21世纪经济报道记者骆轶琪 就在豆包手机助手引发热议并在部分场景遭遇使用阻力后不久，大模型厂商智谱华章选择开源其AI Agent模型AutoGLM，后者也被称为是全球首个具备"Phone Use"（手机操作）能力的AI Agent。 从展示的功能来看，AutoGLM的自动化操作能力与豆包手机助手功能十分相似，其实也与手机厂商近两年来展示的AI功能demo 颇为类似。 根据智谱的说法，此举初衷在于，把这种智能化能力变成整个行业可以共同拥有、共同打磨的公共底座；同时让数据与隐私， 永远留在使用方这一侧。 当然挑战也存在。虽然多个品牌此前展示过用AI手机自主操作发红包、买外卖等能力，但到今天具体手机的使用层面，这种操 作其实还无法完好落地。 有手机行业从业者对记者分析，这背后源于"国内在用户隐私和数据权限边界方面存在标准空白。"这也是当前手机行业在积极 推进完善的工作。 这显示出，在AI手机势不可挡的当下，无论技术还是制度层面，都还亟待"补缺"，需要手机、软件、大模型等多方角色站在一 起，共同探讨推进。 Agent加速普及 12月9日，智谱华章宣布开源AutoGLM，这被行业定义为"全球首个具备Phone Use能 ...

Core Insights - The AI industry is entering a rapid change cycle, with 2025 being a pivotal year for the development of large models, particularly with the emergence of DeepSeek, which is reshaping the global landscape and promoting open-source initiatives [1][10][18] - The dual-core driving force of AI development is characterized by the United States and China, each following distinct paths, with key technologies accelerating towards engineering applications [1][10][11] - Despite advancements in model capabilities, challenges in real-world application remain prevalent, indicating a shift in focus from "large models" to "AI+" [1][10][19] Group 1: Global Large Model Landscape - The global large model development is driven by a dual-core approach, with the U.S. leading in closed-source models and China focusing on open-source models [10][11][13] - OpenAI, Anthropic, and Google represent the leading trio in the large model arena, each adopting differentiated strategic paths [17] - DeepSeek's emergence marks a significant breakthrough for China's large model development, showcasing the potential of open-source models [18][19] Group 2: Key Technological Evolution - The evolution of large models is marked by four major technological trends: native multimodal integration, reasoning capabilities, long context memory, and agentic AI [22][24] - Native multimodal architectures are replacing text-centric models, allowing for seamless integration of various modalities [23] - Reasoning capabilities are becoming a core feature of advanced models, enabling them to demonstrate their thought processes [24][26] Group 3: Industry Chain and Infrastructure - The AI infrastructure is still dominated by Nvidia, with a slow transition towards a multi-polar ecosystem despite the emergence of alternatives like Google’s TPU and AMD’s chips [47][48] - The AI industry is shifting from reliance on a few cloud providers to a more collaborative funding model, with Nvidia and OpenAI acting as dual cores driving the ecosystem [51][52] Group 4: Application Layer Opportunities - Large model companies are positioning themselves as "super assistants" while also aiming to control user entry points through various products and services [53][54] - Independent application companies can find opportunities in vertical markets that require deep industry understanding and complex workflow integration [55][56] - The evolution of AI applications is moving towards intelligent agents capable of autonomous operation, indicating a significant shift in application development paradigms [61][62]

Core Insights - The future of mobile technology is predicted to see the disappearance of traditional smartphones and apps within the next 5-6 years, with AI assistants becoming the main interface for human-computer interaction [1] - Mobile companies are shifting their focus from cloud collaboration to edge computing, enhancing AI capabilities directly on devices [3][4] Group 1: AI Development Trends - The rise of generative AI, particularly after the success of ChatGPT, has led mobile companies to invest heavily in AI technologies, with the first AI smartphones launched in late 2023 [2] - Companies like vivo and OPPO are prioritizing edge models, with vivo introducing a 3B (30 billion parameters) edge multimodal inference model, moving away from large cloud models [3][5] Group 2: User-Centric AI - Edge models are designed to create personalized data models based on user data, ensuring privacy and efficiency, which is crucial for personalized intelligence [4][10] - The integration of AI into user habits is emphasized, with companies like vivo implementing seamless AI features that operate without user intervention [6][7] Group 3: Ecosystem Transformation - The shift towards edge computing will redefine the mobile ecosystem, with companies aiming to control user data and flow to third-party applications [7][8] - Huawei has established an AI collaboration ecosystem through its HarmonyOS, while other companies like OPPO and Honor are also developing their own AI ecosystems [8] Group 4: Challenges Ahead - The transition to edge AI faces challenges, including compatibility between cloud and mobile architectures, as well as the need for sufficient device power and memory [9][10] - The core value of edge AI lies in its ability to access private user data, which is a significant differentiator from traditional apps [10][11]

