消息面上，开源项目Clawdbot在硅谷爆火，7x24的Agent重塑AI体验，可在Mac mini上运行，兼具本地AI智能体和聊天网关双重身 份，通过WhatsApp、iMessage等随时对话。Clawdbot解决了大模型记忆力痛点，能记住两周前的对话，还会主动推送邮件、日程提 醒，并可直接操控电脑执行任务。 业绩方面，1月28日晚间，软件ETF汇添富（159590）标的指数第一大重仓股——科大讯飞发布业绩预告，公司预计2025年净利润为 7.85亿元至9.5亿元，同比增长40%至70%；扣非后净利润为2.45亿元至3.01亿元，同比增长30%至60%。除盈利水平提升外，科大讯 飞多项核心财务指标表现亮眼：2025年公司营收、毛利保持正向增长，销售回款及经营性净现金流均创历史新高，其中销售回款总 额超270亿元，同比增长超40亿元；经营性净现金流超30亿元。对于业绩增长原因，科大讯飞表示主要得益于公司基于自主可控平台 持续夯实行业地位，人工智能应用规模化落地成效日益显现。 软件ETF汇添富（159590）标的指数权重股多数飘红：科大讯飞涨停，三六零、用友网络涨超5%，金山办公涨超4%，深信服涨超 3%，恒 ...

Core Insights - The core observation is that in the AI era, high-quality data assets are becoming the decisive factor for companies to realize AI value, rather than merely competing on model capabilities [1][2] Group 1: AI Model and Data Challenges - The update cycle for top AI models has shortened to a few months, while user retention for specific models can drop significantly within 12 months, indicating that relying solely on proprietary models for competitive advantage is costly and has a short "shelf life" [1] - Many enterprise AI projects face the challenge of "stunning demonstrations but difficult implementations," primarily due to the "sleeping" and "disordered" nature of business data, which includes various documents that are not easily understood by AI [2] Group 2: WPS365 and Data Utilization - WPS365 aims to build a "comprehensive knowledge base" for enterprises, focusing on activating dormant unstructured data through strong document parsing and knowledge governance capabilities [2] - The WPS365 platform provides a full-link solution from data aggregation to intelligent parsing and knowledge governance, exemplified by its intelligent document library that can detect conflicts and convert unstructured data into structured knowledge graphs [3] Group 3: Competitive Landscape and Implementation - The WPS365 "enterprise brain" has been implemented in leading companies in East China, achieving efficiency improvements in knowledge asset management, intelligent compliance review, and automated report generation [4] - The competition in enterprise AI has entered a new phase centered on high-quality data governance capabilities, shifting the focus from merely the intelligence of models to the ability to understand and utilize data effectively [3][4]

作者: 导语：截至2026年1月27日软件ETF汇添富（159590）近一个月净值涨幅高达13.89%，现出了明显的"高 弹性"。 一、 行情透视：AI应用与智能体（Agent）逻辑的超额溢价 基于Wind最新数据显示，从近5个交易日的资金净流向来看，市场情绪表现得极为复杂： | | 份额变动(万 | 资金净流向(万 | | | --- | --- | --- | --- | | 日期 | (સે) | 元) | 盘面特征 | | 2026-01- | -500 | -2473.78 | 缩量调整。 部分获利盘回吐 | | 27 | | | | | 2026-01- | +2500 | +2786.85 | 关键位放量吸筹,多头反攻 | | 21 | | | | | 2026-01 - +4000 | | +4083.61 | 机构大额入场,定调Al 应 | | 1.5 | | | 用春季躁动 | 虽然1月27日份额减少了500万份，出现了约2473.78万元的净流出，但在近一个月涨幅近14%的背景 下，这更倾向于高位获利盘的正常清洗。1月15日与21日合计超6800万元的大额净流入，构成了当前 1.27元净 ...

