Core Viewpoint - There is a growing trend of comparing Advanced Micro Devices (AMD) and Nvidia Corp. (NVDA), which raises questions about the validity of such comparisons given the distinct differences between the two companies [1]. Group 1 - AMD and Nvidia are often compared, but the rationale behind these comparisons is unclear [1]. - The author emphasizes the need for a deeper understanding of each company's unique market position and business model [1].

汇丰：将AMD(AMD.O)目标股价从300美元上调至335美元。 本文源自：金融界AI电报 ...

智通财经APP获悉，东吴证券发布研报称，AIAgent落地速度正逐渐加速，其带来的沙箱创建、负载卸 载等需求有望大幅拉动CPU的需求。但同时，全球算力供应链产能紧张，多个环节涨价，使得CPU制造 成本增加。两方面影响下，该行认为CPU产业有望迎来大周期。 东吴证券主要观点如下： 风险提示：Agent落地进展不及预期；产能供应问题缓解。 DRAM生产转向HBM，消耗更多晶圆；与NAND需求攀升、交货期延长，库存告急，挤占CPU晶圆材 料供给。CPU部件PCB应用及加工材质的转变，使得钻针使用寿命缩短，消耗量暴增；CCL采用的树脂 体系、玻纤布与铜箔匹配复杂，新进入者良率提升缓慢与客户认证周期长，导致有效产能释放缓慢；二 者纷纷涨价，带动CPU价格上涨。 AI推理和Agent发展迅速，拉动高端多核CPU需求 CPU负载正从"人类节奏"转向"机器节奏"。Agentic AI远不是单次推理，而是动态推理+多步决策+外部 工具调用的循环，这比传统大模型调用更耗资源、负载更复杂、成本更高。这种资源调用增长，加之为 了安全防范而产生的高频沙箱隔离开销，使得CPU资源消耗呈现指数级放大。Deepseek提出Engram模 块 ...

J P M O R G A N North America Equity Research 29 January 2026 Semiconductors & Semiconductor Capital Equipment / IT Hardware Harlan Sur AC (1-415) 315-6700 harlan.sur@jpmorgan.com Peter K Peng (1-415) 315-8327 peter.k.peng@jpmchase.com Mayur Ramdhani (1-212) 622-1664 mayur.ramdhani@jpmorgan.com J.P. Morgan Securities LLC See page 2 for analyst certification and important disclosures. J.P. Morgan does and seeks to do business with companies covered in its research reports. As a result, investors should be aw ...

Accessible version Global Memory Tech Weekly theme: memory industry model update, softening DRAM spot, LGE upside Price Objective Change Global DRAM/NAND sales forecasts raised by 25% We updated our global memory+HBM industry model following Samsung Electronics' and SK Hynix's 4Q25 earnings results. Key changes to our 2026 estimates vs earlier include: 1) 20%+ higher ASPs (for both DRAM and NAND); 2) slightly higher bit growth; 3) capex increase, largely driven by HBM and infrastructure investments (shell f ...

公众号记得加星标⭐️，第一时间看推送不会错过。 如果时光倒流十年，当时的AMD并非如今的AMD。彼时，Zen架构尚未出货，EPYC仍在研发路线图 上，而AMD能够同时在CPU、GPU和AI基础设施领域占据高端市场，这在当时还远非主流观点。过 去十年见证了半导体行业最激进的技术革新之一，涵盖核心架构、芯片策略、封装、代工厂合作，以 及如今的机架级系统设计。这一转型并非一蹴而就，而是需要长期的基础设施投入、多代产品的执行 力，以及在成果显现前数年就勇于承担风险的魄力。 Ian Cutress：我们显然进行过很多讨论。我和架构师们交流过，每次我们谈到芯片设计时，都会有 一个两到三年的周期，但之后还需要五到七年的时间进行探索，并与代工厂合作伙伴合作等等。回过 头来看，你认为Zen架构前两代产品的成功，相对于公司未来的发展方向而言，究竟有多少"烟幕 弹"（smoke screen）效应？ Mark Papermaster： 想想我们第一代Zen架构的开发历程——我们在2017年发布了Zen，它被应用 在EPYC服务器和Ryzen PC上。你称之为"烟幕弹"，但我并不认同这种说法。我认为那是我们前两代 Zen架构的试验期 ...

