Summary of Key Points Core Viewpoint - The company held a performance briefing for the fiscal year 2024 and the first quarter of 2025, highlighting significant growth in revenue and net profit, driven by advancements in artificial intelligence and innovative solutions in various sectors [1][2]. Financial Performance - For the fiscal year 2024, the company reported a revenue of 334,431,235.63 yuan, an increase of 13.18% year-on-year - The net profit attributable to shareholders was 10,476,947.66 yuan, up 112.57% from the previous year - The net profit after deducting non-recurring gains and losses was 2,206,158.43 yuan, reflecting a growth of 102.34% [2]. Future Growth Drivers - Future profitability is expected to be driven by self-developed industry models targeting smart security, smart communities, and smart healthcare, providing specialized and lightweight solutions - The company aims to standardize its technology modules to reduce customization costs and enhance operational efficiency [3]. Industry Solutions Development - The company plans to integrate AI large models with ultra-high-definition video and XR technology to offer high-value solutions in niche markets [5]. Dividend Distribution - The company proposed a cash dividend of 1.50 yuan per 10 shares, totaling approximately 40,051,312.35 yuan (including tax), reflecting a significant increase in profit distribution compared to previous years [5]. Collaboration with Huawei - The company is a service provider for Huawei's HarmonyOS and has over 100 certified engineers, actively participating in the development of applications for the HarmonyOS ecosystem [6]. Product Applications in AI Models - The company is developing AI models for various sectors, including safety management in ports and comprehensive media content production systems, enhancing operational efficiency and media capabilities [7][8]. First Quarter Performance - In the first quarter of 2025, the company achieved a revenue of 74,470,050.49 yuan, an increase of 11.11% year-on-year, with a net profit of 1,210,720.75 yuan, up 35.34% [11]. 2025 Development Plans - The company plans to strengthen its video technology capabilities, activate AI model potential, and expand XR technology applications, focusing on digital transformation in various industries [12].

Group 1: Event Overview - The 9th World Drone Conference and International Low-altitude Economy and Unmanned Systems Expo opened in Shenzhen on May 23, showcasing advancements in drone technology and applications [1] - This year's focus is on the integration of drones with embodied intelligence technology, expanding product lines into more niche markets [1] Group 2: Innovations in Drone Technology - Over 5,000 unmanned aerial vehicles (UAVs) including helicopters, fixed-wing aircraft, multi-rotor drones, airships, and underwater drones were displayed at the expo [2] - Daotong Intelligent showcased its industrial-grade drones and highlighted the application of multi-modal large models, aiming to transform drones into intelligent robots capable of autonomous operation and path optimization [2] Group 3: Emergency Rescue Applications - The conference featured a variety of emergency rescue applications, including drone firefighting units, smart remote-controlled rescue stretchers, and flying life-saving devices [3] - A notable innovation is a flying life-saving ring developed by Zhejiang Juxing Power Technology, which can be remotely controlled to drop near drowning victims, enhancing rescue efficiency [3] - Longyi Aviation presented a drone firefighting unit capable of rapid deployment, with the ability to cover an area of 500 square meters in three minutes, significantly improving firefighting efficiency [3]

Core Insights - The article discusses the development and support of the multimodal large model industry in China, highlighting various policies and initiatives at both national and local levels aimed at enhancing AI capabilities and applications [1][4][11]. Policy Development Timeline - In 2023, local policies began to emerge, focusing on computational power to encourage the development of large model technology and innovative application scenarios, starting with Guangdong, Beijing, and Shanghai. By 2024, more regions are expected to introduce relevant policies aimed at improving administrative efficiency [1]. - By 2025, government work reports will emphasize the ongoing promotion of the "Artificial Intelligence +" initiative, with a focus on supporting the widespread application of large models [1]. National Policy Summary - The Chinese government has implemented several measures to support the AI industry, particularly multimodal large models, which are seen as crucial products within the AI sector. The State Council has identified embodied intelligence as a future industry, promoting the integration of digital technology with manufacturing and market advantages [4][5]. - Key national policies include the "Guidelines for the Development of Artificial Intelligence Industry" and the "Three-Year Action Plan for Data Elements," which aim to enhance data utilization and promote high-quality economic development through data-driven initiatives [11][13]. Local Policy Highlights - Various provinces have introduced specific policies to support the development of AI large models. For instance, Guangdong aims to develop a comprehensive technology system for large models with trillion-parameter capabilities, while Beijing targets the creation of 3-5 advanced, controllable foundational model products by the end of 2025 [13][15]. - Local initiatives also include the establishment of intelligent computing centers and the promotion of AI applications in various sectors, such as manufacturing, healthcare, and urban governance [13][14]. Key Development Directions - The article outlines that provinces like Guangdong, Beijing, and Shanghai have set ambitious goals for the development of large models, focusing on creating a robust ecosystem for AI innovation and application [15]. - The emphasis is on fostering collaboration between government, industry, and academia to drive advancements in AI technologies and their practical applications across different sectors [15].

