当地时间1月6日，在CES 2026联想TechWorld大会上，联想集团董事长杨元庆与英伟达CEO黄仁勋共同宣布，两家公司将携手推出"联想人工智能 云超级工厂"，进一步深化合作伙伴关系。 在会上，杨元庆与黄仁勋、超威半导体（AMD）董事会主席兼CEO苏姿丰、英特尔CEO陈立武，以及高通CEO克里斯蒂亚诺·安蒙一齐亮相。 在主题演讲中，杨元庆表示，"联想人工智能云超级工厂"计划旨在赋能AI云服务商以更快的速度上线下一代AI工作负载与应用，推动客户以前所 未有的规模从研发迈向生产阶段。 此外，通过六款新芯片的极致协同设计，大幅缩短了训练时间并降低推理token成本。 CES大会上，黄仁勋展示了新架构Vera Rubin 直播截图 黄仁勋表示，"NVIDIA 正与联想携手交付全栈计算平台，为Agentic AI系统提供动力——覆盖从云端、本地数据中心到边缘及机器人系统的各个 领域。" 联想方面表示，这一计划将帮助云服务提供商极大缩短"time to first token"AI部署的时间，同时可迅速扩展规模至十万台GPU，支持万亿参数级别 的智能体和大语言模型。 在此次大会前夕，联想集团发布了一则杨元庆与黄仁勋的 ...

国芯网[原:中国半导体论坛] 振兴国产半导体产业！ 1月6日消息，英伟达CEO黄仁勋在CES演讲上展示了英伟达新一代AI平台Rubin，包含六款新型芯片，分别为Vera CPU、Rubin GPU、NVLink 6交换机、 ConnectX-9超级网卡、BlueField-4 DPU及Spectrum-6以太网交换机。 其中，Rubin GPU芯片搭载第三代Transformer引擎，NVFP4推理算力是50PFLOPS，是Blackwell的5倍。英伟达Rubin平台已进入全面生产阶段，基于该平 台的产品将于2026年下半年通过合作伙伴面市。 据路透社消息，黄仁勋当天不忘提中国市场，以及此前已获批对华出口的H200人工智能（AI）芯片。 在主旨演讲后，黄仁勋向金融分析师声称，中国市场对H200芯片"有需求"，且"需求强劲"。 英伟达首席财务官科莱特·克雷斯（Colette Kress）则表示，该公司已申请向中国运送H200芯片的许可证，但正在等待美国和其他政府的批准。 上个月，路透社曾援引三位匿名知情人士报道称，英伟达已告知中国客户，计划于2026年2月中旬农历春节假期前，对中国交付H200芯片。 不拘中国 ...

被称为科技春晚的国际消费电子产品展览会(CES)于1月6日正式开幕。 作为英伟达新年战略和新品发布的重要窗口，本次英伟达CEO依然身着去年同款亮面鳄鱼皮衣登场，发表了他在2026年的首场演讲。 在CES开始前，英伟达在社交媒体发布称"(CES2026)不会发布新款GPU。"这也是英伟达五年来首次不在CES发布新款GPU产品。 相比去年CES官宣了RTX50系产品，本次演讲中，黄仁勋把重点放在了新一代计算平台和英伟达在物理AI领域的进展，包括自动驾驶和机器人，相关开源 模型和工具等。 演讲开始。黄仁勋表示过去十年投入的约10万亿美元计算资源，正在被彻底现代化。 但这不仅仅是硬件的升级，更多的是软件范式的转移。 他特别提到了具备自主行为能力(Agentic)的智能体模型，并点名了Cursor，彻底改变了英伟达内部的编程方式。 随后，黄仁勋对2025年开源社区给予了高度评价。他表示，去年DeepSeek的突破让全世界感到意外，它作为第一个开源推理系统，直接激发了整个行业 的发展浪潮。 而在介绍开源生态时，黄仁勋的PPT后出现了三家中国模型的名字，分别是月之暗面的Kimi K2，深度求索的DeepSeek V3.2和 ...