1月27日，金山办公宣布旗下WPS 365"企业大脑"已在华东地区组织级客户中率先落地。作为由全域知识驱动的核心引擎，该方案通过深度赋能，实现了业 务效率显著提升与知识资产的高效盘活，交出了多项关键成果的亮眼答卷。 对此，金山办公助理总裁朱熠锷提出，企业AI应用正加速从"以模型为中心"转向"以数据为中心"，数据质量成为决定企业AI应用效果的关键。WPS 365以知 识增强生成（Knowledge-Augmented Generation）为支撑，让大模型"掌握"企业真正的知识资产。 从 RAG 到 KAG ，让大模型真正 " 掌握 " 企业知识 近三年来，尽管大模型能力飞速发展、算力价格持续下降，但技术层面的进步并未自动转化为企业AI生产力的实质性突破。企业在应用AI时会遭遇复杂文 档解析、知识冲突、企业黑话、问答相关性弱等现实挑战。高质量的数据治理已经成为AI融入千行百业的关键因素。 中金公司研究部执行总经理、计算机行业首席分析师于钟海介绍，头部模型的快速迭代让企业在模型侧作出的努力"保鲜期"极短，越来越多的企业选择外部 采购模型，而数据作为AI时代的企业底盘，将是企业唯一可持续的AI护城河。 针对这一行业 ...

Group 1 - The core viewpoint of the news highlights the performance of the AI sector, particularly the rise of the Zhongzheng AI Theme Index and its constituent stocks, indicating a positive trend in the market [1] - The Zhongzheng AI Theme Index (930713) increased by 0.25%, with notable gains from stocks such as Beijing Junzheng (up 7.93%), Guanghuan Xinnet (up 4.95%), and others [1] - The AI Artificial Intelligence ETF (512930) also saw an increase of 0.38%, with the latest price reported at 2.4 yuan [1] Group 2 - Microsoft has launched its second-generation AI chip, Maia 200, aiming to reduce reliance on Nvidia, with production handled by TSMC and deployment in data centers in Iowa and Phoenix [1] - The chip's control software is now available for developers, although the timeline for Azure cloud service users to access servers based on this chip remains unclear [1] - Microsoft has alternative options for chip designs through its partnership with OpenAI, should internal development face challenges [1] Group 3 - According to GF Securities, the demand for high-frequency random access in vector databases driven by AI agents is significantly increasing, leading to a rise in eSSD capacity requirements for GPUs [2] - The NAND demand growth rate is projected to reach 21% by 2026, while supply constraints due to capital expenditure shifts towards HBM/DRAM are expected to limit capacity expansion [2] - TrendForce forecasts a substantial increase in NAND contract prices, with an annual rise of 105%-110% and a quarterly increase of 55%-60% in Q1 [2] Group 4 - The Zhongzheng AI Theme Index consists of 50 listed companies involved in providing resources, technology, and application support for AI, reflecting the overall performance of AI-related securities [2] - As of December 31, 2025, the top ten weighted stocks in the Zhongzheng AI Theme Index include companies like Zhongji Xuchuang and New Yisheng, accounting for 58.08% of the index [2]

