杭州当虹科技股份有限公司 关于公司拟发行科技创新债券的公告 证券代码：688039 证券简称：当虹科技 公告编号：2025-032 本次拟注册发行科技创新债券的期限为不超过 10 年（含 10 年）。 3、发行利率 本公司董事会及全体董事保证公告内容不存在任何虚假记载、误导性陈述或 者重大遗漏，并对其内容的真实性、准确性和完整性依法承担法律责任。 杭州当虹科技股份有限公司（以下简称"公司"）于 2025 年 12 月 29 日召 开的第三届董事会第十三次会议及第三届董事会审计委员会第十一次会议，审议 通过了《关于公司拟发行科技创新债券的议案》，为积极响应国家科技创新政策 导向，加大科技创新投入力度，同时满足公司经营发展资金需求，进一步拓宽融 资渠道，优化债务结构，降低财务费用，董事会同意公司拟向中国银行间市场交 易商协会申请注册发行最高不超过人民币 5 亿元（含 5 亿元）科技创新债券。现 将相关情况公告如下： 本次拟注册发行科技创新债券的利率将根据公司信用评级、发行时的市场状 况以及监管部门有关规定确定。 一、本次拟注册发行科技创新债券的基本情况 1、注册规模及发行安排 公司本次拟申请注册总额不超过人民币 ...

证券代码：688039 证券简称：当虹科技 公告编号：2025-034 杭州当虹科技股份有限公司 一、 召开会议的基本情况 （一） 股东会类型和届次 2026年第一次临时股东会 召开日期时间：2026 年 1 月 16 日 15 点 30 分 召开地点：杭州市滨江区众创路 309 号当虹大厦 11 楼会议室 股东会召开日期：2026年1月16日 本次股东会采用的网络投票系统：上海证券交易所股东大会网络投票系 统 关于召开2026年第一次临时股东会的通知 本公司董事会及全体董事保证公告内容不存在任何虚假记载、误导性陈述或 者重大遗漏，并对其内容的真实性、准确性和完整性依法承担法律责任。 重要内容提示： 至2026 年 1 月 16 日 采用上海证券交易所网络投票系统，通过交易系统投票平台的投票时间为股 东会召开当日的交易时间段，即 9:15-9:25,9:30-11:30，13:00-15:00；通过互 联网投票平台的投票时间为股东会召开当日的 9:15-15:00。 （六） 融资融券、转融通、约定购回业务账户和沪股通投资者的投票程序 涉及融资融券、转融通业务、约定购回业务相关账户以及沪股通投资者的投 票，应按 ...

一、审议通过了《关于 2026 年度日常关联交易额度预计的议案》 经审议：公司与相关关联方开展的 2026 年度日常关联交易额度预计是基于 业务经营需要，关联交易价格遵循公平、自愿原则进行，不存在违反法律、法规、 《公司章程》及相关制度规定的情况；日常关联交易定价合理、公允，不存在损 害公司和全体股东尤其是中小股东利益的行为；该等关联交易不会对公司的独立 性产生不利影响，公司的主要业务不会因此类交易对关联方形成依赖；该等关联 交易属于公司的正常业务，有利于公司经营业务发展，符合公司和全体股东的利 益；董事会审议该议案时，关联董事已回避表决，审议程序符合相关法律、法规、 《公司章程》的规定，董事会同意该议案。 证券代码：688039 证券简称：当虹科技 公告编号：2025-033 杭州当虹科技股份有限公司 第三届董事会第十三次会议决议公告 本公司董事会及全体董事保证本公告内容不存在任何虚假记载、误导性陈 述或者重大遗漏，并对其内容的真实性、准确性和完整性依法承担法律责任。 杭州当虹科技股份有限公司（以下简称"公司"）第三届董事会第十三次会 议于 2025 年 12 月 24 日以通讯方式发出通知，于 2025 ...

Core Viewpoint - Donghong Technology's stock price has shown significant volatility, with a year-to-date increase of 66.35%, but recent trends indicate a decline over the past 20 and 60 days [1][2]. Group 1: Stock Performance - On December 30, Donghong Technology's stock rose by 2.06%, reaching 45.03 CNY per share, with a trading volume of 56.18 million CNY and a turnover rate of 1.14%, resulting in a total market capitalization of 4.98 billion CNY [1]. - The net inflow of main funds was 5.66 million CNY, with large orders accounting for 28.95% of total purchases and 18.88% of total sales [1]. - The stock has experienced a 6.76% increase over the last five trading days, but a decline of 5.72% over the last 20 days and 11.88% over the last 60 days [1]. Group 2: Company Overview - Donghong Technology, established on May 12, 2010, and listed on December 11, 2019, specializes in intelligent video technology, including algorithm research, video encoding and decoding, intelligent portrait recognition, and video cloud services [2]. - The company's revenue composition includes 73.26% from media culture products and solutions, 17.20% from automotive intelligent cockpit products, 7.27% from industrial and satellite products and solutions, 2.07% from leasing income, and 0.21% from other sources [2]. - As of September 30, 2025, the company reported a revenue of 210 million CNY, a year-on-year increase of 24.02%, but a net profit loss of 21.64 million CNY, which is a 68.84% increase in losses compared to the previous year [2]. Group 3: Shareholder Information - As of September 30, 2025, the number of shareholders increased by 25.03% to 11,300, with an average of 9,822 shares per shareholder, a decrease of 20.96% [2]. - The top ten circulating shareholders include Jin Ying Technology Innovation Stock A and Hua An Small and Medium Growth Mixed Fund, with changes in holdings noted [3].