6 证券研究报告 点评报告 2026 年 1 月 6 日 晨会纪要 核[心Ta观bl点e_：Summary] 一、产业综合组： 1 月 5 日美国 CES 展开始举行，英伟达 CEO 黄仁勋在 CES 演讲上展示了英伟达新一 代 AI 平台 Rubin，其中 Rubin GPU 芯片搭载第三代 Transformer 引|擎，FP4 推理 算力 50PFLOPS，是 Blackwell 的 5 倍。黄仁勋表示 Rubin 平台已进入全面生产阶 段，产品将于 2026 年下半年面市，表明下一代的 AI 芯片进展正常。除了 AI 芯片 外，本届 CES 展的主要技术动态还包括英伟达、高通都发布了机器人芯片套件， 谷歌下属的 Deep Mind 公司也与波士顿动力的 Atlas 人型机器人合作，因此我们 认为黄仁勋所说的机器人领域已经迎来"ChatGPT"时刻的正确性很高，继续看好 今年芯片、人工智能和机器人的产业景气度提升。 1 月 5 日中国证监会召开资本市场财务造假综合惩防体系跨部门工作推进座谈会。 证监会主席、最高人民法院、最高人民检察院、国家发展改革委、公安部、司法 部、财政部、中国人民银行、国务院国资委 ...

格隆汇1月6日｜当地时间1月5日，英伟达CEO黄仁勋在CES 2026演讲上展示了英伟达新一代AI平台Vera Rubin。英伟达高性能计算与AI基础设施解决方案高级总监Dion Harris将Vera Rubin描述为"六颗芯片构 成的一台AI超级计算机"。其中，Rubin GPU芯片搭载第三代Transformer引擎，NVFP4推理算力为 50PFLOPS，是英伟达上一代Blackwell GPU的5倍。据介绍，英伟达Rubin平台已进入全面生产阶段，基 于该平台的产品将于2026年下半年通过合作伙伴面市。 东吴证券表示，Rubin开启AI PCB高端材料新时代，M9与Q布掀起PCB价值跃升潮，该机构关注相关公 司如胜宏科技、菲利华等。山西证券表示，考虑到英伟达GPU芯片主要用于AI服务器中，因此预计AI 服务器技术将持续演进升级、需求将保持强劲增长。PPO树脂、碳氢树脂、low-dk电子布等高频高速覆 铜板核心原材料需求有望快速提升。 ...

Core Insights - Nvidia has become the first company to reach a market capitalization of $5 trillion, closing at $5.03 trillion on October 29, 2023, with a stock price of $207 per share [1] - The company's market value has increased significantly from under $1 trillion two years ago, showcasing a remarkable growth trajectory [1][4] - Nvidia's CEO Jensen Huang has seen his wealth rise, ranking eighth on the Forbes billionaire list as of October 29, 2023 [1] Market Growth - Nvidia's market capitalization surpassed $1 trillion in June 2023, and it reached $2 trillion by March 2024, indicating rapid growth [2] - The company achieved a market value of $3 trillion within three months after hitting $2 trillion, and it took approximately one year to grow from $3 trillion to $4 trillion [3][4] - The leap from $4 trillion to $5 trillion was accomplished in just over three months, setting a new global record [4] AI and Semiconductor Industry Impact - Nvidia is positioned as a key player in the AI era, significantly benefiting from the demand for AI technologies since the launch of ChatGPT [5] - The company's revenue growth rate is projected to be 125% for 2023-2024, while other fabless companies are expected to see around 20% growth [6] - Nvidia's Blackwell and Rubin GPU chips are anticipated to generate over $500 billion in revenue over the next five quarters, with current orders reaching 20 million units, five times the previous generation's total [6] Client Demand and Capital Expenditure - Major cloud service providers like Amazon, Meta, Google, Microsoft, Oracle, and CoreWeave are expected to increase their capital expenditures to $632 billion by 2027 [6] - Goldman Sachs predicts that the total capital expenditure of hyperscalers will approach $1.4 trillion from 2025 to 2027, nearly tripling from the previous period [7] - Nvidia's growth has positively impacted its supply chain, with companies like Industrial Fulian experiencing significant revenue and market value increases due to AI business growth [7][8] Nvidia's Position in the Tech Industry - Nvidia is recognized as a leading provider of AI computing infrastructure, with its technology being fundamental to the current AI landscape [9] - The company has outperformed other tech giants like Microsoft and Apple, with its market value leading the "Big Tech" sector [10] - Despite facing challenges such as competition and the complexities of AI implementation, Nvidia is expected to continue innovating and leading in the tech industry [10][11]

