2026.02. 10 本文字数：2009，阅读时长大约4分钟 作者 | 第一财经 冯迪凡 有报道称，美国计划在下一轮芯片关税中豁免亚马逊、谷歌和微软等美国科技巨头，原因是这些企业 正竞相建设推动人工智能（AI）热潮的数据中心。知情人士透露，美国商务部计划为美国本土超大规 模企业提供的关税豁免将与台积电（TSMC）的投资承诺挂钩。 哈佛大学肯尼迪政府学院教授福尔曼（Jason Furman）在近期的一场研讨会上对第一财经记者表 示，一方面，特朗普政府以关税威胁为手段，推动更多美国制造业的发展，另一方面，从截至目前的 发展看，特朗普政府取消威胁、增加关税豁免的案例比收紧关税的案例要多。 牛津经济研究院也在近期针对全球商业人士的风险调查报告中发现，尽管存在地缘政治风险的担忧， 但全球商业界对美国贸易政策全球影响的忧虑已有所缓解，总体而言，企业并未察觉全球增长前景出 现明显变化。 分配豁免额度？ 据新华社报道，美国白宫1月14日发布声明说，美国将从15日起对部分进口半导体、半导体制造设 备和衍生品加征25%进口从价关税。根据白宫公布的事实清单，英伟达公司H200芯片和超威半导体 声明称，此次加征关税依据的是《196 ...

摩根大通指出，春节AI推广战背后，是用户习惯向聊天机器人迁移，推动Token消耗进入长期高增长。 预测中国Token消耗量年复合增速将达330%。投资焦点应放在 "二阶赢家"：直接受益的AI基础设施商 及广告业务相关公司。 摩根大通指出，虽然春节期间各大科技巨头掀起了激烈的AI应用推广战，但这只是表面现象。透 过"谁将赢得AI应用大战"的短期博弈，市场应关注更深层的趋势：消费者获取信息和消费内容的方式 正在发生结构性改变，这将推动Token（令牌）消耗量开启一个长达多年的高速增长期。 在2026年春节，腾讯、阿里巴巴和百度将通过数十亿元人民币的补贴争夺用户心智。据追风交易台， 2月6日，摩根大通在最新发布的中国互联网与人工智能行业研报中分析认为，这些大型互联网平台投 入真金白银进行推广，是为加速用户从传统搜索转向聊天机器人，并培养智能体商业习惯。这一进程 实际上推升了推理量，从而加快了Token的消耗速度。目前的GPU算力需求过剩或供不应求的状况将 在未来几年持续。 该行预测，2025年至2030年，中国Token消耗量的年复合增长率将高达330%左右。基于此，摩根大通 认为目前判断AI应用端的胜负为时尚早， ...

Core Argument - The ongoing debate between investing in Advanced Micro Devices (AMD) versus Nvidia (NVDA) highlights Nvidia's significant growth advantage since 2023, with a rise of nearly 1,110% compared to AMD's 209% gain [1] Group 1: Performance Comparison - Since 2025, AMD has outperformed Nvidia with a 65% increase in stock price versus Nvidia's 30% [2] - Nvidia has a strong position in the AI market, primarily through its superior graphics processing units (GPUs), which are essential for processing complex AI workloads [2][3] - AMD has recently improved its product offerings, making it a more competitive option, especially as its products are more affordable than Nvidia's [3] Group 2: Revenue Growth - AMD's data center revenue grew by 39% year over year in Q4, while Nvidia's Q3 data center growth was significantly higher at 66%, with expectations for Q4 growth around 67% [4][5] - Nvidia's overall company growth rate in Q3 was 62%, indicating a robust performance compared to AMD's smaller growth [5] Group 3: Valuation Analysis - Despite AMD's slower growth, it trades at a higher valuation of 30 times forward earnings, while Nvidia trades at 23 times forward earnings, suggesting a price discrepancy [5] - The higher valuation of AMD raises questions about its attractiveness as an investment compared to Nvidia, which is seen as the more favorable option for future performance [7]

