黄韦涵S1060523070003（证券投资咨询）HUANGWEIHAN235@pingan.com.cn 证券研究报告 OpenAI发布GPT5.4，英伟达投资北美光 通信公司 计算机行业 强于大市（维持） 刘云坤S1060525120004（证券投资咨询）LIUYUNKUN518@pingan.com.cn 平安证券研究所计算机团队 分析师： 闫磊S1060517070006（证券投资咨询）YANLEI511@pingan.com.cn 2026年03月08日 请务必阅读正文后免责条款 核心摘要 行业要闻及简评：1）OpenAI发布GPT-5.4，专为复杂任务中追求极致性能而设计。2026年3月5日，OpenAI发布GPT-5.4模型，GPT-5.4将 OpenAI在推理、编程及智能体工作流领域的顶尖性能集于一身，整合了GPT-5.3-Codex行业领先的编程能力，同时优化了模型在各类工 具、软件环境以及专业任务（电子表格、演示文稿和文档等）中的表现，GPT-5.4无需大量往复沟通即可准确高效的完成复杂工作并交付 结果。根据ArtificialAnalysis数据，在大模型性能排名方面，GPT-5.4以5 ...

Core Viewpoint - The "Magnificent Seven" stocks, primarily in technology, have significantly contributed to the S&P 500's growth, with a focus on AI, cloud computing, and advertising [1][2]. Group 1: Alphabet - Alphabet holds a dominant position in the search engine market with approximately 90% global market share and reported $82.3 billion in advertising revenue for Q4, a 14% increase year-over-year [4][7]. - Google Cloud revenue surged 48% in Q4 to $17.7 billion, with an annual run rate exceeding $70 billion, highlighting the growing importance of cloud computing for AI applications [7][5]. Group 2: Apple - Apple has a more conservative approach to AI spending, committing $12.7 billion in 2025 and planning $600 billion over four years, focusing on domestic manufacturing rather than data centers [9]. - The company continues to excel in hardware sales, including iPhones and MacBooks, and its Services segment generated $30 billion in revenue in Q1 of fiscal 2026, a 14% increase from the previous year [11]. Group 3: Nvidia - Nvidia is positioned to benefit from the substantial investments in AI infrastructure, with its GPUs being the preferred choice for AI training and operations [12]. - The company reported $68.1 billion in revenue for Q4 of fiscal 2026, a 73% increase year-over-year, with the data center segment contributing $62.3 billion [13][15].

Group 1: Market Overview - The stock market remains volatile due to the U.S.-Iran conflict, fears of AI disruption, and concerns about the sustainability of the AI boom [1] - Investors are advised to consider recommendations from top Wall Street analysts for long-term stock picks [1] Group 2: Nvidia - Nvidia (NVDA) is highlighted as a top stock pick, with UBS analyst Timothy Arcuri maintaining a buy rating and a price target of $245, while TipRanks' AI Analyst has a target of $230 [3][4] - Arcuri is optimistic about Nvidia's networking growth, stating the company aims to be the largest global networking player by surpassing the combined revenue of other suppliers by the end of 2026 [4] - Nvidia expects a long-term gross margin target of 75%, with improvements in compute performance and total cost of ownership (TCO) to sustain margins [5] - The company has a $550 billion backlog for Blackwell and Rubin but will not update this figure as focus shifts to 2027 buildouts [6] - Nvidia's confidence in compute buildouts is supported by the strong financial positions of hyperscalers, with management seeing expanded financing mechanisms [7] Group 3: Palo Alto Networks - Palo Alto Networks (PANW) is another favored stock, with TD Cowen analyst Shaul Eyal reaffirming a buy rating and a price target of $255, while TipRanks' AI Analyst has a target of $181 [9] - Eyal notes strong demand for unified security platforms, particularly in network security and SASE [10] - The growing adoption of agentic AI is seen as a potential catalyst for Palo Alto, as companies may prefer consolidated security tools [11] - Customers who adopted SASE during the pandemic are reevaluating their vendors, allowing Palo Alto to capture market share [12] - Eyal emphasizes the importance of opportunistic acquisitions to achieve $20 billion in next-generation security annual recurring revenue by FY30 [13] Group 4: Micron Technology - Micron Technology (MU) is set to announce its fiscal second-quarter results, benefiting from AI-driven demand and rising memory prices [14] - Stifel analyst Brian Chin has raised his price target for Micron to $550 from $360, while TipRanks' AI Analyst has a target of $497 [14] - Chin highlights that memory pricing is exceeding expectations due to a persistent supply-demand gap [15] - The analyst sees significant growth potential in Micron's DDR5 RDIMM product, expecting gross margins to exceed 80% [15] - Chin believes that memory pricing strength will continue, with consensus expectations being too low for future quarters [16] - Revised estimates reflect growth driven by average selling price increases across Micron's segments [17]