Core Insights - Nvidia has become the first company to reach a market capitalization of $5 trillion, closing at $5.03 trillion on October 29, 2023, with a stock price of $207 per share [3][4] - The company's market value has increased dramatically, surpassing $1 trillion in June 2023, $2 trillion in March 2024, and reaching $3 trillion in just three months [3][4] - Nvidia's CEO Jensen Huang has seen his wealth rise significantly, ranking eighth on the Forbes billionaire list as of October 29, 2023 [3] Market Performance - Nvidia's revenue growth rate for 2023-2024 is projected at 125%, significantly higher than the 20% growth seen by other fabless companies [6] - The demand for Nvidia's Blackwell and Rubin GPU chips is strong, with expected revenues exceeding $500 billion over the next five quarters and an order volume of 20 million units [6][7] - Major cloud service providers are projected to increase their capital expenditures to approximately $1.4 trillion from 2025 to 2027, nearly tripling from the $485 billion spent from 2022 to 2024 [7] Industry Impact - Nvidia's success has positively influenced the AI server and related markets, with companies like Industrial Fulian reporting significant revenue growth due to the expanding AI server market [8] - Industrial Fulian's revenue for the first three quarters of 2023 reached 603.93 billion yuan, a 38.4% year-over-year increase, with net profit growing by 48.52% [7][8] - The AI industry is experiencing a new industrial revolution, with Nvidia positioned as a key player in providing the necessary computational infrastructure [10][11] Future Outlook - Nvidia's market leadership is expected to continue, with analysts projecting a target price increase to $275, which would correspond to a market cap of $6.68 trillion [10] - The company faces challenges in AI implementation, including algorithm ecosystems, energy consumption, and regulatory boundaries, but remains focused on innovation and leadership in the tech industry [11]

Core Viewpoint - Morgan Stanley's recent report indicates that Oracle's orders exceeded expectations, positively impacting the overall AI semiconductor market sentiment, leading to an "overweight" rating for TSMC and an increased target price for King Yuan Electronics to 188 TWD [2] Group 1: Oracle and AI Semiconductor Market - Oracle's financial performance serves as a catalyst for NVIDIA and the AI supply chain, with expectations of 28,000 NVL 72 server cabinet orders by 2026 [2] - Broadcom has secured a $10 billion customized AI chip order from a fourth cloud customer, likely OpenAI, although Morgan Stanley analysts suggest this figure may represent cabinet value rather than guaranteed revenue [2] Group 2: TSMC and CoWoS Supply Chain - TSMC's COO mentioned that semiconductor technology has entered the "Moore's Law 2.0" era, emphasizing system integration over mere chip miniaturization [2] - TSMC's CoWoS capacity is projected to reach 93 kpwm by 2026, with OpenAI accounting for a small portion of approximately 10,000 units [3] Group 3: NVIDIA and Other Collaborations - NVIDIA's Rubin GPU chip is expected to begin mass production in Q2 2026, according to supply chain surveys [3] - Google, in collaboration with Broadcom, is increasing the production forecast for its 3nm TPU v7 to 300,000 units by 2026, while the 3nm TPU v8's output forecast has been revised down to 200,000-300,000 units due to delays [3][4] Group 4: King Yuan Electronics - With the anticipated ramp-up of TSMC's CoWoS TPU, Morgan Stanley expects an upward revision in King Yuan Electronics' 2026 TPU testing volume, raising its target price from 158 TWD to 188 TWD [5]