Core Insights - The article discusses the development and comparison of multimodal large models, emphasizing the integration of visual and language components to enhance understanding and generation capabilities in AI systems [1][7]. Multimodal Model Types - The mainstream approach for visual and language multimodal models involves using pre-trained large language models and image encoders, connected through a feature alignment module to enable deeper question-answer reasoning [1]. - CLIP, developed by OpenAI, utilizes a contrastive learning method to connect image and text feature representations, allowing for zero-shot classification by calculating cosine similarity between text and image embeddings [2]. - Flamingo, introduced in 2022, combines visual and language components, enabling text generation based on visual and textual inputs, and includes various datasets for training [5]. - BLIP, proposed by Salesforce in 2022, aims to unify understanding and generation capabilities for visual language tasks, enhancing model performance through self-supervised learning and addressing complex tasks like image generation and visual question answering [7]. - LLaMA integrates a visual encoder (CLIP ViT-L/14) with a language decoder, utilizing generated data for instruction fine-tuning, ensuring that visual and language tokens exist in the same feature space [8].

Core Viewpoint - The 2025 Zhangjiang Embodied Intelligence Developer Conference and International Humanoid Robot Skills Competition will take place on May 29 in Shanghai, focusing on "open source, openness, and innovation" with over 200 leading companies and 1,000 experts and developers participating [1]. Group 1: Event Overview - The conference will feature a summit, competitions, and exhibitions, aiming to build an ecosystem for humanoid and embodied robot industries [1]. - The exhibition will cover four main areas: embodied intelligence, developer ecosystem, humanoid robot industry chain, and humanoid robot bodies, showcasing the application results of humanoid robot technology [1]. Group 2: Participating Companies and Innovations - Major companies like Kepler Robotics, Zhuoyide Robotics, and Magic Atom have confirmed participation, showcasing advancements in humanoid robots [2]. - Kepler's K2 "Bumblebee" features 52 degrees of freedom, high load capacity, and long endurance, capable of lifting 30kg and operating for 8 hours on a single charge [5][6]. - Zhuoyide's "Walker II" is the world's first modular humanoid robot based on bionic tendon drive technology, with a weight of 30kg and energy consumption reduced by 25% compared to competitors [7]. Group 3: Technological Advancements - The conference highlights the role of embodied intelligence in achieving autonomous decision-making and efficient interaction with environments, with significant investments from Chinese tech companies [14]. - Innovations in core components, such as sensors and actuators, are crucial for the development of humanoid robots, with domestic companies showing competitive capabilities [18]. Group 4: Future Directions - The event aims to present a complete innovation chain and ecosystem for China's humanoid robot industry, marking a shift from being a follower to a rule-maker in the global market [27].

Core Insights - The article emphasizes the growing importance and application of multimodal large models in various industries, highlighting their clearer commercial monetization paths compared to language-only models [1] Multimodal Large Models Applications - Multimodal large models are categorized into 11 application scenarios, with the top five being digital humans, gaming, advertising, social media, and intelligent marketing, indicating a high maturity level and significant attention in these areas [1] - Digital humans leverage multimodal technology to enhance human-computer interaction through natural language processing, voice synthesis, and realistic visual presentation, improving user experience [2][5] - In gaming, multimodal large models enhance interaction by allowing characters to understand player commands and respond contextually, creating a more immersive experience [5] - The advertising industry benefits from multimodal technology by automating content creation, personalizing ad delivery, and enhancing user engagement, leading to a more efficient and intelligent advertising ecosystem [8][10] - Social media platforms are being transformed by multimodal large models, improving content creation, user recommendations, interaction experiences, community governance, and commercialization [11][12]

Investment Rating - Industry investment rating: Overweight [2] Core Views - The report highlights the rapid iteration of multimodal large models in the domestic market, indicating that computing power remains a long-term theme for the computer industry [1][4] - The supply and demand sides of computing power are both favorable, with TSMC planning to open or upgrade nine advanced manufacturing plants in 2025, with an annual budget set between $38 billion and $42 billion [4][5] - The report emphasizes the strong momentum in AI development both domestically and internationally, suggesting potential investment opportunities in related companies [6][8] Weekly Market Review - From May 12 to May 16, 2025, the Shanghai Composite Index rose by 0.76%, the ChiNext Index increased by 1.38%, and the CSI 300 Index gained 1.12%. The Hang Seng Technology Index rose by 1.95%, while the Nasdaq Index surged by 7.15% [3][26] - Sector performance showed the Media Index decreased by 0.67%, while the Hang Seng Internet Technology Index increased by 2.1%. The AI Index fell by 0.95%, and the Computer Index dropped by 1.26% [3][26] AI Developments - Tencent released the Hunyuan Image 2.0 model on May 16, 2025, achieving real-time image generation capabilities, which enhances the creative process for professional designers [4][42] - Alibaba open-sourced the Wan2.1-VACE model on May 14, 2025, which supports video generation and editing, with versions that can run on consumer-grade graphics cards [4][43] Semiconductor Sector - TSMC is accelerating the production of 2nm technology in Taiwan and has completed the second phase of its Arizona plant, with plans for further expansion [5] - AMD achieved a 39.4% revenue share in the global server CPU market in Q1 2025, marking a significant increase from previous quarters [10][43] Investment Recommendations - Focus on overseas AI companies such as Meta, Adobe, Microsoft, Apple, Nvidia, AMD, and Amazon due to their advancements in model iterations [6][8] - In the domestic AI sector, companies like Baidu, Alibaba, Tencent, and Kuaishou are highlighted for their innovative developments [9][10]

