公众号记得加星标⭐️，第一时间看推送不会错过。 美光耗资 1000 亿美元的芯片工厂在纽约州破土动工之际，环保人士、工会和民权组织组成的联盟敦 促这家美国科技巨头签署一项协议，使其一系列做好邻居的承诺具有法律约束力。 美光公司用于生产存储芯片的巨型晶圆厂有望成为该州历史上最大的商业开发项目，也是全美最大的 芯片制造基地。上周五，官员们在锡拉丘兹附近的克莱市举行了奠基仪式。首批芯片有望在五年内交 付，但整个厂区要20年才能完全建成。 由约25个主要为地方倡导团体组成的"纽约中部社区福利联盟"的组织者和成员告诉《连线》杂志，他 们欢迎这个项目。他们也赞赏美光公司已承诺在当地招聘员工，并解决其建设带来的部分环境和社会 影响。但联盟成员认为，目前缺乏监管，美光公司可能会逃脱惩罚，造成环境污染，并加剧该地区的 经济 不平等。 "我们希望有真正、有力、透明且可执行的承诺，"美国就业促进组织（Jobs to Move America）的高 级研究员安娜·史密斯（Anna Smith）说道。该组织是一个亲工会的全国性非营利组织，正在帮助组 织这个联盟。 周三，该联盟公布了一封发给美光公司首席执行官桑杰·梅赫罗特拉的电子邮件， ...

公众号记得加星标⭐️，第一时间看推送不会错过。 被动元件市场掀起涨价潮，华新丽华集团旗下华新科对客户发出涨价通知，将从2月1日起调升所有芯 片电阻产品线报价，业界透露涨幅上看二成，涨价效益将助力华新科营运走强。 华新科是继台湾被动元件龙头国巨、大陆最大芯片电阻厂厚声之后，又一家宣布调升芯片电阻报价的 被动元件大厂，透露这波价格涨势锐不可挡，产业进入新一波涨价周期态势确立。 华新科指出，已正式通知客户，自2月1日起，针对尺寸0201至1206、涵盖全阻值范围的芯片电阻产 品进行价格调整，具体调整幅度将依各产品规格、型号及种类另行提供详细方案，由公司业务团队与 客户进一步说明。 华新科指出，过去一段时间，该公司已透过内部控管与营运优化吸收部分成本，惟随新应用市场持续 发展、产能需求同步提升，整体营运成本逐步累积，经审慎评估后，决定启动价格调整，以反映实际 成本结构与市场状况。 华新科强调，此次价格调整是在多方权衡后所做的必要措施，未来将持续致力于提供稳定供应、高品 质产品与即时支援服务，并与客户携手因应市场变化，维持长期合作关系。 在 华 新 科 之 前 ， 国 巨 已 率 先 宣 布 ， 自 2 月 1 日 起 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 据彭博新闻社援引知情人士报道，人工智能初创公司 SambaNova Systems Inc. 在与英特尔公司的收 购谈判陷入僵局后，正寻求从科技公司和半导体制造商处筹集 3 亿至 5 亿美元资金。 据知情人士透露，这家总部位于加州帕洛阿尔托的公司在与英特尔的融资谈判破裂后，目前正在寻求 各路投资者的资金支持。由于细节尚未公开，这些知情人士要求匿名。据知情人士透露，英特尔正在 考虑是否对这家初创公司追加投资，其首席执行官陈立武同时也是 SambaNova 的董事长。 英特尔此前曾讨论过收购SambaNova，该交易对这家人工智能初创公司的估值约为16亿美元，其中 包括债务。彭博社去年12月报道了此次收购谈判。 此前报道：英特尔计划收购SambaNova 据彭博社去年年底报道，英特尔正在调整其人工智能路线图，并洽谈收购人工智能处理器设计公司 SambaNova。SambaNova是一家定制人工智能芯片开发商，此前因融资失败而寻求收购。 在取消原定于今年发布的Falcon Shores人工智能加速芯片的计划后，英特尔一直在筹划重返人工智 能 市 场 。 作 为 此 次 ...

2025年，韩国各产业部门的境遇喜忧参半。受美国加征关税和俄乌战争的影响，汽车和石化行业举步 维艰；而半导体、造船和国防等行业则因全球需求增长而蓬勃发展。我们将探讨2026年影响全球经济 的关键因素，并预测各行业将受到的影响。[编者注] 公众号记得加星标⭐️，第一时间看推送不会错过。 去年支撑韩国出口增长的是"半导体"行业。随着人工智能（AI）在全球范围内的普及，以三星电子和 SK海力士为首的存储芯片需求激增，并进入繁荣期。半导体出口增长超过30%，抵消了石化和钢铁 等传统制造业的下滑。 今年的前景更为光明。据韩国进出口银行海外经济研究所预测，韩国今年半导体出口额预计将同比增 长11%，达到1880亿美元（约合254万亿韩元），连续第二年创下历史新高。 世界半导体贸易统计（WSTS）预测，2026年全球半导体市场规模将比上年增长超过25%，达到约 9750亿美元。其中，存储器市场预计将实现30%左右的增长，超过整体增速。在DRAM市场，三星 电子和SK海力士合计占据约70%的全球市场份额；在NAND闪存市场，两家公司的市场份额合计约 为50%。 三星电子和SK海力士合并后的营业利润预计将超过200万亿韩元 在 ...