Core Viewpoint - Software stocks are expected to rebound from historic lows as the market has overly pessimistic expectations regarding AI's short-term disruption of the software industry, according to JPMorgan strategists [1][5]. Group 1: Market Sentiment and Stock Performance - The software sector has fallen to its lowest level since the market turmoil in April last year [4]. - Concerns over new AI tools potentially disrupting traditional Software as a Service (SaaS) business models have led to sustained pressure on U.S. software stocks [1][5]. - The recent sell-off did not differentiate between companies with AI partnerships or proprietary data assets, affecting nearly all related software companies equally [1]. Group 2: Company Resilience and Long-term Outlook - Companies like Microsoft and CrowdStrike are highlighted as resilient players in the AI space, likely to benefit from AI-enhanced workflow efficiencies [5]. - The high switching costs and long-term contracts in enterprise software provide a buffer against short-term disruptions [5]. - The long-term fate of traditional software companies in the face of AI remains uncertain, but current market pessimism appears to be an overreaction [5]. Group 3: Analyst Perspectives - Analysts from Morgan Stanley believe that U.S. tech stocks still have room for further gains, and the decline in software stocks has created an attractive entry point [5]. - Wedbush analysts argue that the market's reaction to AI risks is excessive, suggesting that the current sell-off implies an extreme assumption of widespread AI disruption, which is not feasible [6]. - The caution of enterprise clients regarding AI migration is emphasized, as many are reluctant to expose core data to immature new platforms [6]. Group 4: AI Integration and Market Dynamics - The narrative that AI will replace entire enterprise software stacks is overly simplistic; the value density of enterprise software lies in proprietary data and compliance structures [8][9]. - AI is more likely to integrate as embedded tools within existing software platforms rather than completely replacing them [6][9]. - The current sell-off reflects a market response to the question of how much profit pools in SaaS will be redistributed due to AI [10]. Group 5: Future Indicators and Investment Strategy - The rebound in software stocks may depend on two hard indicators: the speed of real deployment and payment expansion by enterprises, and the elasticity of SaaS companies' AI-related product revenues [10]. - Companies with strong data assets and solid fundamentals, such as Microsoft, MongoDB, Snowflake, Palantir, and SAP, are likely to experience a robust rebound post-panic [10].

Investment Rating - The report maintains a "Strong Buy" rating for the domestic computing power industry chain, indicating a positive outlook for investment opportunities in this sector [2]. Core Insights - The report emphasizes that AI inference is leading innovation, with a focus on demand-driven Opex-related areas, specifically storage, power, ASIC, and supernodes [2][28]. - It highlights the expected growth in the domestic storage industry, driven by new products from Yangtze Memory Technologies and Changxin Memory Technologies, alongside a rapid increase in demand for SSDs and HBM [2][29]. - The report anticipates a significant expansion year for storage production, suggesting investors pay close attention to the overall opportunities within the domestic storage industry chain [29]. - The power industry is also highlighted, with a focus on new technologies on both the supply and demand sides, including companies like Sanhua Group and Zhongfu Circuit [30]. - The ASIC segment is expected to see an increase in market share, with a focus on major CSP manufacturers [30]. - The report predicts an evolution in cabinet models, with growth in demand for high-speed interconnects, cabinet OEM, liquid cooling, and PCB [30]. Summary by Sections Market Review - The Shanghai Composite Index fell by 1.27%, while the Shenzhen Component Index dropped by 2.11%, and the overall Shenwan Electronics Index decreased by 5.23% [1][13]. - Year-to-date, the Shenwan Electronics Index has declined by 4.73%, ranking 14th out of 31 sectors [1][13]. Weekly Focus - Nvidia's upcoming Arm architecture-based N1X + N1 processor is discussed, indicating a strategic move into the AI PC and laptop market, competing directly with AMD and Intel [26][27]. - The report notes that Nvidia's new processors will utilize TSMC's 3nm process technology, enhancing performance while maintaining low power consumption [26][27]. Related Research - Previous reports have consistently highlighted the positive outlook for the domestic computing power industry chain, including price increases in semiconductor packaging and storage chips [4][6][5].

然而，随着硬件巨头Intel向"全栈AI公司"快速转型，这种绝对垄断正在被打破。 早在2019年，Intel就发布了oneAPI 跨架构编程模型，旨在让代码在 CPU、GPU、NPU 之间通用。这意味着开发 者用一套代码即可调用 Intel 的所有算力，降低了迁移成本。 来源：快科技官方 一、前言：当前最具性价比的96GB/192GB AI推理卡 oneAPI还允许开发者将原本仅能NVIDIA CUDA环境下运行的代码，通过其迁移工具（SYCLomatic）快速转换到 Intel硬件上，为Arc系列显卡运行主流大模型打下了坚实的软件基础。 凭借深耕多年的CUDA护城河，NVIDIA在AI领域一度拥有"定价权"，这也让这家公司的GPU及相关产品的售价逐 渐脱离普通的消费者。 去年，Intel发布了基于第二代Xe2架构（Battlemage）的专业级显卡—Intel Arc Pro B60。随后，以Maxsun（铭 瑄）、SPARKLE（撼与）、GUNNIR（蓝戟）为代表的核心伙伴正式将其推向全球市场，直指高性能AI推理领 域。 Intel Arc Pro B60与此前发布的消费级Intel Arc B580一样 ...

