扫码可订阅产业日报 1.【美团发布原生多模态 LongCat-Next】3月27日消息，美团发布并全面开源原生多模态大模型 LongCat-Next及其核心组件——离散原生分辨率视觉分词器（dNaViT）。该模型打破了当前大模型 以"语言为中心"的传统拼凑式架构，将图像、语音与文本统一映射为同源的离散Token。通过纯粹 的"下一个Token预测"（Next Token Prediction，NTP）范式，LongCat-Next让视觉与语音成为 AI的"原生母语"。（界面新闻） 2.【ChatGPT广告业务试点项目六周内年化收入超1亿美元，拟将试点范围扩大至更多国家】据报 道，OpenAI的ChatGPT广告业务在美国试点项目推出后，六周内已达到年化收入1亿美元的里程碑。 OpenAI发言人表示，这得益于公司新兴广告业务的强劲早期需求。OpenAI计划在未来几周内将试点 范围扩大到更多的国家，包括澳大利亚、新西兰和加拿大。（财联社） 3.【Anthropic据悉考虑最早10月进行IPO】知情人士透露，Anthropic PBC考虑最早10月上市，在 首次公开募股(IPO)方面与竞争对手OpenAI Inc. ...

Core Viewpoint - Alibaba has established the Alibaba Token Hub (ATH) to focus on "creating, delivering, and applying tokens," positioning itself as the first tech giant organized around the concept of tokens, aiming to reshape its AI business landscape [2][3]. Group 1: ATH Establishment and Structure - ATH is led by CEO Wu Yongming and includes several divisions such as Tongyi Lab, MaaS, Qianwen, Wukong, and AI Innovation, covering a comprehensive layout from foundational model development to AI applications for individuals and enterprises [2][3]. - The establishment of ATH is seen as a strategic response to the evolving AI landscape, similar to Google's consolidation of its AI teams to compete effectively against emerging technologies like ChatGPT [2][3]. Group 2: Token Economy and AI Agents - The rise of AI Agents is expected to lead to significant token consumption, creating opportunities for cloud service providers as these agents will rely on tokens for operation, becoming the primary interface between humans and the digital world [3][11]. - The demand for tokens is projected to surge, with a report indicating that the daily usage of large models in Chinese enterprises skyrocketed to 37 trillion tokens in the second half of 2025, a 263% increase from the first half [5][6]. Group 3: Product Development and Market Position - Alibaba is set to launch a new enterprise-level AI Agent application developed by the DingTalk team, which aims to surpass existing solutions like OpenClaw in capabilities [4]. - Recent product launches include various AI Agent tools, with the Coding Plan experiencing explosive growth in token usage, becoming Alibaba Cloud's fastest-growing product [4][5]. Group 4: Strategic Implications and Future Outlook - The Wukong division is positioned as a key player in monetizing Alibaba's B2B services, integrating large model capabilities into enterprise workflows and replacing traditional digital tools with AI Agents [10]. - The integration of ATH is expected to enhance operational efficiency and clarity of goals, potentially leading to a significant breakthrough akin to Google's Gemini project [12][13]. - Alibaba's cloud market share has shown consistent growth, reaching 36% in the first three quarters of 2025, indicating a strong competitive position in the Chinese cloud market [13].

