [5] - The surgical robot sector is expected to see comprehensive implementation of charging policies before August, with overseas orders doubling and maintaining high growth in 2027. Key types include laparoscopic and orthopedic robots, with strong overseas performance for laparoscopic robots. Recommended companies include MicroPort, Jingfeng Medical, Tianzhihang, and Sanyou Medical. [6]

範式智能技術集團股份有限公司 章 程 二○二六年二月 | 第一章 | 總 | | 則 | | | | | | | 1 | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 第二章 | | | 經營宗旨和範圍 | | | | | | | 2 | | | | | | 第三章 | 股 | | 份 | | | | | | | 3 | | | | | | 第一節 | | | 股份發行 | | | | | | | 3 | | | | | | 第二節 | | | 股份增減和回購 | | | | | | | 7 | | | | | | 第三節 | | | 股份轉讓 | | | | | | | 9 | | | | | | 第四章 | | | 股東和股東會 | | | | | | | 10 | | | | | | 第一節 | | | 東 | | | | | | | 10 | | | 股 | | | 第二節 | | | 控股股東和實際控制人 | | | | | | | 14 | ...

北京第四範式智能技術股份有限公司 （前 稱「北 京 第 四 範 式 智 能 技 術 股 份 有 限 公 司Beijing Fourth Paradigm Technology Co., Ltd.」） （於 中 華 人 民 共 和 國 註 冊 成 立 的 股 份 有 限 公 司） 香港交易及結算所有限公司及香港聯合交易所有限公司對本公告的內容概不 負 責，對 其 準 確 性 或 完 整 性 亦 不 發 表 任 何 聲 明，並 明 確 表 示 概 不 就 因 本 公 告 全 部或任何部份內容而產生或因倚賴該等內容而引致的任何損失承擔任何責任。 Phancy Group Co., Ltd. （股 份 代 號：6682） (1) 於二零二六年二月四日（星 期 三）舉行之 臨 時 股 東 會 之 投 票 表 決 結 果； (2) 更 改 本 公 司 中 文 名 稱； 及 (3) 修改公司章程 茲提述北京第四範式智能技術股份有限公司（「本公司」）日期為二零二六年一 月十五日內容有關本公司臨時股東會（「臨時股東會」）之通函（「通 函」）。除 文 義 另 有 所 指 外，本 公 告 所 用 詞 彙 與 通 函 所 界 定 者 ...

FF301 股份發行人及根據《上市規則》第十九B章上市的香港預託證券發行人的證券變動月報表 截至月份: 2026年1月31日 狀態: 新提交 致：香港交易及結算所有限公司 公司名稱: 北京第四範式智能技術股份有限公司 (於中華人民共和國註冊成立的股份有限公司) 呈交日期: 2026年2月4日 I. 法定/註冊股本變動 | 1. 股份分類 | 普通股 | 股份類別 | H | | | 於香港聯交所上市 (註1) | | 是 | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 證券代號 (如上市) | 06682 | 說明 | | | | | | | | | | | 法定/註冊股份數目 | | | 面值 | | | 法定/註冊股本 | | | 上月底結存 | | | 320,809,496 | RMB | | | 1 RMB | | 320,809,496 | | 增加 / 減少 (-) | | | | 0 | | | RMB | | 0 | | 本月底結存 | | | 320,809,496 | RMB | | | 1 RMB ...