Summary of Conference Call Notes Industry and Company Focus - The conference call primarily discusses the **AI industry** and **commercial aerospace** sector, with a specific emphasis on **Agent technology** and **multi-modal applications**. Key Points and Arguments AI Industry and Agent Technology 1. The call emphasizes a bullish outlook on the **AI industry** and **commercial aerospace**, particularly focusing on the **Agent** sector and its potential impact on various industries [1][2]. 2. Recent product iterations in the **Agent** space, such as **Automic's Skills** and **Minimax's Agent 2.0**, indicate a significant acceleration in product development and capabilities [2][3]. 3. The evolution of **Agent technology** is moving from single-task completion to handling complex tasks and cross-application operations, suggesting that Agents may become the core of the next-generation operating systems [3][4]. 4. The interaction paradigm is shifting from programming languages to natural language, indicating a fundamental change in how users will interact with computers [4][5]. 5. Recent funding rounds for vertical-specific Agents, such as AI intelligence platforms, show a growing interest and investment in this area, with valuations reaching **$50 million** for some startups [5][6]. Product Developments and Market Trends 6. Products like **CoWork** and **Minimax's Agent 2.0** are designed to enhance workflow integration, showcasing the trend towards human-machine collaboration [7][8]. 7. The **Agent** products are expected to evolve to include capabilities for automatic perception and proactive responses, enhancing their utility in various applications [9][10]. 8. The future of **Agent technology** is projected to include multi-agent systems that can collaborate and self-iterate, as well as expand into physical interactions through robotics [11][12]. Multi-Modal Applications 9. The call highlights the importance of **multi-modal** capabilities, with companies like **ByteDance** and **Kuaishou** making significant advancements in this area, leading to increased user engagement and revenue growth [13][14]. 10. The **multi-modal** products are addressing industrial challenges and enhancing the consistency and effectiveness of generated content, which is expected to drive further adoption [14][15]. 11. The **AI industry** is anticipated to see a surge in demand for **multi-modal** applications, particularly in the realm of AI-generated content, such as **AI comics** and videos [16][17]. Investment Opportunities 12. Investment focus areas include: - Large tech companies involved in **CapEx** and ecosystem investments. - Companies in the **Agent** space focusing on digital governance and vertical applications, such as **Kingdee International** and **Dingjie Smart** [19][20]. - Cloud service providers and chip manufacturers, particularly those involved in private cloud deployments and AI applications [20][21]. 13. The **satellite communication** sector is highlighted as a key growth area, particularly with advancements in **LEO satellites** and laser communication technologies, which are expected to address latency and bandwidth issues [22][23]. 14. The market for satellite communication is projected to reach **$83 billion** by 2023, with significant contributions from companies like **SpaceX** and their **Starlink** project [24][25]. Conclusion 15. The call concludes with a strong recommendation to focus on the **Agent** and **multi-modal** sectors for investment, as they are expected to drive significant changes in both B2B and B2C software landscapes [26][27].

相关标的： 【浙商计算机】Clawdbot：现象级开源 AI Agent，重构人机协作与企业智能化落地 2026 年初开源的 Clawdbot 成现象级 AI 项目，上线不到 1 月 GitHub 斩获 57.6K+ stars、6.9K+ forks，被誉 "开源版 贾维斯"，特斯拉前 AI 主管 Karpathy公开点赞，重新定义个人与企业级 AI 助手形态。 Clawdbot 以 "本地优先、反向控制" 【浙商计算机】Clawdbot：现象级开源 AI Agent，重构人机协作与企业智能化落地 2026 年初开源的 Clawdbot 成现象级 AI 项目，上线不到 1 月 GitHub 斩获 57.6K+ stars、6.9K+ forks，被誉 "开源版 贾维斯"，特斯拉前 AI 主管 Karpathy公开点赞，重新定义个人与企业级 AI 助手形态。 Clawdbot 以 "本地优先、反向控制"为核心设计，打破传统 AI 助手场景割裂、被动响应、隐私泄露痛点，实现跨 平台无缝交互与主动式任务执行，成为 AI Agent 从 "对话"到 "干活"的关键落地载体。 轻量化、低门槛、高安全的特性，让 Cla ...

Group 1 - The core viewpoint of the news highlights the performance of the AI sector, with the CSI Artificial Intelligence Theme Index (930713) rising by 0.64% and notable gains in constituent stocks such as Tongfang Co., Ltd. (up 9.56%) and Ruixin Microelectronics (up 3.69%) [1] - The AI Artificial Intelligence ETF (512930) increased by 0.54%, with the latest price reported at 2.4 yuan [1] - The DeepSeek team released a paper titled "DeepSeek-OCR 2: Visual Causal Flow" and open-sourced the DeepSeek-OCR 2 model, which utilizes an innovative DeepEncoder V2 method to dynamically rearrange image components based on their meanings, aligning more closely with human visual encoding logic [1] - The new generation open-source model Kimi K2.5 was officially released, achieving the best results globally in several agent evaluations, including HLE, BrowseComp, and DeepSearchQA [1] Group 2 - Haitong International pointed out that the Engram model, developed in collaboration with Peking University, utilizes a hash table mechanism to decouple static knowledge retrieval from the backbone network, allowing for memory calls in O(1) time, thus enhancing deep reasoning focus [2] - Empirical evidence shows that Engram-27B systematically outperforms MoE baselines on benchmarks like MMLU and BBH, particularly excelling in long-context tasks [2] - The CSI Artificial Intelligence Theme Index comprises 50 listed companies involved in providing foundational resources, technology, and application support for artificial intelligence, reflecting the overall performance of AI-themed listed securities [2] - As of December 31, 2025, the top ten weighted stocks in the CSI Artificial Intelligence Theme Index include companies like Zhongji Xuchuang and New Yisheng, collectively accounting for 58.08% of the index [2]