Core Viewpoint - The company has completed its investment in outdoor large screen construction and operation projects, establishing a nationwide ultra-high-definition outdoor screen operation network [1] Group 1: Company Strategy - The company is implementing a "Hundred Cities, Thousand Screens" ecological layout, focusing on outdoor large screen projects [1] - The company has developed a unique "networked broadcasting control platform + content operation" dual-driven model, creating a complete business model from hardware construction to traffic monetization [1] Group 2: Operational Progress - The company has nearly completed its asset investment in the outdoor large screen projects [1] - The operation network covers multiple cities across the country, indicating a broad geographical reach [1]

每经AI快讯，12月26日，当虹科技在互动平台表示，公司设工业与卫星事业部，核心产品主要应用于 工业场景及基于商业航天的卫星互联网场景。公司的多模态以及超级压缩技术，包括超级图像、语音、 视频压缩，能大幅提升多模态信息的标识与处理的效率，降低卫星通信传输与存储成本，实现远洋、深 海、深空等低延时窄带通信。 ...

Core Insights - Global technology investment enthusiasm remains high in 2025, but there is a structural differentiation characterized by "strong computing power, weak applications" [3] - Major global tech giants are significantly increasing capital expenditures, benefiting upstream computing chips and supporting hardware, while the computer/software sector faces three pressures: 1) Software companies' performance growth is overall lower than that of the computing chain; 2) Software company valuations are not cheap compared to hardware; 3) Concerns about "large models consuming software" continue to grow [3] - By Q3 2025, domestic public funds are underweight in the computer sector by 1.7 percentage points [10] Market Review: AI Applications' Concerns and Breakthroughs - The performance of the computer/software sector in both China and the US has been weak, with software and service-focused indices significantly underperforming hardware and semiconductor indices [5] - As of December 12, 2025, the A-share computer index has risen by 15.15%, lagging behind the 49.16% increase in the ChiNext index and the 45.90% increase in the electronic index [5] - In the US market, the IGV software index has only increased by 7.84%, underperforming the 22.52% rise in the Nasdaq Composite and the 39.75% increase in the SOXX semiconductor index [5] Opportunities in Chinese AI Applications 1. Deepening Industrial Applications - China possesses the world's only complete industrial system, providing high-value real-world data, application scenarios, and energy supply, which will foster diverse agent forms [27] - The construction of a modern industrial system is prioritized in the "14th Five-Year Plan," with expectations for intensive policy support for "Artificial Intelligence+" during this period [39] 2. Expanding Overseas - AI application teams in China can leverage "extreme efficiency + global pricing" arbitrage, with the engineer dividend in China continuing to expand [3] - Many computer companies prioritize overseas strategies, with overseas revenue growth generally outpacing domestic growth [64] 3. Reconstructing Edge AI - Internet giants and AI startups are actively exploring edge forms in the AI era, integrating deeply with existing consumer electronics and developing new AI-native hardware [3] - The shipment volume of related products is expected to have high growth potential as hardware and algorithms mature [3] Investment Recommendations - Focus on companies with industry know-how and data for industrial empowerment, such as Hikvision, Dahua Technology, and Han's Laser [3] - Prioritize AI application companies with high overseas revenue proportions, including Foxit Software and Kingsoft Office [3] - Pay attention to AI hardware brands and algorithm vendors benefiting from the emergence of blockbuster products, such as Rainbow Soft and iFlytek [3]