Core Viewpoint - AMD is optimistic about its growth prospects for 2026, aiming for a significant comeback in the AI accelerator market despite competition from Nvidia and Broadcom [1][2][3]. Group 1: Company Overview - AMD has a diversified chip business, with over 43% of its revenue coming from OEM and gaming, and approximately 47% from data centers, while its embedded processor division contributes about 10% [3]. - The company is less reliant on AI-related sales compared to Nvidia, which may provide some insulation against potential downturns in AI infrastructure spending [4]. Group 2: Growth Projections - AMD's management anticipates a 60% compound annual growth rate (CAGR) for its data center division over the next five years, while other divisions are expected to grow at 10% CAGR [6]. - In Q3, AMD's data center revenue increased by 22% year over year, indicating the need for substantial growth to meet the 60% CAGR target [7]. Group 3: Stock Performance and Valuation - If AMD's stock were to double, it would reach approximately $500 per share, necessitating an earnings per share (EPS) of $10 based on a price-to-earnings (P/E) ratio of 50 [7][8]. - Current analyst EPS estimates for 2026 range from $5.36 to $8.02, which are below the required EPS for a doubled stock price [8]. - AMD's profit margin is currently low compared to Nvidia's, but doubling the profit margin could enhance the likelihood of the stock doubling [10]. Group 4: Competitive Landscape - AMD has yet to demonstrate effective competition with Nvidia in the GPU space, which raises skepticism about its ability to achieve the projected growth [11].

Core Insights - Nvidia has maintained a dominant position in the high-performance GPU market, particularly in the generative AI sector, due to its pioneering role in GPU design [1] - Advanced Micro Devices (AMD) is emerging as a significant competitor in the GPU landscape, with ambitions to transform its market presence by 2026 [2] Group 1: AMD's Competitive Position - AMD's chips are increasingly being adopted by hyperscalers, indicating a growing validation of its products against Nvidia's offerings [4] - Major tech companies like Microsoft, Meta Platforms, Oracle, and OpenAI are integrating AMD's Instinct accelerators alongside Nvidia GPUs, showcasing AMD's credibility in handling large-scale AI applications [5] - AMD's chips are utilized for both training and inference workloads, potentially offering better unit economics and positioning AMD as a lower-cost alternative to Nvidia [6] Group 2: Technological and Strategic Advantages - AMD's ROCm software platform provides developers with greater control compared to Nvidia's CUDA ecosystem, which may serve as a competitive advantage for big tech [7] - The flexibility of AMD's offerings could enhance customers' negotiating leverage over suppliers, indicating a potential disruption to Nvidia's market dominance [8] - AMD's ability to secure significant contracts with hyperscalers may lead to sustained deal flow, reinforcing its position in the market [9]

Group 1: Job Market Insights - The U.S. jobs report for January is anticipated, with December showing signs of labor market weakening, as only 50,000 jobs were added, lower than economists' expectations [3] - Federal Reserve officials are monitoring the labor market closely after keeping interest rates unchanged, citing elevated inflation risks despite a slowdown in hiring [4] Group 2: Earnings Reports Focus - Key earnings reports are expected from major tech firms such as Alphabet and Amazon, with Alphabet recently surpassing $100 billion in revenue [5] - Amazon has also shown strong revenue growth in the previous quarter but announced another round of layoffs [5] - Reports from Advanced Micro Devices indicate brisk sales of data center chips, contributing to positive analyst sentiment, although concerns about inflated valuations for top tech companies persist [6] Group 3: Sector-Specific Earnings - Disney's earnings report will provide insights into its direct-to-consumer segment, which grew 8% in the last quarter but fell short of expectations [7] - Pharmaceutical firms are also in focus, with Eli Lilly's shares rising due to optimism over its weight loss drugs, alongside earnings reports from competitors like Novo Nordisk, Amgen, Merck, AbbVie, and Novartis [7]

Core Viewpoint - Many AI stocks are perceived as expensive, but there are several undervalued options with significant growth potential [1][2]. Group 1: Advanced Micro Devices (AMD) - AMD's forward price-to-earnings ratio is 39.7, and its trailing P/E ratio is 131.6, indicating a high valuation at first glance [3][4]. - The stock's PEG ratio is notably low at 0.5, suggesting it is one of the cheapest AI stocks available [3]. - AMD anticipates revenue from AI data centers to grow at a compound annual growth rate (CAGR) of over 80% in the next three to five years [5]. - The company is gaining market share in server CPUs and is making progress in the GPU market with its Instinct MI350 Series [6]. Group 2: Micron Technology - Micron's PEG ratio is just below 0.7, and shares trade at 12.3 times forward earnings, indicating it is not an expensive commodity [7]. - The company has secured contracts for its entire 2026 high-bandwidth memory (HBM) supply, reflecting strong demand and supply constraints [9]. - Micron expects the total addressable market for HBM to reach $100 billion by 2028, with a CAGR of approximately 40% [9]. Group 3: Nvidia - Nvidia's PEG ratio is 0.7, and it is expected to experience strong growth over the next five years, making its current valuation more justifiable [10]. - The company projects annual AI infrastructure spending to reach $3 trillion to $4 trillion by the end of the decade, driven by emerging technologies [12]. - Nvidia's Blackwell GPUs are currently the most powerful AI chips, with plans to launch even more advanced Rubin GPUs later this year [13].