Core Insights - Upscale AI, a chip startup, has raised $200 million in Series A funding to challenge Nvidia's dominance in rack-level AI systems and compete with companies like Cisco, Broadcom, and AMD [1][3] - The rapid influx of investors reflects a growing consensus that traditional network architectures are inadequate for the demands of AI, which require high scalability and synchronization [1][2] Funding and Market Position - The funding round was led by Tiger Global, Premji Invest, and Xora Innovation, with participation from several notable investors, bringing Upscale AI's total funding to over $300 million [1] - The AI interconnect market is projected to reach $100 billion by the end of the decade, prompting Upscale AI to focus on this growing sector [6] Technology and Product Development - Upscale AI is developing a chip named SkyHammer, optimized for vertical scaling networks, which aims to provide deterministic latency for data transmission within rack components [9][10] - The company emphasizes the importance of heterogeneous computing and networks, believing that no single company can provide all the necessary technologies for AI [10][12] Competitive Landscape - Nvidia's networking revenue has seen a significant increase, with a 162% year-over-year growth, highlighting the competitive pressure in the AI networking space [3] - Upscale AI aims to create a high radix switch and a dedicated ASIC to compete with Nvidia's NVSwitch and other existing solutions [14][16] Strategic Partnerships and Standards - Upscale AI is building its platform on open standards and actively participating in various alliances, including the Ultra Accelerator Link and SONiC Foundation [7][17] - The company plans to expand its product line to include more traditional horizontal scaling switches while maintaining partnerships with major data center operators and GPU suppliers [18]

公众号记得加星标⭐️，第一时间看推送不会错过。 半导体行业观察人士表示，存储器市场已进入长期上升周期，DRAM 和 NAND 的供应短缺状况可能 至少持续到 2028 年。 历史上，内存供应商的供应一直处于供过于求和供不应求的状态，时而需求旺盛，时而供不应求，时 而库存积压，利润暴涨暴跌，供应过剩后又会陷入断供。然而，这一次，人工智能训练和推理工作负 载驱动的需求正在重塑这一周期。内存需求的激增如此强劲，以至于制造商需要数年时间才能扩大产 能以满足需求。 从美光半导体内存产品的收入和利润来看，就可以看出传统的从供过于求到供过于求的模式： 2016年至2018年需求急剧上升，2019年和2020年跌至谷底，2021 年和2022年再次攀升，然后在 2023年再次下滑。此后，需求在2024年和2025年持续增长，进入2026年后仍未见放缓迹象。 美光科技在去年12月发布2026财年第一季度业绩报告时表示，持续的行业需求加上供应限制，导致 市场供应紧张，预计这种情况将持续到2026年以后。这意味着市场上行周期至少会持续三年，甚至可 能更久。 这种情况会持续多久？"存储器价格可能在2027-2028年期间保持高位，只 ...

Core Viewpoint - Nvidia has become TSMC's largest customer, surpassing Apple, which previously held this position for many years [1][3][5]. Group 1: Nvidia's Rise - Nvidia's revenue has surged due to the booming demand for AI GPUs, with enterprise clients willing to spend billions on these processors [2][5]. - Nvidia's sales are projected to grow by 62% by January 2026, while Apple's product revenue is expected to grow only by 3.6% by December 2025 [5][7]. - The demand for high-performance chips driven by AI is significantly outpacing the growth of smartphone sales, which have plateaued [7][8]. Group 2: Apple's Challenges - Apple is facing increased chip prices from TSMC and may no longer have priority production rights, which could hinder its competitiveness [2][3]. - Apple's revenue growth rate has slowed to single digits, contrasting sharply with Nvidia's rapid growth [7][8]. - The competition for TSMC's capacity has intensified, with Nvidia and AMD taking up more space on the production lines, making it harder for Apple to secure its needs [3][11]. Group 3: TSMC's Performance - TSMC reported a 36% revenue growth, reaching $122 billion, with a gross margin of 62.3% in the last quarter [5][8]. - The sales of high-performance computing chips, including AI chips, grew by 48%, while smartphone revenue only increased by 11% [7][8]. - TSMC's capital expenditures are expected to rise significantly, indicating a strong investment in future technologies [7][8]. Group 4: Market Dynamics - The shift in customer dynamics at TSMC reflects broader industry trends, with AI driving demand for advanced chips while smartphone growth stagnates [7][15]. - Nvidia's business model allows for high margins, but it faces risks related to inventory surplus, while TSMC must balance capacity expansion with market demand [18][19]. - The semiconductor industry is experiencing a transformation, with Nvidia's influence growing at the expense of traditional players like Apple [19][20].