Core Insights - Shanghai aims to strengthen its smart terminal industry, targeting a total scale exceeding 300 billion yuan by 2027, with the goal of establishing over three globally influential consumer terminal brands and nurturing two leading enterprises [1] Group 1: Smart Terminal Development - The action plan includes ten initiatives focusing on artificial intelligence computers, AI smartphones, robotic terminal capabilities, smart glasses, and satellite internet terminal products [2][3] - Shanghai plans to cultivate leading brands in AI computers and establish production bases with an annual capacity of ten million units, transitioning traditional computer production to AI computers [2] - The initiative emphasizes the development of AI smartphones to fill existing gaps in Shanghai's smartphone production capabilities [3] Group 2: Robotics and Smart Glasses - The plan aims to create humanoid robots with advanced capabilities, supporting R&D and mass production of these products while enhancing the industrial robot sector [3] - Shanghai will accelerate the development of smart glasses, focusing on AR, MR, and VR technologies, and aims to establish a complete product matrix for various market segments [3] Group 3: Satellite Internet and Elderly Care - The action plan includes the establishment of a satellite terminal production line, targeting a production scale of 100,000 units and fostering the development of supporting industries [4] - Shanghai will support the development of innovative products for the elderly, such as AI hearing aids and smart assistive devices, enhancing the functionality of smart terminals [4] Group 4: Industrial Applications - The plan encourages the development of industrial terminal products that align with advanced AI technology trends, aiming to transform production models and improve efficiency in smart factories [5] Group 5: Key Technology Development - The action plan outlines five initiatives to strengthen key technologies, including enhancing AI chip layouts and promoting next-generation display technologies [6][7] - Shanghai will focus on lightweight, low-power, high-resolution display technologies, supporting innovations in Micro LED and OLED technologies [7] Group 6: Industry Ecosystem Optimization - The plan proposes five measures to optimize the industry ecosystem, including financial support for key components and services related to smart terminals [8] - Shanghai will leverage various funding mechanisms to support terminal enterprises and encourage mergers and acquisitions in the sector [8] Group 7: Application Promotion - The initiative will support the integration of smart terminal products in various sectors such as industry, education, and healthcare, promoting innovative applications [9] - Shanghai aims to develop key areas like Pudong and Songjiang into industrial clusters for smart terminals, fostering competitive enterprises and innovation [9]

Core Insights - Shanghai's Economic and Information Technology Commission has released the "Shanghai Intelligent Terminal Industry High-Quality Development Action Plan (2026-2027)" aiming to enhance the city's intelligent terminal industry and brand recognition by 2027, targeting an overall scale exceeding 300 billion yuan and the establishment of three globally influential consumer terminal brands [20]. Group 1: Key Product Development Actions - Accelerate the development of artificial intelligence computers, aiming to establish a production base with an annual capacity of 10 million units and support the transition of traditional computers to AI computers [21]. - Develop artificial intelligence smartphone products to fill existing gaps in production capacity and enhance interaction methods through AI integration [21]. - Enhance the scale of intelligent computing terminals by promoting edge computing solutions and developing lightweight inference machines [21]. Group 2: Robotics and Smart Devices - Strengthen capabilities in humanoid robots, focusing on emotional intelligence and skills, while supporting the development of core components for mass production [22]. - Promote the advancement of smart glasses across AR, MR, and VR categories, creating a comprehensive product matrix for various market segments [22]. Group 3: Satellite and Elderly Care Technologies - Cultivate satellite internet terminal products to meet deployment needs of low-orbit satellite constellations, aiming for a production scale of 100,000 units [23]. - Stimulate the silver economy by supporting the development of AI-powered assistive devices for the elderly, enhancing their functionality and accessibility [23]. Group 4: Industrial and Future Technologies - Support the development of industrial terminal products that align with advanced AI technology trends, facilitating smart factory transformations [24]. - Accelerate research in next-generation technologies such as mobile communication, quantum computing, and photonic computing [24]. Group 5: Key Technology Foundations - Strengthen the layout of AI chips for various terminal products, focusing on SoC and CPU developments across major technology routes [25]. - Enhance the performance of edge models for intelligent terminals, ensuring competitive capabilities in various application scenarios [25]. Group 6: Industry Ecosystem Optimization - Improve the competitiveness of Shanghai's intelligent terminal brands through targeted policies and support for brand operations [28]. - Accelerate the formation of industrial clusters by fostering collaboration among key districts and promoting the development of competitive enterprises [29].

