格隆汇2月1日｜1月31日晚间，英伟达CEO黄仁勋现身中国台湾台北市砖窑古早味怀旧餐厅，宴请了35 位与英伟达合作的供应链厂商高管，不过实际到场约有接近40位台企高管，有1位是大陆企业高管。 现场唯二女性，除了纬颖董事长洪丽宁外，另一位是显卡大厂同德股份有限公司CEO刘盈君，她上次也 有受邀出席。 此次宴会合影环节，第一排特地安排椅子座位，跟黄仁勋同坐第一排的包括华硕董事长施崇棠、联发科 副董事长蔡力行、台积电董事长魏哲家、广达董事长林百里及副董事长梁次震、纬创董事长林宪铭，业 界戏称是礼让70岁或是头发已白德高望重的供应链大佬优先入座。 第二排之后依序为英业达董事长叶力诚、矽品董事长蔡祺文、鸿海董事长刘扬伟、和硕董事长童子贤、 纬创总经理林建勋、技嘉总经理林英宇及李宜泰、纬颖董事长洪丽宁、奇鋐董事长沈庆行、广达资深副 总杨麒令、微星董事长徐祥、仁宝董事长陈瑞聪、 英业达总经理蔡枝安、工业富联董事长郑孟弘、光 宝总经理邱森彬、台达电董事长郑平、和硕共同执行长郑光志与邓国彦。京元电子总经理张高薰、华擎 总经理许隆伦、永擎总经理沙韦旭、宏碁董事长陈俊圣、华硕共同执行长许先越及胡书宾、研华总经理 张家豪、鸿佰科技董 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 英伟达首席执行官黄仁勋昨晚宴请供应链伙伴高层，「兆元宴」再度登场。黄仁勋表示，台积电今年 必须要非常努力工作，因为英伟达需要很多晶圆，他预期「未来十年台积电的产能可能会成长超过百 分之百，是非常显著的规模扩张，而光是为英伟达的需求就得翻倍」。 黄仁勋再度选在台北砖窑古早味怀旧餐厅举行「兆元宴」，和供应链伙伴餐叙，出席的包括台积电董 事长魏哲家、联发科首席执行官蔡力行、广达董事长林百里、英业达董事长叶力诚、纬创董事长林宪 铭、鸿海董事长刘扬伟、宏碁董事长陈俊圣、矽品董事长蔡祺文、和硕董事长童子贤、共同首席执行 官郑光志及邓国彦、华硕董事长施崇棠、共同首席执行官胡书宾及许先越、纬颖董事长洪丽寗、台达 电董事长郑平、仁宝董事长陈瑞聪以及云达总经理杨麒令等人。餐叙结束后黄仁勋亲自送魏哲家离 开，显见对台积电的重视程度。 黄仁勋说，台积电今年要非常努力工作，因为英伟达需要很多晶圆和CoWoS，他也说，台积电做得 非常好。英伟达已经全面投产Blackwell、Vera Rubin芯片，而Vera Rubin包含六款不同的芯片，每 款都是世界上最先进的芯片。黄仁勋强调，今年英 ...

Group 1 - NVIDIA's CEO Jensen Huang hosted a "trillion-dollar banquet" in Taiwan, attended by key supply chain partners and industry leaders, indicating strong collaboration and networking within the semiconductor industry [1][2] - Huang emphasized the challenges ahead in 2025 with the production of the Grace Blackwell semiconductor system, which has faced significant design modifications and supply chain disruptions [2][3] - The GB300 cabinet has entered the initial production phase, while the GB200 product is in smooth mass production, showcasing NVIDIA's advancements in technology [2] Group 2 - Huang noted a significant shift in the AI industry, highlighting the profitability of tokens as AI technology has become more effective, suggesting a positive outlook for AI-driven business models [3][4] - The banquet featured a diverse array of industry leaders, including representatives from major companies like ASUS, MediaTek, and TSMC, reflecting the importance of NVIDIA's partnerships in the tech ecosystem [3][4] - The event underscored NVIDIA's commitment to innovation and collaboration, with companies like Advantech transitioning to edge AI solutions, indicating a broader industry trend towards integrating AI into various sectors [4][5]

在接受采访时，黄仁勋谈到供应链的动态。他表示，今年英伟达需求非常强劲，英伟达正在全力生产 Blackwell芯片，同时生产Rubin芯片。台积电今年必须要非常努力工作，因为英伟达需要很多晶圆和 CoWoS产能，台积电现在做得非常好。他还表示，台积电在未来10年可能会增加100%产能，这是规模 很大的基础设施投资。 格隆汇1月31日｜据一财，英伟达CEO黄仁勋春节前访华到达上海、北京、深圳后，近日又到达台湾。 此次行程中，黄仁勋与当地供应商的交流备受关注。1月31日晚间，黄仁勋宴请了英伟达的供应链伙 伴，参加晚宴的包括纬创董事长林宪铭、台积电董事长魏哲家、广达创始人林百里及副董事长梁次震、 和硕集团董事长童子贤、鸿海董事长刘扬伟、仁宝董事长陈瑞聪、联发科执行长蔡力行、华硕创始人施 崇棠等。晚宴结束后，黄仁勋送魏哲家离开。 ...

