Core Viewpoint - DeepSeek, a Chinese AI company, is reportedly developing a more advanced AI model with enhanced agent capabilities to compete with industry leaders like OpenAI [1][2] Company Developments - DeepSeek's founder, Liang Wenfeng, is pushing the team to release new software in Q4 of this year [1] - The company previously launched the R1 series model in January, which mimics human reasoning processes at a development cost of only a few million dollars [3] - An upgraded version, DeepSeek-R1-0528, was released on May 28, showing significant improvements in reasoning ability, programming performance, and text processing [3] - The DeepSeek-V3.1 model was officially released on August 21, enhancing code understanding and agent task execution capabilities [3][4] Market Trends - AI concept stocks have recently experienced a downturn, with the DeepSeek index, AI computing power index, and AIGC index dropping by 8.94%, 9.25%, and 7.16% respectively over a few trading days [1] - Analysts attribute this decline to profit-taking, but maintain a positive long-term outlook for the AI industry [1] Industry Insights - The AI agent sector is viewed as a key area of growth following large models, with capabilities for autonomous decision-making and task execution [5] - A government policy issued on August 26 aims to integrate AI into six key sectors by 2027, with a target of over 70% penetration of new intelligent terminals and agents [5] - The market for AI agents is expected to see significant growth, with predictions of a commercial explosion in 2025 [6][7]

Core Insights - The article discusses the current limitations of AI agents and emphasizes the importance of a coordination layer to enhance task execution reliability and cost control [2][3] - Temporal, a company focused on Durable Execution, has gained attention for its ability to ensure reliable workflow execution even in the face of failures [3][6] - Temporal has completed a $146 million Series C funding round, achieving a valuation of $1.72 billion, with notable clients including Nvidia and OpenAI [3][8] Group 1: What is Temporal? - Temporal is an AI infrastructure company founded in 2019, focusing on Durable Execution to ensure reliable workflow execution despite failures [6] - The company has over 2,500 clients, including major firms like Nvidia, Airbnb, and Netflix, with a Net Dollar Retention (NDR) rate of 184% [8] Group 2: Product Architecture - Developers can write business logic in workflow functions, while Temporal guarantees reliable and persistent execution [11] - Temporal uses Event Sourcing to automatically recover workflow states, ensuring execution can continue from the point of failure [11][16] - The architecture allows for asynchronous task execution through a Task Queue, enhancing system stability and simplifying development [16][17] Group 3: Use Cases - Temporal is utilized in various scenarios, including infrastructure orchestration, application deployment, and data processing, demonstrating its versatility [18][19][20] - Specific examples include Uber's machine deployment coordination and Coinbase's transaction reliability in fintech [19][20] Group 4: Open Source and Commercialization - Temporal offers both an open-source version and a managed cloud service, allowing users to switch between deployment modes seamlessly [21][22] - The open-source version is designed to be fully functional, with a focus on maintaining customer trust and avoiding vendor lock-in [24][25] Group 5: Durable Execution - Durable Execution allows developers to manage distributed tasks without worrying about system crashes, as the execution state is persistently stored [34][35] - The system provides runtime visibility, enabling developers to track interactions and quickly identify issues [35][37] Group 6: Future Directions - Future developments for Durable Execution may include the integration of WebAssembly for enhanced performance and the evolution of RPC protocols to support long-running operations [37][39] - Temporal aims to become a core component in the ecosystem of tool invocation, particularly in cross-company interactions [39]

Who’s making an AI Agent trained on Pokémon data? ...