Core Insights - The article highlights the significant advancements in AI large models, particularly in the context of mobile devices, indicating a shift from technical competition to a focus on deep user understanding [1][2][4] - Vivo has developed the world's first 3B (3 billion parameters) model specifically for mobile agents, which showcases capabilities in multimodal processing, reasoning, and long-context understanding [1][2][4] Model Breakthrough - The new 3B model allows for independent operation on mobile devices without relying on cloud resources, enhancing user experience and enabling personalized AI interactions [2][4] - This model excels in language processing, multimodal understanding, and logical reasoning, marking a transition from a "public service" AI model to a "personalized" AI model [4][5] User Experience Enhancements - The model enables instant responses and offline functionality, allowing users to perform tasks without internet connectivity, thus providing reliable assistance anytime and anywhere [5][6] - It can understand and execute commands related to both text and images, evolving from a conversational AI to an actionable assistant capable of performing tasks across applications [6][8] Ecosystem Development - Vivo emphasizes the importance of an open ecosystem to enhance AI capabilities, aiming to connect various intelligent agents with users to create a more personalized experience [8][10] - The company has established the "Blue Heart Personal Intelligence Framework," focusing on perception, memory, planning, and execution to improve user interaction with AI [8][10] Collaborative Opportunities - Vivo's strategy includes opening its AI capabilities to developers, allowing for a broader range of personalized services and applications, thus fostering a collaborative ecosystem [10][12] - Over 50 ecosystem partners have already integrated with Vivo's open platform, indicating a growing network of services that enhance user experience through personalized AI interactions [10][12]

Core Insights - The article discusses the challenges and opportunities in the AI and robotics industry, particularly focusing on the role of Qualcomm and various industry players in shaping the future of embodied intelligence and agent systems [1][4][31]. Group 1: Industry Challenges - The robotics field is currently facing diverse technical routes, leading to a perception of activity without significant progress [5][23]. - There is a critical need for improved communication protocols and reduced cable usage in robotics to enhance performance and reliability [16][17][20]. - The deployment of high computational power in robots is hindered by physical space limitations, battery capacity, and heat dissipation issues [19][20]. Group 2: AI and Robotics Development - The ultimate goal for robotics is to achieve a level of intelligence where robots can understand and execute tasks in unfamiliar environments using natural language instructions [10][11]. - The industry is encouraged to adopt an open-source approach to AI models, similar to OpenAI's early releases, to foster collaboration and accelerate development [25][26]. - The concept of agent systems is emerging as a key component in AI, with a focus on enhancing user experience through improved collaboration between cloud and edge computing [31][32]. Group 3: Future Directions - The future of AI in robotics will require a shift towards a unified operating system that can integrate various hardware and software components, creating a seamless user experience [44][45]. - Collaboration among industry players is essential for building the necessary infrastructure and standards to support the growth of AI and robotics [46][47]. - The focus is shifting from single-device intelligence to inter-device agent collaboration, indicating a trend towards more integrated and cooperative systems [48].

