Market Overview - The market continues to follow the rebound of US stocks, with Hong Kong stocks jumping above 27,000 points but later experiencing volatility, closing up 0.58% [1] - The competition among major companies in large model development is intensifying, transitioning from a "hundred model war" to a focus on commercialization, innovation, and global layout, with the number of capable developers reduced from over 200 to less than 10 [1] AI Model Developments - ByteDance's image generation model Seedream 5.0 has launched on various platforms, enhancing image generation capabilities and allowing precise adjustments [2] - Alibaba has reportedly integrated its new model Qwen 3.5 into the HuggingFace open-source project, indicating an imminent release [2] - The launch of Seedance 2.0 has been highly praised, marking a significant advancement in video generation technology, leading to concerns over privacy and copyright issues [1][2] Company Performance and Strategies - The CEO of Yueda Group has outlined three core strategic directions for the year: "evergreen content, IP+AI, and globalization," which are expected to benefit from the advancements in AI models [2] - Xunce (03317) is positioned as a leader in real-time data infrastructure and analysis solutions, with expectations of being included in the Hong Kong Stock Connect list, leading to a stock increase of over 15% [4] Financial Highlights - Semiconductor company SMIC reported a 2025 Q4 revenue of 17.813 billion yuan, a year-on-year increase of 11.9%, with a net profit of 1.223 billion yuan, up 23.2% [8] - The company anticipates a revenue growth rate higher than the industry average for 2026, with a capital expenditure plan similar to 2025 [8] Industry Trends - The demand for AI-driven applications is driving interest in related stocks, with companies like Xunce and others in the AI data analysis sector gaining attention [3] - The shipping sector is also showing strength, with companies like Pacific Shipping and COSCO Shipping Energy experiencing stock increases due to rising demand for dry bulk shipping [6] Robotics and AI Innovations - Alibaba's Damo Academy has released the RynnBrain model, significantly enhancing robotic capabilities, which is expected to attract investment in the robotics sector [7] - The launch of the Gino1 robot by Geekplus is anticipated to have a broad application in China's logistics sector, contributing to a stock increase of over 11% [6]

Group 1 - The core point of the article is that UBTECH's logistics subsidiary, Youqi, has successfully completed the first batch of PVT vehicles for its autonomous logistics vehicle, Chitu α, at Foxconn's new energy vehicle R&D center in Zhengzhou [1] - The PVT phase is a critical verification stage for vehicle mass production delivery capability, focusing on manufacturing consistency, process stability, and quality traceability under mass production conditions [1] - The production process at Foxconn's automotive-grade intelligent manufacturing system covers the complete closed loop of vehicle assembly, testing, and delivery, laying a solid foundation for the next phase of Chitu α's mass production ramp-up and bulk delivery [1]

