Core Viewpoint - Nvidia's CEO Jensen Huang announced the unveiling of a "world-first" new chip at the upcoming GTC 2026 conference, which is expected to further solidify the company's leading position in the AI infrastructure sector [1][4]. Group 1: Upcoming Conference Details - The GTC 2026 conference will take place on March 15 in San Jose, California, focusing on a new era of AI infrastructure competition [4]. - Huang acknowledged the challenges in developing these new chips, stating that "all technologies are approaching their limits," yet the industry remains optimistic due to Nvidia's track record [4]. Group 2: New Chip Series - Specific models of the new chips have not been disclosed, but speculation suggests they may come from two major series: the Rubin series derivatives and the next-generation Feynman series [4]. - The Rubin series, which includes six new chip designs, has already been fully mass-produced and was showcased at the 2026 CES [4]. - The Feynman series is described as "revolutionary," with Nvidia exploring broad integration using SRAM and potential 3D stacking technology for LPUs, although details remain unconfirmed [4]. Group 3: Addressing AI Compute Needs - Nvidia is adapting to quarterly changes in AI compute demands, shifting focus from the Hopper and Blackwell series, which emphasize model pre-training, to the Grace Blackwell Ultra and Vera Rubin series, which target inference scenarios [4]. - The new products are expected to specifically address latency and memory bandwidth bottlenecks [4]. Group 4: Strategic Positioning - Huang emphasized that extensive collaboration and investment are key to Nvidia's continued leadership, as the company is positioning itself across the entire AI industry chain, including energy, semiconductors, and data centers [4].

" 历时一年，我们在算力产业的褶皱中寻找真相。 " 作者丨 刘伊伦 编辑丨 包永刚 2025年年初，DeepSeek以算力效率的极致优化，打破大模型高算力、高成本的部署壁垒，中国AI产业迎 来了一个极具象征意义的时刻，这不仅是一次技术路线的胜利，更为国产算力基础设施与全球顶尖模型同 频共振提供舞台。 从华为昇腾、寒武纪，到摩尔线程、昆仑芯，国产芯片厂商纷纷抓住这一轮模型革新带来的适配机遇，与 国内开源AI生态形成紧密联动。 然而，技术的浪漫主义终究要让位于产业的冷峻法则。 年初炙手可热的一体机仅四个月便沦为"明日黄花"，智算中心在"算力闲置"与"卖卡求生"中的痛苦挣扎。 GPU贸易从暴利狂欢跌落至薄利常态，存储涨价潮则将中小厂商逼至生死边缘。 退潮中，真正的价值才开始浮现。 我们历时一年，深入调研产业链内多家头部企业、访谈上百位产业专家、追踪国产GPU"四小龙"的上市征 程、见证推理芯片从"围攻英伟达"到"部分卡位"的艰难跃迁，记录存储产业链在成本暴涨下的集体承压。 这些深度内容不是为了渲染焦虑，而是为了在产业的褶皱中寻找真相： 当技术热潮的余温散去，什么才是 真正支撑中国AI算力走向规模化、生态化的底层因素 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 当Meta Platforms与英伟达（Nvidia）达成大规模 AI 系统交易时，通常意味着该公司此前的某些开 放硬件计划已无法满足紧迫的算力需求。这与项目延期不完全是一回事，但效果是一样的。提醒一 下，这类情况我们掌握的数据并不多，而如今这家社交网络巨头、AI 模型厂商与 AI 硬件巨头英伟 达之间宣布的巨额合作，已是第三起。 这笔交易远比 Meta 上一次与英伟达的合作规模更大，对英伟达而言价值至少数百亿美元，再加上原 始设计制造商将英伟达芯片集成到 Meta 系统中所能获得的额外收益。 在前两起案例中（几乎可以确定第三起新案例也是如此），一旦 AI 算力需求足够紧迫，Meta 便愿 意放弃自家开放计算项目（OCP）的设计方案。 在超大规模云厂商与大模型厂商中，Meta 的定位略有不同：它不只是为搜索加入 AI 能力，或是打 造能与 OpenAI、Anthropic 等抗衡的通用大模型，同时还高举开源大旗（至少目前是这样）。该公 司还运营着庞大的高性能集群集群，作为旗下各类服务的推荐引擎。这些系统需要CPU 与加速器紧 密耦合，让加速器能直接访问 CPU ...

