1月23日消息，多位知情人士透露，黄仁勋再度来华，首站到访了英伟达位于上海的新办公室，与员工见面并回答了诸多员工关注的问题， 同时回顾公司2025年主要事件。 "好几个问题质量不高，主要聚焦在2026年重点芯片相关的话题，"一位知情人士说，"(H200)这个还真没人问"。 据了解，黄仁勋此次来华行程与2025年初基本一致，主要参加上海、北京和深圳分公司的新年晚会以及供应商答谢会。 中国大陆作为英伟达最重要的市场之一，黄仁勋每年春节都会例行拜访。刚刚过去的2025年，黄仁勋更是三度拜访，以解决特供中国市场的H20产品的销 售问题，其中7月份来华参加链博会，身着唐装以中文开场进行演讲引发广泛关注。 关于黄仁勋此次来华，业内称其也有为H200对中国大陆出口铺路的目的。 此前，美国商务部已经对H200出口中国发放了特别许可，同时还推动了相关出口管制条款的调整，进而从法规层面为这款芯片的销售"解禁"。 根据腾讯科技、IT之家、CNMO等多家科技媒体1月23日最新报道，英伟达CEO黄仁勋开启2026年首次中国大陆之行，行程公开信息清晰明确，整体低调 务实，延续春节前访华惯例。 一、核心行程总览 1. 首站：上海 1月23 ...

Core Viewpoint - The company, Sibeichi Technology, is restarting its IPO process after changing its advisory institution, demonstrating its determination to enter the capital market despite previous setbacks in its listing attempts [2]. Group 1: Company Background - Sibeichi Technology was founded in 2007 in Cambridge, UK, by a team of graduates from prestigious universities, focusing on the intelligent voice language field [4]. - The company's core business covers three main areas: smart travel, smart IoT, and smart office, emphasizing a combination of software and hardware in AI technology and product services [4]. - Over the past decade, Sibeichi has completed multiple rounds of financing, raising several billion yuan, with investments from various well-known industrial and market-oriented investment institutions [4]. Group 2: Technology and Business Model - As a domestic leader in large model conversational AI platforms, Sibeichi has developed a "1+N distributed model intelligent system" under the leadership of Yu Kai, and is actively deploying in the field of embodied intelligence [5]. - The company has established solid partnerships with numerous well-known automotive manufacturers in the smart travel sector, significantly contributing to the production of mass-market vehicles [5]. - In the smart IoT and office sectors, Sibeichi's technology and services are widely applied across various terminal devices and government-enterprise client groups, creating a mature and diversified business layout [5]. Group 3: Financial Challenges and Future Outlook - Despite showing signs of revenue growth and narrowing losses in recent years, the company faces ongoing profitability pressures due to high R&D investments [5]. - The decision to restart the IPO process amid the AI industry wave is seen as a critical move for the company to seek capital support and accelerate development [5]. - The company needs to continue addressing concerns regarding its long-term development capabilities, focusing on core technology, competitive advantages, and optimizing its profit model [5].

Core Viewpoint - The article highlights the significant progress of Shanghai Xihe Technology Co., Ltd. in the silicon photonics sector, marking its initiation of the A-share IPO process, which positions it as a key player in the domestic silicon photonics market amid a global AI computing power race and surging demand for high-speed optical communication [1]. Group 1: Company Overview - Shanghai Xihe Technology was established in May 2021 with a registered capital of 97.64 million yuan, and its core team includes experts from the Chinese Academy of Sciences and overseas technology sectors [2][5]. - The company has achieved recognition as a national-level "specialized and innovative" small giant enterprise by 2025, reflecting its strong competitive edge in niche markets [2][5]. - The company’s major shareholder, Shanghai Xijing Management Consulting Partnership, holds 37.24% of the shares, while the chairman, Wu Aimin, controls 48.23% of the voting rights, ensuring a stable governance structure [2]. Group 2: Product and Technology Development - Xihe Technology focuses on silicon photonics technology, offering a full-service capability that spans design, manufacturing, and testing, with a focus on high-speed interconnects, intelligent sensing, and consumer applications [4]. - The company has launched a series of high-speed silicon photonic integrated chips, including 400G, 800G, and 1.6T products, which have been successfully delivered in millions, validating its commercial technology capabilities [4][5]. - Xihe Technology is advancing the development of 3.2T high-speed silicon photonic chips and is expanding its product matrix to include applications in autonomous driving and smart industrial sectors [4]. Group 3: Market Position and Future Outlook - The global silicon photonics market is experiencing a boom, with the penetration rate of 800G optical modules exceeding 30% by 2025, and demand for 1.6T and above products rapidly increasing [5]. - Xihe Technology is well-positioned to benefit from the high growth cycle in the industry, leveraging its established delivery capabilities and cutting-edge technology to break the overseas monopoly in the domestic market [5].