Summary of the Conference Call for UBTECH Company Overview - **Company**: UBTECH - **Technology**: Development of the Thinker large model for humanoid robots, showcasing advanced capabilities in task understanding, environmental perception, cognitive decision-making, and task planning [2][4] Key Points Industry and Technology Advancements 1. **Performance in International Evaluations**: The Thinker model has achieved 9 global firsts in international evaluations, demonstrating its technological leadership [2][4] 2. **Robust Execution in Industrial Applications**: The humanoid robots exhibit high efficiency and precision in tasks such as handling and sorting, aided by the integration of the Thinker model with Visual Object Analysis (VOA) technology [2][4][5] 3. **Self-Correction Capabilities**: The robots can autonomously adjust during operations, showcasing strong generalization abilities even in unforeseen circumstances [5][10] Model Development and Data Handling 1. **Data-Centric Approach**: The development has shifted from model-centric to data-centric, achieving nearly 100% automation in data annotation processes, emphasizing the importance of high-quality data [3][6] 2. **Diverse Data Sources**: The training data primarily consists of open-source data, supplemented by real machine and customer site data, effectively addressing hardware heterogeneity issues [5][6] 3. **Cost Control**: Utilizing open-source data significantly reduces research and development costs [5][6] Engineering and Optimization 1. **Efficient Model Deployment**: The Thinker model operates efficiently on limited computational resources through techniques like model distillation and quantization [5][6] 2. **Performance Improvement Factors**: Key factors include multi-modal data fusion, advanced algorithms, and engineering optimizations that enhance stability and reliability in real-world applications [6][10] Future Directions and Goals 1. **Upcoming Model Updates**: In the next 3-6 months, UBTECH aims to explore visual language action (VLA) and world model directions, with plans to release related results soon [7][10] 2. **Adaptability for Specific Models**: The current model serves as a foundational model, which will be fine-tuned for specific robot models like Work S2 [7][10] Collaboration and Open Source 1. **Support for Developers**: The open-source model allows developers to perform secondary fine-tuning and includes tools for hardware optimization, facilitating efficient deployment on UBTECH robots [9][10] 2. **Data Security Measures**: UBTECH ensures that sensitive information in training data is protected and complies with cross-border data transmission policies [9][10] Unique Selling Propositions 1. **Compact Model Size**: The Thinker model is notably small (4B), enhancing deployment efficiency compared to competitors [11][12] 2. **Real-World Data Utilization**: The reliance on real industrial scenario data increases the model's reliability and applicability in practical environments [11][12] Challenges and Solutions 1. **Long-Tail Problem Management**: UBTECH addresses the long-tail problem by focusing on failure cases, collecting and annotating them for model iteration, thus creating a data feedback loop to enhance overall performance [10][12] Conclusion UBTECH's advancements in humanoid robotics through the Thinker model highlight its commitment to innovation, efficiency, and practical application in industrial settings, positioning the company favorably within the robotics industry.

Core Insights - Zhuoshijia Technology (Hainan) Co., Ltd. has raised several hundred million RMB in Pre-IPO financing, with participation from institutions such as Guotai Junan Venture Capital, UBTECH Technology, Guoxin Guozheng Investment, and Zhejiang Huayu [1] Company Overview - Zhuoshijia Technology is a leading artificial intelligence technology innovation enterprise in China, recognized for its strong technological and research capabilities [1] - Since its establishment in 2019, the company has adhered to the philosophy of "AI for All," focusing on AI-to-B (AI empowering industry digitization) [1] - The company specializes in the development and application of industry-specific large models, providing comprehensive solutions across the entire chain, scenarios, and lifecycle [1] - Zhuoshijia Technology is considered one of the most promising leaders in the industry large model sector in China [1]

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: 深圳市優必選科技股份有限公司（於中華人民共和國注冊成立之股份有限公司） 呈交日期: 2026年2月2日 I. 法定/註冊股本變動 | 1. 股份分類 | 普通股 | 股份類別 | 其他類別 (請註明) | | 於香港聯交所上市 (註1) | | 否 | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | 證券代號 (如上市) | N/A | 說明 | 内資股 | | | | | | | | | 法定/註冊股份數目 | | 面值 | | | 法定/註冊股本 | | | 上月底結存 | | | 70,665,977 RMB | | | 1 RMB | | 70,665,977 | | 增加 / 減少 (-) | | | | | | RMB | | | | 本月底結存 | | | 70,665,977 RMB | | | 1 RMB | | 70,665,977 | | 2. 股 ...

