Core Viewpoint - The article highlights significant advancements in two leading Chinese large model companies, DeepSeek and Zhiyu, with new model releases expected to enhance their capabilities and market position [2][3]. Group 1: DeepSeek Developments - DeepSeek announced the upload of its new model, DeepSeek-V3.2, to the HuggingFace community platform on September 29 [2]. - The previous version, DeepSeek-V3.1, released in August, introduced a mixed inference architecture that supports both thinking and non-thinking modes, improved thinking efficiency, and enhanced agent capabilities through post-training optimization [3]. Group 2: Zhiyu Developments - Zhiyu is set to release its new model, GLM-4.6, with some users already able to access it via API [2]. - The flagship model GLM-4.5, launched in July, integrates reasoning, coding, and agent capabilities into a single model to meet complex application needs [3]. - In August, Zhiyu also released the GLM-4.5V, a high-performance open-source visual reasoning model with a total of 106 billion parameters and 12 billion active parameters [3].

Core Insights - The new AI model DeepSeek-V3.2 was uploaded to the community platform HuggingFace but was subsequently removed [1] - The new model GLM-4.6 from Zhipu is expected to be released soon and is currently accessible via API [1]

Core Insights - DeepSeek has uploaded the v3.2-base model to its official HuggingFace page, although the model file is currently offline [1] Summary by Categories - **Company Developments** - DeepSeek has made progress by uploading the v3.2-base model to its official HuggingFace page [1]

Core Insights - The AI community has discovered that the new model DeepSeek-V3.2 was uploaded to the HuggingFace platform but was subsequently removed [1] - Additionally, the new model GLM-4.6 from Zhipu is expected to be released soon and is currently accessible via API [1]

Core Insights - The new generation of electric vehicles is being shaped by a deep understanding of consumer needs, with companies like BMW adapting to the Chinese market and consumer preferences [1][12][24] Group 1: Collaborations and Innovations - BMW collaborates with Momenta to create a locally tailored intelligent driving assistance solution, providing comprehensive navigation features for various driving conditions [3] - A partnership with Alibaba aims to enhance the new BMW intelligent personal assistant using AI language models, set to debut in 2026 on domestic models [5] - Collaboration with Huawei introduces features like BMW digital key and HUAWEI HiCar, enhancing the in-car digital experience, with integration planned for 2026 [8] - BMW teams up with DeepSeek to improve human-machine interaction centered around the intelligent personal assistant, reflecting a commitment to meeting local user demands [10] Group 2: Product Features and Testing - The BMW iX3, set to launch in China, features a new electric platform and advanced systems, achieving a maximum range of over 900 kilometers under CLTC conditions [12][22] - The "Driving Control Super Brain" integrates key vehicle control systems for real-time adjustments, enhancing driving dynamics with a response time of less than one millisecond [14] - The BMW iX3 has undergone extensive testing, including extreme temperature trials, ensuring stability and performance under harsh conditions [21][22] - The vehicle's battery system supports rapid charging, achieving over 400 kilometers of range in just 10 minutes, ensuring consistent performance even in high temperatures [22] Group 3: Market Positioning - The upcoming BMW iX3 represents the company's commitment to leading the electric vehicle market and setting new luxury standards in the industry [24]

Investment Rating - The electronic industry is rated as "Outperform" compared to the market [1] Core Insights - The demand for NAND Flash is expected to surge due to the increasing needs from AI servers and the latest smartphone models, particularly with the iPhone upgrading its minimum storage capacity from 128GB to 256GB [2][5] - The NAND Flash market is projected to experience significant price increases of 15%-20% in Q4 2025, breaking the traditional year-end price decline [5][11] - The market size for NAND Flash is anticipated to reach $38.73 billion in 2023 and $65.64 billion in 2024, driven by supply-demand balance improvements and high-capacity storage needs from AI servers and data centers [11][12] Summary by Sections NAND Flash Market Overview - NAND Flash is a non-volatile memory technology primarily used for data storage, offering advantages over DRAM and SRAM in terms of data retention and cost [6][9] - The NAND Flash market is characterized by a fluctuating growth trend, with major manufacturers reducing production to improve supply-demand balance [11][12] Price Trends and Market Dynamics - As of September 22, 2025, the prices for various NAND Flash products are as follows: MLC 256GB at $12.20, MLC 128GB at $9.45, SLC 16GB at $9.94, and SLC 8GB at $3.47 [5][11] - The top five companies in the NAND Flash market by market share in Q2 2025 are Samsung (32.9%), SK Group (21.1%), Kioxia (13.5%), Micron (13.3%), and SanDisk (12.0%) [12] Company Spotlight: Jiangbo Long - Jiangbo Long is recognized as a leading semiconductor storage brand, providing high-end, flexible, and efficient customized services, with a focus on NAND Flash and DRAM products [21][22] - The company's revenue grew from 7.276 billion yuan in 2020 to 17.464 billion yuan in 2024, with a compound annual growth rate of 24.47% [22][30] - Jiangbo Long's product offerings include embedded storage, solid-state drives, mobile storage, and memory modules, with embedded storage being the core product [27][30] Recent Developments in the Industry - The electronic industry index increased by 3.51% in the week of September 22-26, 2025, outperforming the Shanghai Composite Index, which rose by 1.07% [2][46] - The semiconductor equipment sector has shown strong performance, with a 15.56% increase in the semiconductor equipment index [2][50]