英伟达目前是全球市值最高的企业，市值逼近5 万亿美元；自搭上人工智能的发展快车后，公司市值便一路飙升。如今几乎所有人工智能基础设施 的算力都由英伟达提供，该企业也是当下人工智能革命的核心力量。 谈及英伟达的股价走势，不少人都为自己未能及早抓住人工智能的风口深感惋惜，软银集团首席执行官便是典型代表：他曾抛售的英伟达股份，如 今价值已高达数百亿美元。不过单从基本面来看，黄仁勋表示，在超大规模云服务商的全面布局推动下，全球正迎来 "人类历史上最大规模的基础 设施建设浪潮"；而前沿人工智能模型的迭代升级，以及各类上层应用的开发，都对算力提出了极高的需求。 "关键在于，这套人工智能计算平台的搭建，需要底层全产业链的支撑。这场基建浪潮已经开启，大家现在都有目共睹 —— 这是人类历史上规模最大的一 次。目前相关投入已达数千亿美元规模，数千亿美元啊。我和拉里合作过诸多项目，我们都清楚，未来还有数万亿美元的基础设施亟待建设，而这一投入极 具合理性。因为人工智能要生成支撑各类上层应用的智能算力，就必须对海量场景数据进行处理分析，这是一切的基础。"—— 英伟达首席执行官 黄仁勋 在世界经济论坛的活动中，英伟达首席执行官黄仁勋与贝莱 ...

Core Viewpoint - Shanghai Suiruan Technology has officially submitted its IPO application to the Shanghai Stock Exchange, aiming to raise 6 billion yuan for the development and industrialization of fifth and sixth-generation AI chips, targeting the global AI computing power market [1]. Group 1: Company Overview - Suiruan Technology is a leading player in the domestic cloud AI chip sector, focusing on "original innovation and independent research and development" to become a leader in general artificial intelligence infrastructure [5]. - The company has established a comprehensive core technology system covering hardware, software, and cluster solutions, with 262 domestic invention patents and participation in over 10 national and local technology projects [5][11]. - Suiruan Technology has developed GCU-CARE acceleration computing units and GCU-LARE interconnect technology, comparable to NVIDIA's Tensor Core and NVlink, ensuring programming flexibility and high parallel computing adaptability for AI large models [5][6]. Group 2: Product and Technology Development - The company has created a full-stack AI computing and programming software platform called "Yusuan TopsRider," which supports over 400 mainstream models and 300 application scenarios, significantly reducing development difficulty and migration costs for AI large models [6]. - Suiruan Technology's AI acceleration cards and modules are a core revenue source, providing high-performance domestic computing power for various AI applications [8]. - The CloudBlazer POD and distributed architecture of the CloudBlazer cluster enable efficient support for large model training, with products designed for different scales and requirements [9]. Group 3: Market Opportunity and Strategy - The global AI accelerator card market is projected to grow from over $100 billion in 2024 to over $500 billion by 2028, with China's market demand expected to exceed 1 trillion yuan, accounting for about 30% of the global market [12]. - The company plans to use the raised funds to address technology and market gaps, focusing on developing high-performance, efficient, and secure fifth-generation AI chips and enhancing interconnect capabilities in the sixth-generation chips [13]. - Suiruan Technology aims to lead the domestic AI chip sector by adhering to the principles of "national core, national manufacturing" and actively participating in major technological projects [14].