行业主要公司：阿里巴巴(09988.HK，BABA.US);百度(09888.HK，BIDU.US);腾讯(00700.HK， TCEHY);科大讯飞(002230.SZ);三六零(601360.SH);云从科技(688327.SH)等 2025年开始投融资呈爆发式增长 截至2025年4月，多模态大模型投融事件数量接近50件，其中国2021年投融资金额出现了高峰，达19.1 亿元，尽管当年投资事件数量为5件。2024年开始新一轮的投资周期，共有11件投资事件，金额达5.16 亿元。2025年前4个月，共有17件投资事件，金额为16亿元，后续多模态大模型题材的投资将呈现爆发 式增长。 投资目的地为北京 根据企业投融资目的地来看，目前行业内资金主要流向北京，占全部项目的一半。其次是深圳，占比 10%，上海占比8%。北京具有良好的互联网科技、人工智能产业发展基础，企业对于多模态大模型需求 较高，投资吸引力强。此外还有宁波、三亚、苏州三市的项目，这些地方具有较好的营商环境。 多模态大模型生成生活相关场景 智能营销、教学辅助、3D建模以及智能驾驶等应用场景是生产生活中的重要领域，也是目前多模态大 模型可以切入并且精准赋 ...

Core Insights - The rapid development of humanoid robots is driven by embodied intelligence, which is crucial for commercial viability [1] - The market for humanoid robots is projected to exceed one trillion yuan by 2045, with current market size under ten billion yuan [1] Group 1: Market Potential - Humanoid robots possess human-like perception, body structure, and movement, making them highly adaptable to various applications in manufacturing, social services, and hazardous operations [1] - According to the "Humanoid Robot Industry Development Research Report (2024)", the overall intelligence level of humanoid robots in China will remain at Level 1 from 2024 to 2028, with only a few products exploring Level 2 [1] - The evolution towards embodied intelligence is expected to break the limitations of specific scenarios and tasks, leading to comprehensive coverage across industries [1] Group 2: Technological Advancements - Multi-modal large models are key to enhancing human-robot interaction efficiency and situational understanding, with companies like NVIDIA and Tesla actively integrating multi-modal perception [2] - Reinforcement learning is anticipated to become a primary paradigm for motion algorithms, enabling efficient learning of gaits and running through reward functions [2] - The integration of pure visual solutions, six-dimensional force sensors, and electronic skin is expected to set a standard for sensory solutions, significantly improving perception sensitivity [2] Group 3: Communication and Computing - Real-time control requires efficient communication protocols and robust hardware computing power, with EtherCAT expected to become the mainstream communication protocol due to its high real-time performance and low latency [2] - As robot intelligence evolves towards embodied intelligence, the demand for edge computing power is projected to continue growing, driving performance upgrades in edge-side chips [2]

Core Insights - ByteDance has launched an advanced visual-language multimodal model, Seed 1.5-VL, showcasing significant improvements in multimodal understanding and reasoning capabilities [1][2][3]. Group 1: Model Features - Seed 1.5-VL demonstrates enhanced visual localization and reasoning, with the ability to quickly and accurately identify various elements in images and videos [3][4]. - The model can process a single image and a prompt to identify and classify multiple objects, providing precise coordinates [4]. - It can analyze video footage to answer specific questions, showcasing its advanced video understanding capabilities [5]. Group 2: Performance Metrics - Despite having only 20 billion activation parameters, Seed 1.5-VL performs comparably to Gemini 2.5 Pro, achieving state-of-the-art results in 38 out of 60 public evaluation benchmarks [6]. - The inference cost is competitive, with input priced at 0.003 yuan per 1,000 tokens and output at 0.009 yuan per 1,000 tokens [7]. Group 3: Practical Applications - Developers can access Seed 1.5-VL through an API, enabling the creation of AI visual assistants, inspection systems, and interactive agents [7]. - The model's capabilities extend to complex tasks such as identifying emotions in images and solving visual puzzles, demonstrating its versatility [17][20]. Group 4: Technical Architecture - Seed 1.5-VL consists of three core components: a visual encoding module (SeedViT), a multi-layer perceptron (MLP) adapter, and a large language model (Seed1.5-LLM) [27]. - The model has undergone a unique training process, including multi-modal pre-training and reinforcement learning strategies, enhancing its performance while reducing inference costs [29][30]. Group 5: Industry Impact - The advancements presented at the Shanghai event indicate that ByteDance is building a comprehensive AI ecosystem, integrating various technologies from video generation to deep visual understanding [32]. - The emergence of Seed 1.5-VL signifies a step towards a true multimodal intelligent era, reshaping interactions with visual data [32][33].