公众号记得加星标⭐️，第一时间看推送不会错过。 英伟达本周已出售所持 ARM 的最后剩余股份，与几年前曾试图收购该公司的情景已相去甚远。 英伟达与 ARM 的合作，在当代 AI 基础设施建设中至关重要 —— 正是凭借 ARM 的 CPU 架构，英 伟达才得以推出 Grace Hopper、Blackwell 等系列重磅产品。更重要的是，ARM 还将在英伟达即将 推出的Vera CPU中扮演关键角色，这类处理器的重要性正在急剧提升。 据彭博社报道，根据最新提交给美国 SEC 的文件，英伟达已出售其持有的 ARM 剩余股份，价值约 1.4 亿美元。耐人寻味的是，此举恰好发生在ARM 在未来 AI 竞赛中的地位开始受到质疑的节点。 很多人尚未意识到：CPU 近期正迎来空前重要的地位提升。原因在于推理 workload，尤其是智能体 （agentic）相关负载—— 这类场景的重心正从 GPU 计算转向更依赖 CPU 的任务，例如工具调用、 API 请求、内存查找与调度逻辑。 这一转向已非常明显：英特尔、AMD 均表示，超大规模云厂商对其数据中心 CPU 需求暴增，背后 正是 CPU 整体市场（TAM）的高速扩张。与此 ...

当地时间周三，美联储公布的1月货币政策会议纪要未改变市场对于6月将启动降息的预期，另外当天美 国公布的多项经济数据表现强劲，叠加科技以及能源股走强，推动美国三大股指集体收涨。其中，道指 上涨0.26%，标普500指数上涨0.56%，纳指上涨0.78%。个股方面，受Meta计划大规模部署英伟达芯片 的消息提振，英伟达股价周三上涨超1.8%；美光科技股价当天上涨5.3%，其股票近期被知名对冲基金 经理大卫·泰珀旗下的阿帕卢萨管理公司增持。 存储芯片巨头西部数据计划出售闪迪部分股份 交易规模超30亿美元 另外，存储芯片巨头西部数据日前宣布，计划出售其所持有的部分闪迪股份，交易规模超30亿美元。分 析认为，此次股份出售将增强西部数据的资本配置灵活性。西部数据股价周三上涨近4.4%。 会议纪要显示美联储内部对于未来货币政策路径存分歧 美联储最新公布的会议纪要显示，其内部仍然对货币政策路径存在分歧，明确提到有联储官员讨论加息 的可能性，在就业下行风险有所缓和的背景下，通胀风险依旧是核心关注点。与会者普遍预计，通胀将 在年内回落，"但回落速度和时点仍不确定"，并预计关税对价格的影响将在年内逐步减弱。 贵金属方面，周三，投资 ...

Core Insights - Nvidia's CEO Jensen Huang announced a new chip set to be revealed at the upcoming GTC 2026 conference, which is expected to further solidify the company's leadership in AI infrastructure [1][4] - The GTC 2026 conference will take place on March 15 in San Jose, California, focusing on a new era of AI infrastructure competition [4] Product Development - The specific model of the new chips has not been disclosed, but speculation suggests they may belong to two major series: the Rubin series and the next-generation Feynman series [4] - The Rubin series includes six new chip designs that have already entered mass production, while the Feynman series is described as "revolutionary" and may utilize SRAM and 3D stacking technology [4] Market Positioning - Nvidia is adapting to quarterly changes in AI computing demands, with a focus on overcoming latency and memory bandwidth bottlenecks in its upcoming products [4] - The company emphasizes that extensive collaboration and investment are key to maintaining its leading position, with a strategy that spans the entire AI industry chain, including energy, semiconductors, and data centers [4]