1月22日，安谋科技与香港科技大学正式签署合作备忘录，双方将围绕芯片IP设计、AI计算、具身智能、机器人等关键方向，推进产学研 协同创新，加速前沿技术应用落地。 安谋科技与香港科技大学签署合作备忘录 根据合作备忘录，双方将充分整合安谋科技的产业资源、技术能力与香港科技大学的科研实力，以AI计算与具身智能、机器人为重点，开 展协同研发与技术攻关。此次合作也致力于建立长期、稳定、深入的产学研融合机制，加快创新成果与实际场景结合，实现技术的普惠价 值，推动半导体与AI生态持续繁荣。 加入"中国IC独角兽联盟"，请点击进入 是说芯语，欢迎关注分享 合作洽谈，进入公众号：服务—>商务合作 安谋科技CEO陈锋（右），香港科技大学副校长（研究与发展）郑光廷教授（左） 此次合作备忘录由安谋科技CEO陈锋与香港科技大学副校长（研究与发展）郑光廷教授共同签署。 安谋科技CEO陈锋表示 ："在'AI Arm CHINA'战略发展方向指引下，公司正全力推进AI领域的前沿IP设计，持续深化在物理AI、边缘与端 侧AI、基础设施等方向的布局。本次安谋科技与香港科技大学的合作，将结合双方在集成电路与人工智能领域的核心优势，共同加速前沿 技 ...

Core Viewpoint - The article highlights the significant strategic financing of nearly 3 billion for Hangzhou-based GPU company Xiwang, aimed at advancing the development, production, and ecosystem of next-generation inference GPUs, positioning it strongly in the competitive landscape of AI computing power [1]. Group 1: Financing and Support - Xiwang secured nearly 30 billion in strategic financing within a year, with participation from industrial capital, leading VC/PE firms, and state-owned enterprises, indicating strong confidence in its technology and commercialization capabilities [1]. - The financing arrangement provides comprehensive support from technology research and development to market expansion, enhancing Xiwang's growth potential [2]. Group 2: Leadership and Team - Xiwang is led by a dual CEO team, with Wang Yong, a veteran in the chip industry, and Wang Zhan, a core member of Baidu's founding team, bringing extensive experience in both chip development and commercialization [3]. - The core team consists of around 300 members, many from top companies like NVIDIA and AMD, with an average of 15 years of industry experience and over 200 core patents, addressing the challenges of slow R&D and commercialization in domestic chip development [5]. Group 3: Differentiation Strategy - Unlike many competitors focusing on training and inference integration, Xiwang targets inference scenarios with a restructured native architecture, optimizing key components to reduce inference costs significantly [6]. - The company aims to lower the cost and accessibility of large model inference, making computing power widely available [7]. Group 4: Product Development - Xiwang has developed a three-generation chip matrix, with the S1 chip launched in 2020 for visual inference, the S2 chip set for 2024, and the upcoming S3 chip targeting a new benchmark of "one cent for a million tokens" [8]. - The S3 chip is designed to support low-precision inference, significantly reducing cost and energy consumption, which could redefine the cost structure of AI commercialization [8]. Group 5: Ecosystem Collaboration - Xiwang emphasizes ecosystem collaboration over zero-sum competition, positioning itself as a "cost optimization layer" within existing computing systems, fostering partnerships with local chip manufacturers [10]. - This strategy aims to create a virtuous cycle of broader application, refined technology, and lower costs, ultimately enhancing the overall strength of domestic computing power [10].