Core Viewpoint - Huang Renxun's recent visit to Taiwan highlights Nvidia's strong demand and ongoing investments in AI infrastructure, emphasizing the company's commitment to expanding its production capabilities and partnerships in the AI sector [3][4][6]. Group 1: Nvidia's Demand and Production - Nvidia is experiencing robust demand this year and is fully engaged in the production of Blackwell and Rubin chips, indicating a strong growth trajectory for the company [4]. - TSMC is expected to significantly increase its production capacity by 100% over the next decade, which represents a substantial infrastructure investment to meet Nvidia's needs for wafers and advanced packaging technology [4]. Group 2: Competitive Landscape - Huang Renxun asserts that while ASICs (Application-Specific Integrated Circuits) have demand, Nvidia's approach is unique as it encompasses the entire AI infrastructure, producing a range of products including CPUs, GPUs, and networking chips [5]. - The assertion that ASIC shipments will surpass those of GPUs is dismissed, with Huang emphasizing that achieving better ASICs than Nvidia's products requires superior R&D capabilities, which many companies are attempting but have not yet succeeded [5]. Group 3: Investment in OpenAI - Reports suggest that Nvidia's $100 billion investment plan in OpenAI has stalled due to Huang's concerns, but he clarified that the partnership remains intact and Nvidia is considering participating in OpenAI's new funding round [6]. - Huang noted that the AI infrastructure development is at its inception and will take approximately 10 years, necessitating global computing facilities, including new factories in Taiwan, the US, Europe, Japan, and Southeast Asia [6].

Core Viewpoint - Nvidia's CEO Jensen Huang hosted a dinner in Taipei with key supply chain executives, emphasizing the importance of collaboration and the challenges ahead in AI technology development [1][4][10]. Group 1: Event Details - The dinner included approximately 40 executives from Taiwanese companies, with only one representative from a mainland Chinese firm [1]. - Notable attendees included leaders from Asus, MediaTek, TSMC, Quanta, and Wistron, highlighting the significance of these partnerships [1][3]. Group 2: Nvidia's Future Plans - Huang discussed the upcoming challenges in 2025 with the production of the Grace Blackwell architecture, indicating that it presents more difficulties than previous models [7]. - The GB300 cabinet has entered initial mass production, and the GB200 product is being produced smoothly, while the Vera Rubin platform is expected to simplify future production processes [8][9]. Group 3: AI Industry Insights - Huang noted that AI has become increasingly useful, with large language models now generating revenue, contrasting with previous years when AI was less effective [8]. - He projected that 2026 will be a critical year for the AI industry, with unprecedented demand for high-bandwidth memory (HBM) and LPDDR, leading to significant supply chain pressures [9]. Group 4: Nvidia's Strategic Investments - Nvidia plans to participate in OpenAI's next funding round, potentially marking its largest strategic investment to date, reflecting the company's commitment to AI development [9]. - Huang emphasized that Nvidia's comprehensive AI infrastructure, which includes CPUs, GPUs, and networking chips, cannot be easily replaced by specialized AI chips (ASICs) [9]. Group 5: Importance of Taiwanese Supply Chain - Huang stated that Nvidia's existence is heavily reliant on Taiwan's technological capabilities and engineering culture, particularly praising TSMC's role in advanced manufacturing processes [10]. - He anticipates that TSMC's capacity will grow significantly over the next decade, contributing to a major expansion in global technology infrastructure [10][12].