Core Insights - The memory shortage is expected to persist until 2027 due to the strong demand for DRAM driven by artificial intelligence data centers, leading to increased prices across the memory market [1][3] - The current price dynamics are influenced by concerns over future supply shortages, prompting customers to secure memory supplies in advance, which exacerbates the shortage and drives up spot prices [1][3] Group 1: Market Dynamics - The latest round of DRAM price increases began in Q3 2025, with a 13.5% quarter-over-quarter rise, indicating a peak in the market cycle and potential for a correction [3] - Early signals from company earnings reports suggest that prices may rise further by 30% in Q4, driven by fears of supply shortages [3] - The spot price for DDR5 memory used in servers has surged by 100% in some cases, impacting PC manufacturers like HP and Dell, who may remove certain laptop models from their product lines due to high DRAM prices [3] Group 2: AI Infrastructure Impact - The core imbalance in the memory market is attributed to the construction of AI infrastructure, with data center operators heavily investing in AI accelerators that require high-bandwidth memory (HBM) and standard DDR5 memory [4] - An AI server with eight accelerators requires approximately 1.6TB of HBM and 3TB of DDR5 memory, significantly more than a typical non-AI server, leading to a rapid increase in memory demand that exceeds supply capabilities [4] Group 3: Broader Market Effects - The automotive sector, which uses LPDDR4 and LPDDR5 memory, is strategically important for memory suppliers, especially with the rise of autonomous vehicles requiring more memory [5] - The production processes for LPDDR and DDR memory are about 80% similar, meaning that if DRAM companies prioritize AI server production, LPDDR supply will also be affected [5] - Micron's decision to gradually shut down its Crucial consumer business reflects a strategic shift towards higher-margin AI-driven demand rather than consumer products [5] Group 4: Demand and Supply Outlook - Data centers dominate DRAM demand, accounting for about 50% of total bit demand, with AI workloads representing approximately 30% of that demand [6] - Historical trends show that DRAM market cycles can change rapidly, but the typical self-regulating mechanism of high prices leading to reduced demand has not yet occurred due to the insensitivity of data center operators to price increases [6] - Structural constraints in supply relief are evident, as building or expanding a DRAM factory typically takes 2-3 years to reach mass production, with limited new supply expected until 2026 [6][7] Group 5: Future Supply Developments - Companies like CXMT are expanding capacity primarily for domestic clients, while Samsung is prioritizing HBM production over broader DRAM products [7] - SK Hynix's M15X factory is expected to start production in late 2026, and Micron's new factory in Boise is anticipated to increase capacity in 2027 [7] - Until significant capacity increases occur, smartphone and PC manufacturers may need to slow down memory capacity growth or AI infrastructure spending to alleviate price pressures [7]

公众号记得加星标⭐️，第一时间看推送不会错过。 日前，有媒体报道指出，1月15日，广汽集团董事长冯兴亚率队到访格力电器，格力电器董事长董明 珠接待了来访一行。在接待广汽集团董事长冯兴亚时，格力电器董事长董明珠几天前笑言，未来广汽 的汽车芯片中一半由格力产品替代。 于是，市面上有了广汽将一半芯片将用格力产品的报道见诸报端。对此，1月20日，广汽集团官方微 博发文辟谣，称近期不少朋友关注广汽与格力的合作动态，网传"未来广汽的汽车芯片中一半由格力 产品替代"，相关表述并非事实。 广汽集团表示，1月15日，广汽集团董事长冯兴亚率队访问格力电器，与董事长董明珠及管理团队共 商"人车家"智慧生态融合发展，探讨产业协同。后续如有合作的具体进展，将会第一时间通过官方渠 道发布。请勿轻信或传播未经证实的信息。 值得一提的是，在日前举办的大湾区化合物半导体生态应用大会上，格力电器总裁助理、珠海格力电 子元器件有限公司总经理冯尹也透露，格力电器碳化硅芯片工厂在量产家电用碳化硅芯片后，今年光 伏储能用、物流车用碳化硅芯片也将量产。 据了解，碳化硅（SiC）是一种化合物半导体材料，优势在于耐高压、耐高频、耐高温。2019年，格 力开始 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 在全球人工智能（AI）与高效能运算（HPC）需求呈现爆炸性成长的推动下，晶圆代工大厂联电正积 极执行成熟制程升级计画，正式卡位共同封装光学（Co-Packaged Optics，CPO）世代。根据经济日 报 引 用 供 应 链 最 新 消 息 指 出 ， 联 电 已 将 新 加 坡 Fab 12i P3 新 厂 确 立 为 承 载 硅 光 子 （ Silicon Photonics）与CPO 技术的核心基地，内部已进入实质技术导入与试产准备阶段，目标于2027年实现 量产。 报导表示，联电近年来致力于提升成熟制程的附加价值，将22/28奈米制程导向更具竞争力的特殊制 程应用。供应链透露，联电新加坡Fab 12i P3新厂在此战略中扮演关键角色。该厂不仅以22/28奈米 制程做为营运主轴，服务通讯、车用、物联网与AI等应用市场，更进一步强化在CPO应用的布局。 据了解，该厂的22奈米产线已具备衔接硅光子产品的成熟制程基础。针对外界关注的布局进度，联电 官方回应证实，在特殊制程的推进上，硅光子确实是公司积极投入的重点技术之一。供应链进一步指 出 ， 目 前 该 厂 ...