Core Insights - Zhuoshit Technology has completed a Pre-IPO financing round amounting to several hundred million yuan, with investments from Guotai Junan Innovation Investment, UBTECH Technology, and others, aimed at enhancing R&D in industry large models and intelligent agents [1][4] - The company focuses on the "AI to B" sector, leveraging its self-developed "Xuanji Yuheng" MoE large model with hundreds of billions of parameters to provide scalable and cost-effective AI solutions across various industries [3][4] Financing and Investment - The recent financing will primarily be used to increase R&D investment in industry large models and intelligent agents, and to expand application scenarios [1][4] - The successful financing round signifies a strategic shift for the company from having product technology advantages to achieving large-scale industrial empowerment [4][5] Company Overview - Founded in 2018, Zhuoshit Technology has established itself as a leader in the AI industry, with a comprehensive technology stack that includes model platforms, data platforms, and intelligent agent platforms [3][4] - The company has received multiple recognitions, including being listed as a national high-tech enterprise and holding over a hundred AI model invention patents [4] Strategic Partnerships - Strategic partners, including Guotai Junan Innovation Investment and UBTECH Technology, have expressed high recognition of Zhuoshit Technology's technical strength and strategic value [5] - The collaboration between Zhuoshit Technology and UBTECH is seen as a perfect combination of embodied intelligence and general AI brain, crucial for the intelligentization of humanoid robots [5] Future Outlook - Zhuoshit Technology aims to accelerate the commercialization of embodied intelligence in industries such as manufacturing, leveraging its large model and intelligent agent capabilities [1][5] - The company is positioned to play a significant role in the AI industrialization wave in China, focusing on transforming AI technology into productivity across various sectors [5]

Core Viewpoint - The report from China Merchants Securities International predicts that the humanoid robot industry in China will experience significant growth by 2026 due to advancements in AI model capabilities, hardware technology maturity, cost reductions, and accelerated application scenarios, along with a surge in orders in the fourth quarter [1] Group 1: Recommended Companies - Recommended humanoid robot manufacturers include UBTECH Robotics (09880), Yujin Robot (02432), and Geek+ (02590) [1] - Cross-industry collaborative companies recommended are Tesla (TSLA.US) and Xpeng Motors (09868) [1] - Ecological companies to watch include Xiaomi Group (01810) and Midea Group (00300) [1] Group 2: Supply Chain Insights - The supply chain outlook is positive, focusing on technology convergence, with a favorable view on Minth Group (00425) in the Hong Kong stock market [1] - Companies expanding from intelligent driving to robotics, such as Hesai Technology (02525) and Horizon Robotics (09660), are also highlighted as promising [1] Group 3: Industry Trends - The humanoid robot industry in China is characterized by rapid incremental advancements and a full-chain explosion, particularly with the trend of large-scale production by overseas leaders by the end of the year [1] - Catalysts for growth include the release of Tesla's V3 and the listing of Yushu Technology, along with a significant number of new robot companies expected to go public in the Hong Kong market this year [1]

Core Viewpoint - The recent trading activities of two major stocks, Fenglong Co. and Jiamei Packaging, have been significantly influenced by their respective acquisition announcements, leading to a sharp rise in stock prices followed by a notable correction after their resumption of trading [1][2]. Group 1: Stock Performance - Fenglong Co. and Jiamei Packaging both faced a trading halt before resuming on February 2, with initial price drops followed by recovery, closing at 95.77 CNY (down 3.78%) and 22.79 CNY (down 1.64%) respectively [1]. - Prior to the halt, Fenglong Co. experienced an impressive 18 consecutive trading days of gains, with a price increase exceeding 400% over the last 20 trading days [1]. - Jiamei Packaging's stock surged by 408.11% from December 17, 2025, to January 23, 2026, driven by acquisition expectations [1]. Group 2: Acquisition Details - Fenglong Co.'s stock surge was primarily driven by its acquisition of a 43% stake in the humanoid robot company, UBTECH, for 1.665 billion CNY, making it the first A-share listed subsidiary of UBTECH [1][2]. - Jiamei Packaging's rise was linked to the anticipated acquisition by Zhuyue Hongzhi, controlled by the founder of the leading robot vacuum company, Dreame Technology, for approximately 2.28 billion CNY, aiming to gain control of Jiamei Packaging [2]. Group 3: Market Reactions and Announcements - Both companies issued announcements during their trading suspension that tempered the overly optimistic market expectations regarding their acquisitions [3]. - Fenglong Co. clarified that there are no plans to inject assets into the company within 36 months post-acquisition, warning of significant trading risks due to the stock's price surge [3]. - Jiamei Packaging explicitly stated that its core business remains in food and beverage packaging, with a projected net profit decline of 43.02% to 53.38% for 2025, indicating a significant disconnect between stock price and fundamental performance [3].