Core Insights - The article emphasizes the strategic decision of SenseTime to build its own computing infrastructure instead of selling to tech giants, which has become a crucial asset during the AI industry's value verification phase [4][6] - SenseTime's approach contrasts with the trend of many companies opting for quick profits through acquisitions, highlighting the importance of establishing AI as a foundational infrastructure [4][6] Group 1: Market Transition - The AI industry is undergoing a significant market logic shift, moving from a phase where capital was willing to pay premiums for technology concepts to a phase focused on value verification and stable cash flows [6][8] - SenseTime's resilience amidst this transition serves as a testament to its strategic foresight and comprehensive stack layout [6][8] Group 2: Financial Performance - By mid-2025, SenseTime's revenue from generative AI has surged to 77%, indicating a successful transition from traditional business models [11] - The company reported a cash reserve of 13.16 billion yuan by mid-2025, which is crucial for risk mitigation and valuation reconstruction during a capital downturn [12] Group 3: Infrastructure Investment - SenseTime began investing in computing infrastructure as early as 2018, recognizing the need for AI-specific cloud services, which were scarce at the time [14][16] - The company has developed a significant operational computing power of approximately 25,000 PetaFLOPS by August 2025, enhancing its capacity to support AI applications [16] Group 4: Competitive Positioning - SenseTime's strategy of building a comprehensive technology base allows it to avoid being trapped in low-end markets and instead pursue high-value growth avenues [18][22] - The company's dual-driven business model (B2B and B2C) is structurally more rational, with a current trading multiple of 13 times compared to competitors with much higher multiples [22] Group 5: Long-term Value Creation - The article suggests that mastering underlying computing power and full-stack implementation capabilities is becoming essential for companies to anchor long-term value amid market uncertainties [23]

1月22日消息，据复旦大学公众号介绍，今天凌晨，国际顶级学术期刊《自然》主刊发表了复旦大学彭慧胜/陈培宁团队的最新研究成果《基于多层旋叠 架构的纤维集成电路》， 团队成功在柔软的高分子纤维内制造出大规模集成电路，创造出世界首款"纤维芯片"。 有望为脑机接口、电子织物、虚拟现实等新兴产业提供强有力的技术支撑。 "纤维芯片"概念图 柔软的"纤维芯片"在手指上打结照片 经过近五年时间，团队先后攻克了高分子表面平整化、耐溶剂侵蚀、形变下电路稳定等多个技术难题，最终成功制备出具有信息处理功能的"纤维芯 片"。 该"纤维芯片"不仅保持了纤维柔软、可编织的本征特性， 更实现了电阻、电容、二极管、晶体管等电子元件的高精度互连，光刻精度达到了实验室级光 刻机最高水平。 "纤维芯片"及其内部局部电路光学照片 这意味着，基于"纤维芯片"， 未来可将发光、传感等模块直接集成在一根纤维上，形成无需外接设备的全闭环系统，甚至实现自供能。 据介绍，传统芯片的光刻工艺普遍依赖平整的硅晶圆衬底，而纤维不仅具有曲面结构，表面积极小，用于制备纤维器件的弹性高分子基底，也很难耐受光 刻过程中的各类极性溶剂。 同时还要保证在拉伸、扭转等变形中保持电路 ...

2026年1月22日，摩尔线程（688795）披露2025年年度业绩预告，公司预计实现营业收入14.5亿元至15.2亿元，较上年同期的4.38亿元同比增长230.70%至 246.67%；归属于母公司所有者的净利润预计亏损9.5亿元至10.6亿元，较上年同期16.18亿元的亏损额收窄34.50%至41.30%，扣非后净利润亏损也同步收窄 29.59%至36.32%。 业绩高增长与亏幅收窄的核心驱动力，来自公司在全功能GPU领域的技术突破与商业化落地。报告期内，摩尔线程持续推进产品架构迭代，成功推出旗舰 级训推一体全功能GPU智算卡MTT S5000，该产品性能达市场领先水平并实现规模量产。基于其构建的大规模集群已完成建设并上线服务，可高效支持万 亿参数大模型训练，计算效率达到同等规模国外同代系GPU集群先进水平。 得益于人工智能产业蓬勃发展及市场对高性能GPU的强劲需求，公司产品竞争优势扩大，市场认可度持续提升，推动收入与毛利增长。同时，公司仍处于 持续研发投入期，保持高研发投入力度，目前尚未盈利且存在累计未弥补亏损。公告同时提及，与部分国际巨头相比，公司在综合研发实力、核心技术积 累、产品客户生态等方面仍存 ...