Investment Rating - The report maintains a "Strong Buy" rating for the domestic computing power industry chain, indicating a positive outlook for investment opportunities in this sector [2]. Core Insights - The report emphasizes that AI inference is leading innovation, with a focus on demand-driven Opex-related areas, specifically storage, power, ASIC, and supernodes [2][28]. - It highlights the expected growth in the storage sector due to new product breakthroughs from Yangtze Memory Technologies and Changxin Memory Technologies, predicting a significant expansion year for domestic storage capacity [29]. - The report also notes the anticipated increase in ASIC market share and the evolution of cabinet models, with a positive outlook on high-speed interconnects, cabinet manufacturing, liquid cooling, and PCB demand [30]. Summary by Sections Market Review - The Shanghai Composite Index fell by 1.27%, while the Shenzhen Component Index dropped by 2.11%. The Shenwan Electronics Index decreased by 5.23%, ranking 29th among 31 Shenwan industries [1][13]. Weekly Focus - Nvidia is set to launch new Arm architecture-based chips, specifically the N1X and N1 processors, designed for AI computing, which will compete directly with AMD and Intel in the laptop market [26][27]. Storage Sector - The report identifies key players in the NAND & DRAM semiconductor industry, including Zhongwei Company, Tuojing Technology, and Anji Technology, and suggests focusing on the overall opportunities within the domestic storage industry chain [29]. Power Sector - The report recommends attention to new technologies in both the power generation and consumption sides, highlighting companies like Sanhuan Group and Zhongfu Circuit [30]. ASIC and Supernodes - The report anticipates an increase in ASIC's market share and suggests monitoring major CSP manufacturers, while also noting the expected growth in demand for high-speed interconnects and related technologies [30]. Domestic Computing Power Chain - The report underscores the improvement in domestic advanced process yields and capacity, which is expected to enhance the supply side of domestic computing power chips, alongside a clearer commercialization model for domestic CSP manufacturers [30][31].

Nvidia stock is now hovering near its highest levels this year so far. ...

Core View - NVIDIA Corporation (NASDAQ:NVDA) is recognized as a leading player in the AI sector, with Bank of America maintaining a "Buy" rating and a price target of $275, reflecting confidence in its strong position in the AI compute and networking markets [1][5]. Financial Analysis - The $275 price target is based on a forward valuation of 28x CY27E PE excluding cash, which falls within NVIDIA's historical PE range of 25x-56x, justified by its leading market share in rapidly growing AI sectors [2]. - Wall Street analysts have a consensus "Buy" rating for NVIDIA, with an average price target of $250 indicating a potential upside of 30.23%. The highest target of $432.78 suggests a significant upside of 125.44% [2]. Company Overview - NVIDIA specializes in AI-driven solutions, providing platforms for data centers, self-driving cars, robotics, and cloud services, highlighting its diverse applications in the technology landscape [3].

Group 1 - The current Trump administration's "mercantilist" approach is evident in its chip tariff strategy, which is more about negotiating with chip manufacturers than a definitive goal [1] - Reports indicate that the U.S. plans to exempt major tech companies like Amazon, Google, and Microsoft from upcoming chip tariffs due to their investments in AI data centers, with exemptions linked to TSMC's investment commitments [1][4] - The U.S. government has announced a 25% tariff on certain imported semiconductors and related equipment, citing national security threats, while noting that only 10% of the chips needed are produced domestically [3] Group 2 - The new plan allows TSMC to allocate exemption quotas for its U.S. clients based on their investment levels in the U.S., aiming to encourage more domestic chip manufacturing [4][5] - TSMC has committed to investing $165 billion in expanding its U.S. production capacity, with the U.S. government closely monitoring the integrity of tariff and rebate measures [5] - Major U.S. tech companies are projected to invest $700 billion in AI infrastructure by 2026, nearly double last year's total, with significant portions allocated for data centers and advanced chips [6] Group 3 - The AI investment surge is reshaping the economic landscape, leading to increased demand for memory chips and computing components, which is causing shortages in parts needed for traditional consumer electronics [6] - European investments in AI, such as the €20 billion announced by France, are significantly lower than the investments by U.S. tech giants, highlighting challenges in financing AI development in Europe [7] - The reallocation of labor and materials in the U.S. to support large tech projects is driving up construction costs and exacerbating housing crises in various cities [7]