Investment Rating - The report maintains a standard rating for the electronic sector, highlighting structural opportunities despite challenges in demand and pricing pressures [5]. Core Insights - The OpenClaw initiative is driving a surge in AI Agent applications, leading to increased demand for reasoning tokens and subsequently boosting the computing power across the AI industry chain. Major domestic CSPs have launched related Agent products [5]. - AI demand is projected to drive a 26.3% year-on-year growth in global wafer foundry output value in 2025, reaching a historical high of $169.5 billion. However, high storage prices may impact terminal demand and wafer fab capacity utilization in 2026 [5]. - The electronic sector underperformed the market this week, with the Shenwan Electronics Index declining by 1.23% compared to a 0.19% increase in the CSI 300 Index [5]. - Investment suggestions include focusing on companies benefiting from strong domestic and international AIOT demand, AI innovation-driven sectors, and upstream supply chain opportunities related to semiconductor equipment and materials [6]. Summary by Sections Industry Overview - The report emphasizes the transformative impact of OpenClaw on AI Agent applications, which are set to redefine human-computer interaction and necessitate upgrades in security systems [5]. - The global wafer foundry market is expected to see significant growth driven by AI server GPU and TPU demand, with TSMC maintaining a dominant market share [5]. Market Performance - The electronic sector's performance this week was below the market average, with various sub-sectors showing mixed results, particularly in semiconductors and consumer electronics [20][22]. Investment Recommendations - Suggested companies include those in the AIOT sector such as Lexin Technology and Hanguang Technology, as well as firms in the AI innovation space like Cambricon and Moore Threads [6].

Investment Rating - The report maintains a standard rating for the electronic sector, indicating a cautious but optimistic outlook on investment opportunities driven by AI demand and structural changes in the industry [5]. Core Insights - The OpenClaw initiative is driving a surge in AI Agent applications, leading to increased demand for reasoning tokens and subsequently boosting the computing power across the AI industry chain. Major domestic CSPs have launched related Agent products to capitalize on this trend [5]. - AI demand is projected to drive a 26.3% year-on-year growth in global wafer foundry output value in 2025, reaching a historical high of $169.5 billion. However, high storage prices may impact terminal demand and wafer factory capacity utilization in 2026 [5]. - The electronic sector underperformed the market this week, with the Shenwan Electronics Index declining by 1.23%, while the CSI 300 Index rose by 0.19% [20]. - Despite strong AI-driven demand, high storage prices are significantly suppressing demand, and there may be a temporary cooling in AI investment trends. The report suggests focusing on structural opportunities in the market [5][6]. Summary by Sections Industry Overview - The report highlights that OpenClaw, an AI Agent platform, is set to redefine human-computer interaction and drive significant growth in the AI industry chain due to its high demand for reasoning tokens [5]. - The global wafer foundry market is expected to see substantial growth, with TSMC maintaining a dominant market share of 70% and advanced process nodes experiencing high demand due to AI server GPUs and mobile SoC production [5]. Market Performance - The electronic sector's performance this week was below the market average, with various sub-sectors showing mixed results. Semiconductor stocks fell by 2.60%, while electronic components rose by 1.44% [20][22]. - The report notes that the overall PE ratio for the electronic sector stands at 68.37, indicating a relatively high valuation compared to historical averages [24]. Investment Recommendations - The report suggests focusing on companies benefiting from strong domestic and international AIOT demand, such as Lexin Technology, Hengxuan Technology, and others in the AI innovation-driven sector [6]. - It also recommends monitoring companies involved in semiconductor equipment and materials, as well as those positioned for recovery in pricing, such as new energy and power components [6].

Group 1 - The article discusses the recent public testing of Tabbit, an AI browser developed by Meituan, and its implications for investment in the Hang Seng Technology Index [3][4][7] - Tabbit's design is compared to other AI agents, highlighting its limitations in task execution and user interaction, which may affect its competitiveness in the market [13][14][20] - The article raises concerns about the potential bias in information sources used by AI agents like Tabbit, which could lead to cognitive biases for users [20] Group 2 - The competitive landscape for AI browsers is intensifying, with companies like OpenAI and Alibaba also launching similar products, indicating a race for market share [28][29] - Meituan's strategic decision to invest in an AI browser team, despite it being unrelated to its core business, is seen as a way to explore new opportunities and mitigate risks [30][32] - The article emphasizes the importance of establishing a foothold in the rapidly evolving AI landscape, suggesting that frequent innovation can positively impact Meituan's stock performance [33]