Core Insights - The article provides a comparative analysis of several leading AI service providers, focusing on their practical applicability and business adaptability in the AI landscape [1][3][4]. Group 1: AI Service Providers Overview - The analysis includes major players such as Baidu Smart Cloud, Alibaba Cloud, iFlytek, and Fourth Paradigm, each evaluated based on their strengths and weaknesses [3][4]. - The ranking methodology emphasizes not only technical capabilities but also the commercial adaptability of the solutions offered [4]. Group 2: Evaluation Criteria - Practical Implementation Ability (Weight: 40%): This dimension assesses whether AI solutions are user-friendly and derived from real industry scenarios, with verifiable success cases [5]. - Technical Architecture and Forward-Looking Capability (Weight: 35%): Focuses on the core engine's self-development, alignment with cutting-edge technology, and support for private deployment [6][8]. - Comprehensive Empowerment (Weight: 25%): Evaluates whether the product can provide a one-stop solution covering various business functions [10]. Group 3: Individual Company Analysis - Wenzhou ByteCube: Recognized for its practical approach, combining AI with GEO optimization, and has proven adaptability across diverse scenarios, including government services [11][12]. - Baidu Smart Cloud: Positioned as a versatile player with a rich product matrix, but may be too complex for smaller enterprises [13][14]. - Alibaba Cloud: A leader in cloud computing, offering robust infrastructure but requiring strong business development capabilities from clients [14][15]. - iFlytek: An expert in voice recognition and cognitive intelligence, particularly strong in vertical industries like education and healthcare [16][17]. - Fourth Paradigm: Specializes in decision-making AI, excelling in financial risk control and supply chain optimization [18][19]. Group 4: Suitability Rankings - For immediate AI implementation and growth, Wenzhou ByteCube is recommended for its practical solutions and focus on GEO as a new traffic entry point [22]. - For building a comprehensive AI technical foundation, Baidu Smart Cloud and Alibaba Cloud are suitable for larger enterprises with strong technical teams [22]. - For vertical industry intelligence, iFlytek is ideal for clients needing specialized solutions in voice interaction and industry knowledge [23]. - For backend decision-making and operational optimization, Fourth Paradigm is best suited for industries focused on complex decision-making [24].

Core Viewpoint - The collaboration between Paradigm Intelligence and Sunrise aims to significantly reduce AI inference costs to 1 cent per million tokens, facilitating the large-scale application of AI technology [1][9]. Group 1: Cost Reduction Initiative - The "1 cent per million tokens" inference cost plan is designed to address the high costs associated with large model inference, which currently ranges from 0.4 to 2 yuan per million input tokens and 1 to 4 yuan per million output tokens [3]. - The partnership leverages Sunrise's new generation inference GPU chip, the Yingwang S3, and Paradigm's PhanthyCloud cloud service to achieve a 90% reduction in unit token costs in typical scenarios [3][9]. Group 2: Industry Impact - This initiative is seen as a potential turning point for AI infrastructure development, moving the industry from "technical validation" to "scale application" [9]. - The collaboration is expected to enhance the practical application value of domestic computing power, supporting small and medium enterprises and government agencies in adopting intelligent solutions at low costs [7][9]. Group 3: Company Background - Sunrise is a Chinese AI computing chip company focused on developing large model inference GPUs, with a strategic financing of approximately 3 billion yuan completed in the past year [10]. - Paradigm Intelligence, founded in 2014, is a leading global AI technology company with a mission of "AI for Everyone," having implemented over 10,000 AI applications worldwide across various sectors [11].

Group 1 - The company Paradigm Intelligence (06682) announced that Mr. Peng Jun has submitted his resignation as co-company secretary due to job changes, effective from January 26, 2026 [1] - Ms. Yang Xiaohui, another co-company secretary, will assume the role of the sole company secretary [1] - The company's Chief Financial Officer, Mr. Liu Nan, will act as the contact person for Ms. Yang in the company [1]

香港交易及結算所有限公司及香港聯合交易所有限公司對本公告的內容概不 負 責，對 其 準 確 性 或 完 整 性 亦 不 發 表 任 何 聲 明，並 明 確 表 示，概 不 對 因 本 公 告 全部或任何部份內容而產生或因倚賴該等內容而引致的任何損失承擔任何責 任。 Phancy Group Co., Ltd. 北京第四範式智能技術股份有限公司 （前 稱「北 京 第 四 範 式 智 能 技 術 股 份 有 限 公 司Beijing Fourth Paradigm Technology Co., Ltd.」） （於 中 華 人 民 共 和 國 註 冊 成 立 的 股 份 有 限 公 司） （股 份 代 號：6682） 聯席公司秘書卸任 北京第四範式智能技術股份有限公司（「本公司」）董 事（「董 事」）會（「董事會」）謹 此 宣 佈，彭 駿 先 生（「彭先生」）因 工 作 變 動，已 提 呈 卸 任 本 公 司 聯 席 公 司 秘 書 職 務，自2026年1月26日 生 效。 彭 先 生 已 確 認 彼 與 董 事 會 並 無 意 見 分 歧，亦 無 有 關 其 上 述 卸 任 而 需 提 請 香 港 聯 合 交 ...