Core Viewpoint - Minglue Technology has submitted a new application for listing on the Hong Kong Stock Exchange, aiming to issue up to 8,301,800 shares, with the backing of major investors like Tencent and Kuaishou [1][3][11] Company Overview - Established in 2006, Minglue Technology has nearly 20 years of experience in the digital marketing sector and has served 135 Fortune Global 500 clients [4][6] - The company is recognized as China's largest data intelligence application software supplier based on 2023 revenue [3][6] Business Performance - From 2021 to 2023, the Chinese data intelligence application software market grew at a CAGR of 20.4%, with Minglue's revenue expanding to 1.462 billion yuan in 2023, marking a growth of 15.2% [6][7] - Despite the growth, the company faces challenges with declining revenues in its marketing intelligence segment, which fell from 750 million yuan in 2023 to 730 million yuan in 2024 [8][9] Technology and Innovation - Minglue holds 2,177 patents and has received over 450 domestic and international awards, focusing on multi-modal data intelligence, enterprise knowledge graphs, and data privacy technologies [5][6] - The company has launched AI-driven products like AdEff, which utilizes its proprietary multi-modal large language model to optimize advertising effectiveness [12] Market Position and Challenges - The company has experienced a significant valuation drop from over $30 billion in 2020 to approximately $15 billion in early 2024, reflecting pressures from market competition and changing customer demands [11] - The operational efficiency of its industry AI solutions remains low, with revenue contribution consistently below 10% and negative gross margins [9][10] Investment Appeal - Minglue's strong investor backing, including major players in the AI sector, enhances its attractiveness in the market, despite current operational challenges [11][13] - The anticipated growth of the AI Agent market, projected to reach $50.31 billion by 2030, presents potential opportunities for Minglue to capitalize on [11]

Group 1 - The Hang Seng Tech Index fell by 1.13% as of September 4, 2025, with mixed performance among constituent stocks [3] - Alibaba Health led the gains with an increase of 2.35%, while Horizon Robotics saw the largest decline at 6.33% [3] - The Hang Seng Tech Index ETF (159742) decreased by 0.78%, trading at a premium with a latest price of 0.77 yuan [3] Group 2 - AI revenue has significantly improved operational performance for several leading companies, with AI applications and Agent commercialization accelerating [3][4] - The Chinese government has issued guidelines aiming for over 70% penetration of new intelligent terminals and agents by 2027, and over 90% by 2030 [3] - Alibaba has committed to investing 380 billion yuan in AI development over the next three years and is testing a new AI chip compatible with Nvidia [3] Group 3 - Kuaishou's AI revenue has exceeded 250 million yuan, while Meituan's AI revenue accounts for over 20% of its total [4] - The Hang Seng Tech Index ETF has reached a latest scale of 3.303 billion yuan and a record high of 4.285 billion shares [4] - The ETF has seen continuous net inflows over the past eight days, totaling 391 million yuan, with an average daily net inflow of approximately 48.84 million yuan [4] Group 4 - The top ten weighted stocks in the Hang Seng Tech Index account for 68.71% of the index, including Tencent, Alibaba, and Netease [4]

Group 1 - The establishment of a Tech communication group focused on large models, inviting participants to discuss topics such as RAG, AI Agents, multimodal large models, and deployment of large models [1] - Interested individuals can join the group by adding a designated WeChat assistant and providing their nickname along with a request to join the large model discussion group [2]

【深圳商报讯】（首席记者谢惠茜）近日，AIGC软件A股上市公司万兴科技接受摩根士丹利、中金公 司等49家知名机构调研。万兴科技董事长吴太兵率高管团队出席，就公司2025年上半年在产品、研发及 经营管理等方面取得的成果，向投资者与分析师进行介绍，并对未来发展战略与前景作出展望。 吴太兵指出，未来10年是AI技术与产业深度融合的黄金窗口，AI大模型是新的生产力工具，将推动创 意软件产业全面升级，由功能被动调用转向"主动协作"；Agent正推动内容创作从"素材+编辑"的手工作 坊模式，进入到"意图+生成"的自动化时代。万兴科技未来将聚焦三大方向：一是坚持"双模驱动"战 略，持续布局多模态、智能体等下一代先进技术，加快万兴天幕大模型的迭代，并以AI Agent拓宽用户 边界，推动"创意平权"；二是坚持多端矩阵化产品战略，全线推进产品移动化和AI化，孵化AI新品；三 是推进AI应用生态共建，持续与全球AI应用产业开放合作，加强软硬件产业联动，共同提升AI创新与 用户体验。 据统计，近1年来，万兴科技共接待168家机构调研，其AI布局受到市场关注。万兴科技上半年实现营 收7.6亿元，同比增长7.77%。 公开资料显示，万 ...