Core Insights - Anker Innovation, founded in 2011, has achieved annual revenue exceeding 24.7 billion and a market capitalization of over 70 billion, with over 95% of its revenue coming from overseas markets [4] - The company has adopted a "Shallow Sea Strategy," focusing on smaller consumer electronics markets through product innovation and channel advantages [5][10] - Anker's CEO, Yang Meng, emphasizes the importance of long-term sustainability over short-term gains, reflecting a shift in management philosophy and organizational structure [18][30] Company Overview - Anker Innovation's recent half-year report for 2025 shows revenue of 12.867 billion, a year-on-year increase of 33.36%, and a net profit of 1.167 billion, up 33.8% [4] - The company has faced challenges, including a recall of charging products and systemic failures in 2022, which led to a significant organizational review [8][34] Strategic Development - The "Shallow Sea Strategy" was born out of Yang Meng's concerns about the rapid obsolescence of consumer electronics, prompting a focus on sustainable business models [11][16] - Anker aims to become a "third-class company" by integrating multiple shallow product categories for long-term growth, rather than competing in deep-sea markets dominated by giants like Apple and Tesla [25][26] Organizational Structure - Anker has implemented a three-tier management structure to support its multi-category strategy, allowing for greater autonomy at the product line level [32][36] - The company has expanded its product lines significantly, from five business units to as many as 27, before consolidating to focus on core areas [31][58] Crisis Management - In response to the challenges faced between 2020 and 2022, Anker undertook a major restructuring, reducing its product lines from 27 to 17 and freezing aggressive hiring plans [58][61] - The company has emphasized a cultural shift towards focusing on user value rather than merely competing with peers, aiming to foster innovation [69][70] Talent Acquisition and Retention - Anker recognizes that attracting and retaining top talent aligned with its values is crucial for executing its strategy effectively [112] - The company offers competitive profit-sharing models and opportunities for rapid career advancement, particularly for new graduates [114][120] Brand Consolidation - Anker has streamlined its brand portfolio from several independent brands to three main brands: Anker for energy-related products, Soundcore for audio products, and Eufy for home automation [124][126] - This consolidation aims to enhance resource allocation and operational efficiency across its product lines [125][130]

Core Insights - The article highlights the debut of the first national administrative review vertical large model developed by Mianbi Intelligent at the 2025 China International Service Trade Fair, showcasing its capabilities in the administrative review process [2][3] - Mianbi Intelligent has adopted a differentiated strategy focusing on smaller models, which are more efficient and cost-effective for the legal and administrative sectors [5][7] - The company has made significant strides in the automotive sector, establishing partnerships with major automotive brands to implement its edge models [8][9] Company Developments - Mianbi Intelligent's administrative review vertical large model covers the entire case handling process, including case element extraction and document generation, and is currently utilized by the Beijing Judicial Bureau and its 16 district bureaus [3][4] - The company has completed multiple rounds of financing since its establishment in August 2022, with significant investments from various venture capital firms [6] - Mianbi Intelligent's focus on edge models distinguishes it from other large model companies, emphasizing speed and efficiency in processing [7][8] Industry Context - The article notes the increasing competition in the small model market, with major tech companies like Alibaba and Tencent entering the space, which poses challenges for Mianbi Intelligent [9] - The demand for edge models is expected to grow due to the proliferation of smart devices, offering unique advantages in privacy, data protection, and operational performance [7][8] - Mianbi Intelligent's strategic partnerships in the automotive sector aim to enhance its presence and application of edge models in smart vehicle technologies [8][9]

Core Viewpoint - The article discusses the rise of Multiverse Computing, a Spanish AI startup that has developed a compression technology called CompactifAI, which significantly reduces the size of AI models while maintaining performance, positioning itself as a leader in the AI efficiency race amidst growing competition from tech giants and startups alike [6][10][22]. Summary by Sections Company Background - Multiverse Computing was founded in 2019, initially focusing on quantum computing software for financial applications. It quickly gained recognition and funding, being named a "Cool Vendor" by Gartner, which is a prestigious acknowledgment in the tech innovation space [9]. - The company has completed five rounds of financing, with its valuation increasing from $108 million in 2024 to $500 million in 2025, making it one of the largest AI startups in Spain [6][8]. Technology Development - The company pivoted to AI model compression in 2023, leveraging its expertise in quantum tensor networks to address the rising computational costs associated with large AI models. This led to the development of CompactifAI, which can compress model sizes by 80-95% with minimal performance loss [10][13]. - The newly launched models, "SuperFly" and "ChickBrain," are touted as the smallest and highest-performing models, with SuperFly having 94 million parameters and ChickBrain having 3.2 billion parameters [15][17]. Market Position and Competition - Multiverse's technology has attracted interest from major hardware companies like Apple, Samsung, and Sony, aiming to integrate its small models into next-generation devices. This aligns with Apple's strategy to prioritize lightweight local models over large, general-purpose models [22]. - The competitive landscape is intensifying, with tech giants like Meta, Google, and Microsoft also entering the small model space, alongside startups like Neural Magic and Deci, all targeting improved AI performance and cost efficiency [21][23]. Business Model and Applications - Multiverse offers three commercial service models: API access to compressed models, private deployment licenses, and model compression services for clients. Its primary customers include large internet and software companies utilizing AI for various applications [17][18]. - The CompactifAI technology allows for significant cost savings in AI deployment, reducing inference costs by 50-80% and enabling models to run on less powerful hardware, thus broadening accessibility [20][17].