Core Insights - The first MUSA Developer Conference showcased the launch of the new GPU architecture "Huagang" by Moore Threads, along with AI training and inference chip "Huashan" and high-performance graphics rendering chip "Lushan" [1][4][5][7] - Moore Threads introduced the "Kua'e" supercomputing cluster, featuring the self-developed "Yangtze" intelligent SoC chip, aimed at enhancing AI computing capabilities [1][9] Group 1: New GPU Architecture and Chips - The "Huagang" GPU architecture features a 50% increase in computing density and supports full precision end-to-end calculations from FP4 to FP64, with new asynchronous programming models and MTLink high-speed interconnect technology for scaling over 100,000 cards [4][14] - The "Huashan" chip focuses on AI training and inference, integrating new asynchronous programming and full precision tensor computing units, supporting large-scale intelligent computing clusters [5] - The "Lushan" chip specializes in high-performance graphics rendering, achieving a 64x increase in AI computing performance, 16x in geometric processing, and 50x in ray tracing performance, catering to AAA games and high-end graphics creation [7] Group 2: Collaborations and Ecosystem Development - Several companies listed on the Sci-Tech Innovation Board, including Dahong Technology and Zhongwang Software, are collaborating with Moore Threads to leverage its GPU for high-performance needs such as ultra-high-definition live streaming and offline video enhancement [3] - The MUSA software architecture has been upgraded to version 5.0, enhancing compatibility with programming languages like TileLang and Triton, and achieving over 98% efficiency in core computing libraries [12] Group 3: Industry Challenges and Future Directions - The need for a unified or highly compatible interface standard for domestic GPU chips is emphasized to avoid fragmentation and inefficiencies in the software ecosystem [13] - The transition from "usable" to "willing to use" domestic GPU platforms hinges on improving developer experience and reducing migration costs [12] - The engineering challenges of building large-scale systems without proprietary interconnects are highlighted, with a focus on achieving reliable low-latency communication and operational efficiency [14]

Core Insights - The AI inference chip market is experiencing explosive growth due to a structural shift in computing power demand from training to inference, driven by the maturation of open-source large models and the increasing need for real-time applications [1][5] - The Chinese market, being the largest AI application market globally, shows a significant demand for advanced AI chips, supported by favorable government policies [1][5] - The market size for AI inference chip products and services in China is projected to reach 162.6 billion yuan in 2024 and 310.6 billion yuan in 2025, with GPUs dominating the market share [5] Industry Overview - AI inference chips are specialized high-performance computing chips designed to accelerate the inference of AI models, focusing on energy efficiency, low latency, and high throughput [1][4] - The AI chip category includes various types such as GPUs, FPGAs, and ASICs, with core functions for executing data training and inference tasks [2][4] Market Size and Growth - The market for AI inference chips in China is expected to grow from 162.6 billion yuan in 2024, an increase of 105.2 billion yuan from 2023, to 310.6 billion yuan in 2025 [5] - In terms of product structure, GPUs are projected to hold 81.6% of the market share in 2024, decreasing to 78.4% in 2025, while NPUs are expected to increase from 16% to 19.3% [5] Industry Chain - The AI inference chip industry chain consists of upstream components like AI algorithms, semiconductor materials, and EDA tools, midstream activities including design, manufacturing, and testing, and downstream applications in data centers and smart devices [6] Policy Environment - The Chinese government regards AI chips as a core infrastructure for AI and has implemented various policies to promote the development of the AI chip industry, including the "14th Five-Year Plan" and local initiatives [8] Competitive Landscape - The competitive landscape of the AI inference chip industry in China is characterized by international giants dominating the general market while domestic players rapidly penetrate niche segments [9] - Major domestic companies include Huawei, Baidu, and several others, with a trend towards differentiated competition [9] Development Trends - The AI inference chip industry is seen as a strategic high ground, with ongoing technological innovations focusing on energy efficiency, architectural advancements, and ecosystem integration [10] - By 2026, the competition is expected to intensify, with key players like NVIDIA, Google, Qualcomm, and Huawei shaping the market dynamics [9][10]

Core Insights - Yunshen Technology, one of the "Six Little Dragons of Hangzhou," has completed a C-round financing of over 500 million RMB, led by CMB International and Huaxia Fund, with participation from various strategic investors [1][2]. Company Overview - Founded in 2017, Yunshen Technology focuses on the independent research and development, lean production, global sales, and professional services of humanoid and quadruped robots and their core components, leading in B-end applications such as power inspection and emergency rescue [2]. - The founding team originates from Zhejiang University and has over 10 years of experience in legged robot research [2]. Recent Developments - The recent financing coincides with the completion of the company's share reform, transitioning from a limited liability company to a joint-stock company (non-listed) [2]. - This move has sparked market expectations for an IPO, following similar actions by other companies in the embodied intelligence sector this year [2]. Product Innovations - In April, Yunshen Technology launched the M20, the world's first industry application quadruped robot designed for complex terrains and hazardous environments, weighing 33 kg [3]. - In October, the company introduced the DR02, the world's first all-weather humanoid robot, which operates in a wide temperature range of -20°C to 55°C and has an IP66 protection rating [3]. Application and Impact - Yunshen Technology's robotic dogs have been deployed in power inspection and security patrols, covering areas that are difficult or risky for human inspection, resulting in a 70% reduction in operational costs and over 50% improvement in efficiency [3]. - The company has also expanded its applications, with its robotic dog operating in the remote Hoh Xil region, providing valuable data for wildlife research [4]. Strategic Partnerships - Several listed companies, including Ruixin Microelectronics, Danghong Technology, and Meian Technology, have disclosed business cooperation with Yunshen Technology, with some holding indirect stakes in the company [4].