Summary of Key Points from the Conference Call Industry Overview - The AI Agent industry is experiencing increasing application penetration, with core components including user input, environmental sensors, control centers, and perception effectors, enabling autonomous decision-making and actions [1][3] - The global market for AI Agents is projected to reach $52.6 billion by 2030, with significant growth expected in high-frequency, structured data scenarios [1][7] Core Insights and Arguments - The AI industry has shown strong performance in the secondary market, particularly in AI computing themes such as Chiplet Packaging Optimization and liquid cooling [2] - AI Agents are categorized into general agents and industry-specific agents, with applications across gaming, healthcare, marketing, education, finance, and office environments [7] - In the gaming sector, AI is primarily used to enhance development efficiency, with 80% of surveyed developers reporting efficiency improvements of over 20% [8] - In healthcare, AI applications include medical imaging analysis and electronic medical record management, with the AI healthcare market in China expected to grow from 97.3 billion RMB in 2023 to 159.8 billion RMB by 2028 [9] - Marketing challenges such as customer insight and personalized marketing are being addressed through AI-driven data integration and personalized recommendations [10] Important but Overlooked Content - The current AI Agent architecture is based on large language models (LAM) and features a modular design that integrates perception, action, and core modules [6] - The three modes of human-AI collaboration are embedding, copilot, and agent modes, with the agent mode allowing AI to perform most tasks while humans set goals and evaluate outcomes [5] - Recent market dynamics indicate a high level of activity in the A-share market, with significant capital flow into consumer electronics and AI-related concepts, and a record high margin balance of 2.24 trillion RMB [16] - The Chinese government has issued policies to promote AI development, aiming for over 90% penetration of new intelligent terminals by 2030 [17][19]

Investment Rating - The investment rating for the company is "Buy" (maintained) [4] Core Views - The company's performance is stable and meets expectations, with a revenue increase of 16% and a net profit increase of 26% [1] - The integration of AI products and new categories is driving revenue growth, with significant contributions from the AI business [1][3] - The company is expected to benefit from AI-driven efficiencies and scale effects, with projected revenue growth of 14% to 10% from 2025 to 2027 [1][8] Financial Performance - For H1 2025, the company reported revenue of 915 million yuan, a year-on-year increase of 15.9%, and a net profit of 295 million yuan, up 26.1% [1] - In Q2 2025, revenue reached 476 million yuan, with a year-on-year growth of 16.3%, and net profit of 183 million yuan, up 16.4% [1] - The company's gross margin was 80.1%, with a net margin of 32.2%, reflecting a year-on-year increase of 3.5 percentage points [1] Business Segments - Revenue from the China Manufacturing Network was 760 million yuan, a year-on-year increase of 17%, while AI business revenue was 21.74 million yuan, up 92% [1] - The company has launched the "New Maritime Plan" to enhance online promotion in the light industry sector, attracting more sellers [1][2] Membership and AI Integration - As of H1 2025, the number of paid members on the China Manufacturing Network platform reached 28,699, an increase of 581 from the end of Q1 2025 [2] - The AI business has seen a membership increase to over 13,000, with a penetration rate of 45%, up 6 percentage points from the previous quarter [3] Earnings Forecast and Valuation - The company is projected to achieve revenues of 1.9 billion yuan in 2025, 2.1 billion yuan in 2026, and 2.3 billion yuan in 2027, with corresponding net profits of 544 million yuan, 643 million yuan, and 750 million yuan [8] - The price-to-earnings ratio is expected to decrease from 29 in 2025 to 21 in 2027, indicating a favorable valuation trend [1][8]

Investment Rating - The industry investment rating is "Outperform" [1] Core Insights - The report highlights the rapid development of large models and the acceleration of multimodal applications, with closed-source models gradually regaining an advantage over open-source ones [2][4] Summary by Sections Large Models - The development of multimodal technology is still evolving, with both domestic and international major players continuously refreshing state-of-the-art (SOAT) benchmarks [4][11] - Closed-source models are beginning to maintain performance advantages after the open-source wave, with significant advancements in video and image generation capabilities [4][19] Computing Power - Capital expenditures (Capex) from major cloud service providers (CSPs) are increasing, with chip manufacturers accelerating the release of new versions of chips [29][32] - Major CSPs like Microsoft, Google, Meta, and Amazon have raised their Capex guidance, indicating a strong focus on AI infrastructure [32][35] Applications - There is a noticeable acceleration in application differentiation overseas, particularly in complex B2B scenarios that integrate with data [47][64] - Companies that effectively leverage AI to enhance their core business operations are seeing significant revenue growth, particularly in the B2B sector [47][56]