Market Performance - The three major U.S. stock indices closed higher, with the Dow Jones Industrial Average rising by 129.47 points, an increase of 0.26% [1] - The Nasdaq Composite gained 175.25 points, reflecting a rise of 0.78% [1] - The S&P 500 index increased by 38.09 points, marking a 0.56% rise [1] Key Stocks Movement - The "Magnificent Seven" tech stocks collectively saw gains, with Amazon rising nearly 2% [1] - Nvidia increased by over 1% [1] - Microsoft rose by 0.69%, Facebook by 0.61%, Google by 0.43%, Apple by 0.18%, and Tesla by 0.14% [1] Oil Prices - International oil prices experienced a significant increase on the 18th, with light crude oil futures for March delivery rising by $2.86 to $65.19 per barrel, a gain of 4.59% [1] - April delivery Brent crude oil futures rose by $2.93 to $70.35 per barrel, reflecting a 4.35% increase [1]

Group 1 - Nvidia announced a "multi-generational" partnership with Meta to deploy millions of Nvidia's current and next-generation chips for AI training and inference data centers [1][4] - This collaboration marks Nvidia's entry into the CPU market, traditionally dominated by Intel and AMD, as Meta will utilize both Nvidia's GPUs and CPUs [1][4] - The partnership highlights Meta's increasing reliance on Nvidia, despite its own chip development and collaborations with competitors like AMD and Google [4] Group 2 - The collaboration includes Nvidia's networking equipment and confidential computing technology for WhatsApp's AI features, as well as the deployment of the next-generation Vera CPU to replace the current Grace CPU [4] - By choosing a single supplier for a full-stack chip solution, Meta aims to reduce operational complexity, while Nvidia's move indicates its ambitions in the CPU space [4] - The current demand for AI infrastructure is strong, suggesting that competitors like Intel and AMD may not experience significant declines in the short term despite increased competition [4]

过去三年， 英伟达（NVIDIA）已成为全球科技产业的核心驱动力 。2025财年，公司全年营收突破约 1300 亿美元 ，数据中心业务占比接近 90% ，毛利率维持在 70%–75% 区间。Blackwell GPU 和相关平台成为推动这轮 AI 基础设施升级的主要引擎。 这意味着下游行业——从云服务到企业 IT、从软件平台到终端应用——正在围绕这股算力浪潮重新定义其架构、投资与商业模式。 一、英伟达营收与算力扩张的下游拉动机制 数据中心业务的爆发式增长 英伟达数据中心业务是其增长核心： 2025 财年数据中心营收约 115.2 亿美元 ，同比增长 142% 。 Blackwell 架构单季曾贡献约 110 亿美元营收 ，是公司历史上最快的产品销售增长纪录。 在多个季度中，数据中心收入占比一度超过 88%~90% 。 Blackwell GPU 贡献了数据中心业务约 70% 的计算收入 。 这些数字说明，英伟达的增长几乎完全由 AI 计算需求驱动，而这种需求本质上来自下游企业对大规模模型训练、实时推理、AI 工厂和智能应用的投 入。 下游需求爆发背后的三大趋势 1. 算力成为生产力基础设施 下游企业不再把 A ...

Core Insights - H&H International Investment, managed by investor Duan Yongping, has made significant adjustments to its portfolio in Q4 2025, notably reducing its stake in Apple while increasing investments in AI-related stocks [1][2]. Portfolio Adjustments - In Q4 2025, Duan Yongping reduced his holdings in several stocks, including Apple (AAPL) by 10.12%, Occidental Petroleum (OXY) by 1.27%, Alibaba (BABA) by 1.24%, ASML Holding (ASML) by 0.47%, and Walt Disney (DIS) by 0.1% [2]. - Despite the reduction, Apple remains the largest holding in Duan's portfolio, with a market value of $8.797 billion, accounting for 50.3% of the total portfolio, although he sold 2.4706 million shares [3]. Investment in AI - Duan Yongping significantly increased his position in NVIDIA, acquiring 6.6393 million shares, marking a 1110.62% increase, with a total market value of $1.35 billion, making it the third-largest holding at 7.72% of the portfolio [3]. - Additionally, Duan initiated positions in three AI-related companies: CoreWeave (299,900 shares), Credo (141,300 shares), and Tempus (110,000 shares), targeting key areas in the AI supply chain [3]. Strategic Outlook - In early January, Duan expressed a strong interest in learning about AI, indicating a belief that the technology will significantly transform the world in the coming years, potentially more than in the past decades [3].