黄仁勋表示，英伟达与OpenAI的合作关系没有改变，正在考虑参与OpenAI的新一轮融资。 针对ASIC（专用集成电路）带来的竞争，黄仁勋表示，ASIC一直有需求，但英伟达做的事情非常不一 样。一方面，英伟达不只做一种芯片，而是参与到整个AI基础设施建设过程，做的产品包括CPU、 GPU、网络芯片、交换器芯片等。另一方面，英伟达与几乎所有AI公司合作，包括谷歌。此外，英伟 达与每一个云都相关，一些云计算厂商在与英伟达竞争，但这样也没有关系，英伟达还是无处不在，在 电脑系统、机器人和车里。 黄仁勋表示，ASIC出货量将比GPU更大是无稽之谈，要做到比英伟达产品更好的ASIC，要有比英伟达 更好的研发人员，许多公司正在尝试，但英伟达仍走在前面。 黄仁勋强调，英伟达年研发成本近两百亿美元。科技变得越来越复杂，英伟达此前的芯片架构Hopper很 简单、Blackwell太难了，现在做Rubin则几乎接近不可能，未来英伟达研发成本每年还会增长50%。 近日有消息称，英伟达千亿美元投资OpenAI的计划陷入停滞，原因是黄仁勋对OpenAI有疑虑。黄仁勋 回应称，双方的合作关系没有改变，英伟达正在考虑参与OpenAI的新一 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 近年生成式人工智能（Generative AI）与高效运算（HPC）需求持续爆发，推动先进封装一跃成为 全球半导体供应链中最关键且最紧张的产能瓶颈。 据 供 应 链 业 内 人 士 透 露 ， 台 积 电 南 科 AP8 厂 区 将 新 增 P2 工 厂 ， 两 座 工 厂 均 以 晶 圆 级 系 统 集 成 （CoWoS）技术为主；而原本聚焦晶圆级多芯片封装（WMCM）、系统级集成芯片（SoIC）及面板 级系统集成封装（CoPoS）的嘉义 AP7 厂区，也将把 SoIC 产线调整为 CoWoS 产线。换言之，台 积电未来两年将全面提升 CoWoS 产能，这也促使稳居台积电 CoWoS 供应链的中国台湾设备与材料 厂商加速扩产，以应对订单满载的市场盛况。 供应链指出，除英伟达已成为 CoWoS 最大客户外，谷歌等定制芯片（ASIC）客户也频繁抛出紧急 订单争抢产能。在这双重需求的推动下，台积电近期已敲定全面上调 2026—2027 年 CoWoS 产能目 标，并重新审视原有的先进封装扩产蓝图。 其中，2026 年底 CoWoS 月产能将突破 14 万片，2027 ...

Group 1 - NVIDIA's CEO Jensen Huang attended the company's year-end party in Taiwan, highlighting the collaboration with MediaTek to develop the N1 series processor, which is designed for powerful AI computers with low power consumption [1] - The theme of the year-end party was "NVIDIA Shines," and Huang expressed gratitude for the hard work of employees and the support from Taiwanese partners, noting the rapid growth of NVIDIA's operations in Taiwan [1] - NVIDIA's product offerings have expanded from GPUs to include network chips, switch chips, smart data processors, and CPUs, with future plans to launch the world's smallest AI supercomputer, DGX Spark, in collaboration with MediaTek [1] Group 2 - Huang discussed the development of quantum computing, emphasizing that while quantum computing can simulate nature, traditional CPUs and GPUs are still essential, and AI will remain a crucial computational model [2] - NVIDIA is working on integrating GPU and QPU technologies to create hybrid supercomputers, with significant breakthroughs in quantum bit error correction expected to lead to practical applications in the coming years [2] - Huang mentioned his meetings with supply chain partners in Taiwan, with expectations for a significant gathering referred to as the "Trillion Dinner," attended by high-level executives from the supply chain, including Foxconn's chairman [2]

Core Insights - Counterpoint Research predicts that MediaTek will lead the global smartphone SoC market in 2025 with a shipment share of 34.4%, followed by Qualcomm (25.1%), Apple (18.1%), Unisoc (12.1%), and Samsung (5.7%) [1] - The global smartphone SoC market is expected to slow down in 2026, with a projected year-on-year decline in shipments of 7% [1] - Despite the overall decline in shipments, the market is shifting towards higher-end products, with one in three smartphones expected to sell for over $500 by 2026 [4] Market Dynamics - The rise in storage prices poses a significant challenge for the smartphone industry, particularly affecting entry-level products priced below $150 [4] - Companies with in-house SoC capabilities, such as Samsung, Google, Huawei, and Xiaomi, are better positioned to navigate market challenges compared to those reliant on 4G and entry-level 5G SoCs [4] - OEMs are adjusting their product offerings and exploring cloud offloading strategies amid ongoing supply constraints [5] Technological Advancements - Leading high-end smartphone SoC manufacturers are expected to transition from 3nm to 2nm process nodes by 2026, with Samsung already set to launch the first 2nm smartphone SoC, Exynos 2600, in December 2025 [5] - The smartphone SoC market is anticipated to achieve double-digit revenue growth in 2026, driven by ongoing premiumization, rising storage prices, and the rapid adoption of AI features in smartphones [5] AI Integration - By 2026, edge AI performance is expected to reach around 100 TOPS, with nearly 90% of high-end smartphones supporting edge AI capabilities [6] - Mid-range smartphones priced between $100 and $500 may increasingly rely on cloud-based AI processing to manage costs amid ongoing pressure from storage prices [6]