| 信報: | | | --- | --- | | ◼ | 內地 GenAI 用戶逾 6 億 普及率 43% | | ◼ | 貝桑何立峰擬本月中旬會晤 | | 經濟日報: | | | ◼ | 阿里雲推 QoderWork 文字指令控 AI 任務 | | ◼ | 傳 xAI 擬還債 175 億美元 分析：利 SpaceX 上市 | | 明報: | | | ◼ | 亞股下瀉拖累 恒指挫近 300 點 | | ◼ | 韓股挫 7%年半最傷 三星 SK 海力士跌一成 | | 星島日報: | | | ◼ | 亞股續捱沽韓股單日暴瀉逾 7% | | 證券日報: | | | ◼ | 人行：1 月末債券市場托管余額 197.7 萬億元 | | 即日港股重點消息 | | --- | 港股昨日回顧 | 香港指數 | 上日收市(點) | 日變化(%) | 年初至今(%) | | --- | --- | --- | --- | | 恒生指數 | 25768.08 | -1.12 | 0.54 | | 國企指數 | 8608.71 | -1.07 | -3.42 | | 科技指數 | 4876.53 | -2.26 | `1 -11. ...

Core Insights - The article highlights the rapid development of artificial intelligence (AI) technology, presenting significant opportunities in the market [3]. Group 1: AI Applications and Developments - Alibaba's QoderWork has been fully launched, providing a dedicated AI assistant for Mac and Windows users, enabling tasks like document organization and data processing through natural language interaction [5]. - The AI product rankings show that Qianwen has reached 203 million monthly active users (MAU), marking a 552% increase, making it the fastest-growing AI application globally [6]. - Meituan's new Tabbit AI browser has entered public testing, designed for professionals and students, offering features like automatic webpage opening and data extraction [8]. Group 2: Market Dynamics and Competition - Following a controversy involving OpenAI, many users have switched to Anthropic's Claude, which has topped the U.S. app download charts, while OpenAI's ChatGPT and Google's Gemini follow in second and third place [10]. - Apple has launched the iPhone 17e and a faster iPad Air, with the iPhone priced at $599 and the iPad Air at $799, indicating ongoing competition in the consumer electronics market [11]. - Apple is exploring a partnership with Google to enhance Siri using Google's AI models and cloud services, which may improve Siri's capabilities while adhering to strict privacy standards [12].

Core Insights - The competition in the AI agent sector has intensified with the launch of products like Claude Cowork and OpenClaw, indicating a potential shift in how users interact with computers [1][2] - Domestic companies are rapidly developing their versions of desktop AI agents, suggesting a forthcoming battle for dominance in the "universal desktop agent" market [2][7] Group 1: Product Launches and Features - OpenClaw has gained significant traction, achieving over 100,000 GitHub stars and 2 million visitors within a week, becoming a phenomenon in Silicon Valley [3] - Cowork, developed by Anthropic, allows users to perform non-technical tasks easily, such as document creation and data analysis, by accessing user files with permission [5][6] - Domestic products like QoderWork from Alibaba and others are emerging to compete with Cowork, aiming to enhance productivity through user-friendly interfaces [7][8] Group 2: Market Dynamics and User Behavior - The popularity of OpenClaw has led to increased sales of the Mac mini, particularly the M4 chip version, which is seen as an ideal "AI server" due to its low power consumption and high performance [3][4] - Users are purchasing new Mac minis for running OpenClaw to mitigate data security risks associated with the software's operation [4][6] - The introduction of desktop AI agents is expected to revolutionize user interaction with computers, potentially leading to the development of new operating systems and interaction paradigms [8] Group 3: Security and Operational Differences - OpenClaw's operational model is more open, allowing users to install various software and agents, which raises concerns about data security and user control [6][7] - Cowork operates in a sandbox environment, requiring explicit user permission for file access, thus providing a more secure alternative to OpenClaw [6][7] - The emergence of AI agents is anticipated to catalyze growth in related sectors such as cybersecurity, IT operations, and data governance [7]