Core Viewpoint - Nvidia's investment thesis remains strong despite a recent 4% decline in stock price, as the company is positioning itself not just as a GPU seller but as a creator of the operating system for the AI economy [1] Company Analysis - Nvidia is expanding its role in the AI sector beyond hardware, focusing on developing the foundational software that will drive the AI economy [1] Investment Strategy - The investment approach emphasizes identifying high-potential opportunities with significant upside potential, aiming for a return of 3-5 times the initial investment while managing downside risks [2] - The strategy includes leveraging market inefficiencies and contrarian insights to maximize long-term compounding and protect against capital impairment [2] - A 2-3 year investment horizon is adopted to endure market volatility, emphasizing patience, discipline, and intelligent capital allocation for superior returns [2]

证券研究报告 | 行业周报 gszqdatemark 2026 03 08 年 月 日 通信 GTC 大会临近，算力再预热 英伟达 GTC 2026 大会将于 3 月 16 日-19 日正式举办，作为全球 AI 算力 领域的风向标，本届大会将集中展示下一代 GPU 架构、CPO 共封装光学、 电源液冷等前沿突破，有望成为 AI 算力产业链新一轮技术迭代的起点， 再次催化算力板块热度。 【新一代 GPURubin 架构：平台、Feynman 预告与 LPU 推理芯片】 黄仁勋称将于 GTC 大会展示"世界前所未见"的全新芯片，业界推测本 次大会预计正式揭晓Rubin及下一代Feynman架构GPU的核心技术细节， 并推出整合了 LPU 技术的全新推理芯片。 【算力基础设施变革：光通信、电源、液冷】 随着 GPU 单芯片功率迈入千瓦时代，算力基础设施正经历颠覆性升级， 本次 GTC 预计将集中展示 CPO 交换机、电源架构、液冷散热等环节的技 术突破与商业化落地进展。 行业走势 -20% 4% 28% 52% 76% 100% 2025-03 2025-07 2025-11 2026-03 通信 沪深300 作者 ...

When Jeff Bezos said that one breakthrough technology would shape Amazon’s destiny, even Wall Street’s biggest analysts were caught off guard. Fast forward a year and Amazon’s new CEO Andy Jassy described generative AI as a “once-in-a-lifetime” technology that is already being used across Amazon to reinvent customer experiences. At the 8th Future Investment Initiative conference, Elon Musk predicted that by 2040 there would be at least 10 billion humanoid robots, with each priced between $20,000 and $25,000 ...

When Jeff Bezos said that one breakthrough technology would shape Amazon’s destiny, even Wall Street’s biggest analysts were caught off guard. Fast forward a year and Amazon’s new CEO Andy Jassy described generative AI as a “once-in-a-lifetime” technology that is already being used across Amazon to reinvent customer experiences. At the 8th Future Investment Initiative conference, Elon Musk predicted that by 2040 there would be at least 10 billion humanoid robots, with each priced between $20,000 and $25,000 ...

When Jeff Bezos said that one breakthrough technology would shape Amazon’s destiny, even Wall Street’s biggest analysts were caught off guard. Fast forward a year and Amazon’s new CEO Andy Jassy described generative AI as a “once-in-a-lifetime” technology that is already being used across Amazon to reinvent customer experiences. At the 8th Future Investment Initiative conference, Elon Musk predicted that by 2040 there would be at least 10 billion humanoid robots, with each priced between $20,000 and $25,000 ...

When Jeff Bezos said that one breakthrough technology would shape Amazon’s destiny, even Wall Street’s biggest analysts were caught off guard. Fast forward a year and Amazon’s new CEO Andy Jassy described generative AI as a “once-in-a-lifetime” technology that is already being used across Amazon to reinvent customer experiences. At the 8th Future Investment Initiative conference, Elon Musk predicted that by 2040 there would be at least 10 billion humanoid robots, with each priced between $20,000 and $25,000 ...

When Jeff Bezos said that one breakthrough technology would shape Amazon’s destiny, even Wall Street’s biggest analysts were caught off guard. Fast forward a year and Amazon’s new CEO Andy Jassy described generative AI as a “once-in-a-lifetime” technology that is already being used across Amazon to reinvent customer experiences. At the 8th Future Investment Initiative conference, Elon Musk predicted that by 2040 there would be at least 10 billion humanoid robots, with each priced between $20,000 and $25,000 ...