Core Insights - The article highlights the shift in enterprise demand for cloud services from resource elasticity to intelligent, secure, and scenario-based capabilities, marking a new phase in cloud services with the introduction of Phanthy Cloud by Fan Shi Intelligence [1][9]. Group 1: Product Overview - Phanthy Cloud represents a comprehensive cloud solution that integrates AI engineering, industry modeling, and computing resource scheduling, addressing the urgent market need for quickly deployable and continuously evolving intelligent cloud platforms [3]. - The platform features a three-layer architecture: cloud layer, computing layer, and capability layer, providing a full-stack solution from resource management to intelligent application [5]. Group 2: Cloud Layer - The cloud layer of Phanthy Cloud supports seamless compatibility with public, private, and hybrid deployment models, ensuring scalability while meeting the security and compliance requirements of government and enterprise clients [5]. Group 3: Computing Layer - The computing layer incorporates a high-performance heterogeneous vGPU scheduling engine, supported by the global active computing management community HAMi, enhancing resource utilization and delivery efficiency in multi-team and multi-task scenarios [5]. Group 4: Capability Layer - The capability layer integrates three core pillars: 1. Model Ecosystem: Addresses challenges in domestic replacement by connecting to a leading community of certified models, rapidly increasing from 0 to over 10,000 models in three months [6]. 2. Engineering Platform: Utilizes the industry-leading AI development platform, enabling enterprises to develop, deploy, evaluate, and govern models autonomously [7]. 3. Vertical World Models: Transforms industry knowledge and high-quality data into reusable model assets, allowing AI to perform precise reasoning and optimization in complex decision-making scenarios [8]. Group 5: Future Outlook - The launch of Phanthy Cloud signifies a critical step towards the era of "AI-as-a-Service," focusing on building usable, trustworthy, and evolvable enterprise-level intelligent systems [9][10].

Core Insights - The article highlights the shift in enterprise demand for cloud services from resource elasticity to intelligent, secure, and scenario-based capabilities, marking a new phase in cloud services with the introduction of Phanthy Cloud by Fan Shi Intelligent [1][8] Group 1: Phanthy Cloud Overview - Phanthy Cloud represents a comprehensive cloud solution that integrates AI engineering, industry modeling, and resource scheduling, addressing the urgent market need for quickly deployable and continuously evolving intelligent cloud platforms [3] - The platform is designed to be "ready-to-use and continuously evolving," catering to enterprises' demands for a robust AI infrastructure [3] Group 2: Three-Layer Architecture - Phanthy Cloud features a three-layer integrated design that provides a full-stack solution from foundational resources to top-level intelligence [5] - Cloud Layer: Seamlessly compatible with public, private, and hybrid deployment models, ensuring security and compliance for government and enterprise clients [5] - Computing Layer: Incorporates a high-performance heterogeneous vGPU scheduling engine, enhancing resource utilization and delivery efficiency in multi-team and fluctuating load scenarios [5] - Capability Layer: Integrates ten years of AI core capabilities, including a model ecosystem, engineering platform, and vertical world models [5][6][7] Group 3: Model Ecosystem - The model ecosystem addresses common challenges in domestic replacement processes, allowing enterprises to quickly access verified model assets and significantly reduce selection and integration cycles [6] - The Phanthy Cloud has connected to a leading domestic model community, achieving a rapid increase in certified models from 0 to over 10,000 in just three months [6] Group 4: Engineering Platform - Phanthy Cloud integrates the industry-leading AI development platform, enabling enterprises to develop, deploy, evaluate, and govern models autonomously, ensuring a closed-loop and evolving AI production pipeline [6] Group 5: Vertical World Models - The platform innovatively transforms ten years of vertical world model capabilities into reusable model assets, allowing AI to perform precise reasoning, evaluation, and optimization in complex decision-making scenarios across various industries [7] Group 6: Future of AI Services - The launch of Phanthy Cloud signifies a critical step towards transforming intelligent capabilities into a new type of enterprise infrastructure, aiming to enable every business to build its own future-oriented AI capability stack with lower barriers and higher efficiency [8]