Core Viewpoint - The article emphasizes the challenges faced by AI companies in global expansion, particularly in infrastructure, talent, and business models, with a focus on the integration of high-quality large models and products for effective globalization [2][6]. Group 1: Development and Technical Route of Mureka Model - The Mureka model was initiated in 2020, leveraging existing music processing technologies and data accumulated from a music social product, Starmaker, which holds a significant market position overseas [7]. - The decision to enter the music generation field was based on the observation that the scale of music data is smaller compared to text and video data, leading to lower required investment and training resources [7]. - The company initially explored various technical routes for music generation, ultimately adopting the Diffusion Transformer (DIT) approach, which significantly improved the model's performance [9]. Group 2: Global Promotion Challenges and Non-Acquisition Growth - After developing the model, the company faced challenges in global promotion, particularly the reliance on user acquisition (UA) models, which are less effective in the AI startup landscape [11]. - Non-acquisition growth strategies include leveraging core technological breakthroughs for viral growth, SEO for user acquisition, and GEO optimization, which are essential for companies to explore beyond traditional UA methods [12][13]. Group 3: Product Judgement Standards and Market Opportunities - The article outlines two core judgments for the feasibility of To B products: they serve as "efficiency multipliers" and act as "workflow adhesives" to enhance automation [15]. - For To C products, the focus is on reducing production costs significantly, with AI music generation costing less than 0.1 RMB per song compared to traditional methods costing around 100,000 RMB [16]. - The article highlights the growing competitiveness of Chinese open-source large models, indicating that small and medium enterprises can leverage these models to build new ecosystems and tap into vast market opportunities [17].

Core Viewpoint - The Chinese public fund industry is experiencing a technological revolution led by artificial intelligence, which is expected to significantly impact global economic and geopolitical dynamics in the coming years [1] Group 1: Technological Revolution and Economic Impact - Each technological revolution has historically led to a transformation in human civilization, with knowledge density being the root of competitive advantage [2] - The current AI revolution is expected to create a surplus in knowledge and data products, fundamentally reshaping global division of labor and accelerating significant changes in the geopolitical landscape [3] Group 2: AI Integration and Economic Challenges - The integration of AI with the real economy is seen as a key to addressing China's current challenges, including avoiding the "Thucydides Trap," preventing manufacturing hollowing, and mitigating the effects of an aging population [4][6] - The high profit margins in AI-related investments, with a theoretical cost-profit ratio of 545%, indicate a significant demand for computing power, which is expected to attract substantial capital investment [4][5] Group 3: Policy and Reform for Growth - Comprehensive reforms to promote the development of the private economy are essential for China to succeed in the AI and digital economy era [6][7] - The Chinese public fund industry is encouraged to leverage its capital allocation capabilities to support the integration of AI with the real economy, thereby contributing to national strategic transformation and innovation [7]

Core Insights - DeepSeek has gained significant recognition by being featured on the cover of Nature magazine, highlighting its leading position in the AI field [4][19] - The article emphasizes the importance of peer review in the AI industry, noting that DeepSeek is the first major model to undergo this rigorous process, filling a critical gap in the sector [5][6][19] Group 1: Industry Impact - DeepSeek's peer-reviewed status is seen as a breakthrough, contrasting with the trend of other AI models that have not been subjected to such scrutiny, which has often led to a lack of transparency [6][7][19] - The traditional approach to AI training has been to use supervised fine-tuning (SFT), where models learn from human-generated solutions. DeepSeek challenges this by allowing its model to learn independently through reinforcement learning [8][19] Group 2: Technological Innovation - DeepSeek's model, DeepSeek-R1-Zero, was trained using a unique method that involved presenting it with difficult problems without any human guidance, simulating a high-pressure learning environment [11][12] - The model demonstrated advanced reasoning capabilities, including self-reflection and error correction, which were not previously expected from AI systems trained without human input [15][16] Group 3: Strategic Decisions - The decision to open-source its findings and model is framed as a long-term strategy to build trust, accelerate innovation, and attract top talent in the AI field [17][18] - By publishing in Nature, DeepSeek aims to establish itself as a credible player in the AI landscape, emphasizing the importance of transparency in gaining societal trust [19]

每经记者｜刘艳美 每经编辑｜杨欢 这几天，一场科技盛宴正在杭州上演。 9月25日起，为期五天的第四届全球数字贸易博览会（以下简称"数贸会"）正式拉开帷幕。 作为我国唯一以数字贸易为主题的国家级、国际性、专业型展会，本届数贸会的一大亮点，就是围绕数字贸易核心优势，不仅注重改造提升传统产业，也注 重培育壮大新兴产业，布局建设未来产业。 在15.5万平方米的展会现场，超1800家参展商携前沿技术、创新产品和重磅首秀亮相。其中，低空经济作为新质生产力的典型代表和战略性新兴产业，在此 次数贸会上备受关注。 9月27日，第四届数贸会"低空经济新兴技术与产业发展对接会"举行。会上，由每日经济新闻·每经智库、城市进化论与火石创造联合编写的《城市低空经 济"链接力"指数（浙江2025）》报告正式发布。 报告从企业聚集度、创新聚集度、资本活跃度、环境友好度和AI低空融合度等五大维度，对浙江各市低空经济发展水平进行了系统性评估，为城市找准"坐 标"、协同发展提供参照。 在对接会圆桌对话环节，每经智库总裁、城市进化论主编江然还与德国不来梅州经济促进局亚洲部部长Matthias Hempen，浙科投资董事长、浙江省创业投 资协会副会长 ...