Core Insights - The emergence of desktop AI agents is becoming a competitive battleground for AI giants by 2026, driven by the launch of Anthropic's Claude Cowork, which redefines the potential of AI assistants [1][2] - Major Chinese AI players are responding to this trend with their own desktop agent tools, indicating a significant shift in the AI landscape [1][4] Group 1: Product Developments - Anthropic's Claude Cowork allows users to organize desktops and perform tasks like document creation and data analysis with a single command, disrupting traditional office software markets [2][3] - Chinese companies like Alibaba and MiniMax are launching similar desktop agents, such as QoderWork and MiniMax Agent, which can operate local files and software without complex setups [1][4] - The evolution of desktop agents is marked by different approaches, with Claude Cowork focusing on collaborative tasks and OpenClaw emphasizing local control and self-evolution capabilities [3][4] Group 2: Market Reactions - The introduction of Claude Cowork has led to panic selling in the software stock market, with a reported 11% decline in a basket of software service stocks tracked by Morgan Stanley [2] - OpenClaw, an open-source project, has gained popularity by demonstrating the feasibility of agents replacing repetitive tasks, further fueling interest in desktop agents [10] Group 3: User Experience and Functionality - MiniMax's desktop agent can perform various tasks such as chatting, coding, and document processing, showcasing its ability to handle complex workflows [5][6] - The desktop agents are designed to overcome limitations of web-based AI, particularly in managing large files and executing automated tasks seamlessly [5][6] Group 4: Security and Compliance Concerns - The increased capabilities of desktop agents raise concerns about data privacy and system security, as they require higher permissions to access and modify local files [2][11] - Companies are implementing strict isolation strategies and permission management to mitigate risks associated with desktop agents, ensuring user data remains secure [11][12] Group 5: Business Models and Future Directions - The business models for desktop agents are becoming clearer, with subscription-based services and defined pricing structures emerging [13] - The future of AI assistants is expected to integrate deeper into local applications, combining cloud intelligence with local execution capabilities, reflecting evolving user demands [7][9]

Core Insights - The emergence of desktop AI agents is becoming a competitive battleground for AI giants by 2026, driven by the launch of Anthropic's Claude Cowork, which redefines the potential of AI assistants and ignites global market interest in "desktop agents" [1][3][22] Group 1: Market Developments - Anthropic's Claude Cowork, launched in January 2026, can organize desktops and create documents based on user instructions, significantly impacting the traditional office software market [3][4] - Following Claude Cowork's release, major Chinese AI players like Alibaba and MiniMax have announced similar desktop agent tools, such as QoderWork and MiniMax Agent, respectively [1][9] - The market reaction to these developments has been notable, with a reported 11% decline in a basket of software service stocks tracked by Morgan Stanley after the launch of Claude Cowork [5] Group 2: Product Features and Comparisons - Claude Cowork focuses on collaborative office tasks and workflow understanding, while OpenClaw emphasizes local control and self-evolution capabilities [8][15] - MiniMax's desktop agent aims to evolve AI from a mere conversationalist to an active executor of tasks, capable of handling complex workflows and long-term memory accumulation [10][17] - The desktop agents differ from traditional AI assistants, which rely heavily on cloud-based operations and have limited system permissions [15][22] Group 3: User Experience and Security - MiniMax's desktop agent has been tested for its ability to perform tasks such as organizing files and executing commands, showcasing its practical application in real-world scenarios [11][14] - Security concerns arise with desktop agents having direct access to local files and applications, prompting discussions on data privacy and system security [3][19] - Companies are implementing strict isolation strategies and user confirmation mechanisms to mitigate risks associated with desktop agents [19][22] Group 4: Future Outlook - The rapid development of desktop agents is expected to continue, with predictions that AI assistants will increasingly integrate into local applications, forming a "cloud brain + local hands" model [15][22] - The balance between functionality, user experience, and security will be crucial as desktop agents evolve and become more embedded in workflows [22]