新华财经上海1月28日电（记者高少华）金山办公27日在上海宣布旗下WPS 365"企业大脑"已在华东地区客户中率先落地。作为由全域知识驱动的核心引 擎，该方案通过深度赋能，实现业务效率显著提升与知识资产的高效盘活。 金山办公助理总裁朱熠锷现场分享 （受访者供图） 金山办公助理总裁朱熠锷表示，企业AI应用正加速从"以模型为中心"转向"以数据为中心"，数据质量成为决定企业AI应用效果的关键。WPS 365以知识增强 生成（Knowledge-Augmented Generation）为支撑，让大模型"掌握"企业真正的知识资产。 转自：新华财经 针对这一行业痛点，金山办公提出了知识增强生成的新模式。与传统的检索增强生成（Retrieval-Augmented Generation）仅让大模型"看到"文档不同，知识 增强生成架构系统性融合多模态、多结构的知识资产，使AI能够真正"掌握"企业知识的内在逻辑与关联。 金山办公副总裁吴庆云表示，高质量数据治理已经成为保障企业级AI应用效果的关键，企业需要优先完成非结构化数据的收集与治理。他提出"三步走"的实 施路径：首先，通过WPS 365办公软件一体化的特性将散落各处的非 ...

Core Insights - The year 2023 marked the emergence of AI, particularly with ChatGPT, leading to significant advancements in AI technology and applications [1][3] - By 2024, the focus has shifted from merely showcasing AI capabilities to practical applications and real-world utility [3][4] Industry Developments - AI models are evolving rapidly, with a shift from merely increasing model parameters to enhancing overall capabilities, including text processing, image understanding, video generation, and complex reasoning [3][5] - The competitive landscape is transforming into a comprehensive evaluation of technology, practical applications, and industry ecosystems, particularly in China, where intelligent computing power is growing rapidly [3] Future Projections - 2025 is anticipated to be a pivotal year for AI application acceleration, with a faster penetration into various industries than during the internet boom [4] - Two key technological advancements are driving this change: multimodal integration, allowing AI to understand text, images, and sounds, and tool invocation capabilities, enabling AI to perform tasks autonomously [5][6] Commercial Impact - AI is becoming a new engine for corporate growth, as evidenced by financial reports from Q3 2025 showing significant revenue increases across various sectors [7] - In the office sector, Kingsoft Office's WPS AI paid membership surpassed 7 million, leading to quarterly subscription revenue exceeding 2.1 billion yuan, with an 85% year-on-year increase in average revenue per user [8] - In enterprise services, companies like Deepin Technology and Yonyou Network reported substantial revenue growth and customer acquisition [9] - In finance, banks and investment firms are leveraging AI to enhance efficiency and user experience [10] Cost and Investment Trends - The rapid adoption of AI is largely due to decreasing costs of AI model inference, making it accessible for small and medium enterprises [11] - A capital frenzy around AI has begun, with at least six core AI companies successfully going public within a month, reflecting strong market confidence [11] Long-term Outlook - The development of AI models is expected to progress through three key phases over the next decade, with increasing general capabilities and significant transformations in manufacturing and supply chains [13] - The concept of "human-machine coexistence" is projected to become the norm, with predictions indicating that 85% of jobs will be reshaped as AI costs decrease [13][15] - Companies that adopt AI early are gaining competitive advantages, prompting a reevaluation of core competencies in a rapidly evolving technological landscape [15]