Market Review - The A-share market for artificial intelligence and robotics showed divergent trends, with related theme ETFs performing inconsistently. The Sci-Tech AI ETF (589010) experienced fluctuations, closing at 1.620 yuan, up 0.809% from the opening price. Among its 30 constituent stocks, 20 declined, with Youkede leading the drop at 5.70%. Conversely, a few stocks surged, including Sikan Technology, which hit the daily limit, and Lanke Technology, which rose over 12% [1] - The Robotics ETF (562500) initially fell but rebounded slightly, closing at 1.064 yuan, down 0.468% from the opening price. Out of 66 constituent stocks, 34 increased, with Lide Harmony leading the gain at 7.78%. However, Huachangda (300278) and Tuobang Co. (002139) fell by 6.68% and 5.08%, respectively, having limited impact on the overall market. The liquidity was robust, with a trading volume of 1.675 billion yuan and a turnover rate of 6.85% [1] Hot News - The AI-exclusive social network Moltbook has rapidly gained popularity, attracting over one million AI agents that engage in various discussions and activities, including controversial topics. This platform is associated with the recently popular open-source AI agent OpenClaw, which offers extensive functionalities for users [2] - On February 1, Tencent launched its AI application Yuanbao, initiating a 1 billion yuan cash red envelope campaign for the Spring Festival. Users can receive red envelopes and earn additional lottery chances through tasks, with the highest single reward reaching 10,000 yuan [2] - On February 1, UBTECH announced the open-sourcing of its self-developed embodied intelligent model Thinker, designed to provide rapid responses and precise spatial awareness for industrial humanoid robots [2] Institutional Views - CITIC Securities reported that with the arrival of key traffic windows like the Spring Festival, AI applications are expected to see significant operational catalysts. Tencent's Yuanbao red envelope activity is anticipated to enhance user engagement and frequency of use, potentially accelerating the penetration and commercialization of AI products [3] Popular ETFs - The Robotics ETF (562500) is the only ETF in the market with a scale exceeding 20 billion yuan, offering the best liquidity and comprehensive coverage of the Chinese robotics industry chain, facilitating investors' access to the sector [4] - The Sci-Tech AI ETF (589010) is described as the "brain" of robotics, with a 20% fluctuation limit and small-cap elasticity, aimed at capturing the "singularity moment" in the AI industry [4]

Group 1 - The core viewpoint of the news highlights the strong performance of the AI sector, with the CSI Artificial Intelligence Theme Index rising by 1.18% and notable gains in constituent stocks such as NewEase and 360 [1] - The AI Artificial Intelligence ETF (512930) also saw an increase of 1.14%, with the latest price reported at 2.4 yuan [1] - UBTECH announced the open-sourcing of its self-developed embodied intelligent model, Thinker, aimed at enhancing industrial humanoid robots' responsiveness and spatial awareness [1] Group 2 - The CSI Artificial Intelligence Theme Index (930713) includes 50 listed companies that provide foundational resources, technology, and application support for artificial intelligence, reflecting the overall performance of AI-related securities [2] - As of January 30, 2026, the top ten weighted stocks in the CSI Artificial Intelligence Theme Index accounted for 57.27% of the index, with companies like Zhongji Xuchuang and NewEase leading the list [2] - The AI Artificial Intelligence ETF closely tracks the CSI Artificial Intelligence Theme Index and has several connection funds available for investors [2] Group 3 - Guojin Securities pointed out that the DeepSeek Engram architecture may further promote the concept of "computing with storage," effectively decoupling computation and memory in large models [1] - The Engram architecture allows for the computation of all Transformer operators on GPU/accelerator cards while running a 100 billion parameter Engram table in CPU memory, which could enhance the demand for storage-based computing [1] - Anthropic's Claude Cowork introduces a new permanent memory method designed for Claude, which could also help overcome GPU memory limitations [1]