超越英伟达的依然是TA自己—— 2026财年，英伟达最后一份季度财报出炉。不出意料，又是连破三个纪录： 单季度营收达681亿美元（约4678亿元），同比增长73%，环比增长20%。 数据中心季度营收623亿美元（约4280亿元），同比增长75%，环比增长22%。 公司全年营收2159亿美元（约14831亿元），同比增长65%。 这份成绩可以说让资本市场吃了"定心丸"，"AI鬼故事"的阴霾一扫而空……财报后，英伟达盘后股价一度涨超3.7%。 英伟达最新季度财报 从财务指标入手，"猛涨"依旧是英伟达财报的基调。 2026财年（即2025年）第四季度，公司营业收入为681亿美元（约4678亿元），同比增长73%，环比增长20%。 2026财年全年营收2159亿美元（约14831亿元），同比增长65%。 透过业务基本盘进一步感知，还是从数据中心、游戏、专业可视化，以及汽车四个方面去展开。 英伟达业务的绝对核心数据中心，第四财季营收623亿美元，同比增长75%，环比增长22%；全年营收1937亿美元，同比增长68%。 数据中心业务的增长，主要来自平台转型在背后推动，Blackwell架构全面放量，推理算力需求呈现指数级 ...

英伟达交出远超预期的业绩与指引，却没能再度点燃股价，市场对AI需求可持续性与供给约束的疑虑，正在压过"数字本身"的亮眼。 英伟达给出的财年第一季度营收指引约为780亿美元，明显高于华尔街平均预期的728亿美元，但公司在分析师电话会期间股价一度下跌约1.5%，到当晚整 体几乎持平。 $${\frac{\mp\pm}{P}}\langle\mp j\rangle\pm$$ 加上本次财报，英伟达已经连续14个季度营收超出预期。分析指出，投资者对英伟达的"固有高预期"已抬至新高度，单纯的"再一次击败预期"已很难带来额 外溢价，市场更想要的是对AI繁荣能否持续的更强保证。 管理层试图正面回应这些担忧。黄仁勋强调客户正在用新增算力赚钱，投资将维持在高位；CFO Colette Kress则称公司已锁定关键供给并将出货承诺延伸到 更长周期。 指引很强，市场仍觉得"差一口气" 英伟达第一季度业绩指引营收约780亿美元。尽管这一数字超过分析师一致预期，但也低于部分分析师接近800亿美元的更激进预测，这使得"超预期"在边际 上不再足够惊艳。 增长主引擎仍是数据中心业务，该部门当季营收为623亿美元，高于分析师平均预期的604亿美 ...

炒股就看金麒麟分析师研报，权威，专业，及时，全面，助您挖掘潜力主题机会！ 来源：华尔街见闻 在最近Anthropic一系列产品发布和Citrini的"末日报告"加剧投资者恐慌之际，人工智能（AI）热潮经受 住直接检验，英伟达带来"炸场"的业绩，证明AI产生的需求持续强劲。 美东时间25日周三，英伟达公布，在截至2026年1月31日的公司2026财年第四财季（"四季度"），营收 达创纪录的681亿美元，同比大增约七成，贡献超九成营收的核心业务数据中心也创单季收入新高，均 较分析师预期水平高逾3%。 四季度英伟达的盈利同样强劲。非GAAP口径下，调整后的每股收益（EPS）同比增长超八成，较分析 师预期高约5.9%，毛利率也超预期攀升至75.2%，创一年半来新高。 更令投资者鼓舞的是，英伟达的2027财年第一财季（一季度）指引也强于预期。营收料再创新高，指引 区间中值较分析师预期中值高7.1%，甚至比买方的乐观预期高4%，同比增速较四季度加快至近77%。 英伟达指出，该指引不含中国市场的数据中心计算收入。 本周三的业绩电话会上，英伟达CEO黄仁勋还上调了此前作出的芯片收入预期，称我们将超越5000亿美 元的目标。供 ...

大家好，继续关注海外市场今晚的表现。 美股上涨 2月25日晚间，美股三大指数上涨，科技股涨幅居前，道琼斯指数微涨，纳指涨近1%，标普指数涨约0.4%。 今晚，全世界的投资人都在等待英伟达收盘之后发布的最新财报，英伟达股价涨超1%，投资人正重新评估科技股的高估值，并对超大规 模云服务商高昂的AI资本支出愈发怀疑。在人工智能技术焦虑蔓延之际，市场希望看到AI前景依然稳固的证据。 稳定币第一股Circle股价飙升超过20%。此前该公司表示，在其稳定币的强劲需求推动下，尽管第四季度数字资产市场处于低迷期，但公 司的利润和营收依然实现了增长。 在过去几年引领市场走高后，英伟达股价近几个月陷入停滞，自去年第四季度至今仅上涨了3.8%，原因是投资者开始质疑Alphabet和微软 等客户在AI上投入的数千亿美元是否物有所值。与此同时，投资者纷纷逃离那些被认为可能受到AI颠覆威胁的行业。 投资者将密切关注的另一个指标是毛利率。去年，由于Blackwell芯片的高生产成本，该盈利指标一度承压。市场预计英伟达去年第四季度 调整后毛利率为75%，为一年多来的最高水平，并预计在本财年维持在该水平附近。在内存芯片涨价和其他成本上升的情况 ...

Core Insights - TSMC's market capitalization reached $2 trillion, doubling in just over a year since it first surpassed $1 trillion in October 2024 [4] - The surge in TSMC's valuation is largely driven by strong demand in the AI sector, with the company forecasting Q1 2024 revenue between $34.6 billion and $35.8 billion [4] - Major tech companies like Amazon and Meta are significantly investing in AI infrastructure, with Amazon planning $12 billion for new data centers and Meta over $115 billion [5] Group 1: TSMC's Market Performance - TSMC's stock price increased by 4.25% to $385.75 per share, marking a significant milestone in its market valuation [4] - The company has indicated robust AI demand, with clients expressing strong needs for production capacity [4][5] - TSMC is expected to increase its production capacity by 100% over the next decade to meet growing demands [6] Group 2: AI Infrastructure Investments - Amazon's capital expenditure for AI infrastructure is projected to be $200 billion this year, while Meta's is expected to exceed $115 billion [5] - NVIDIA's CEO highlighted the importance of TSMC in producing essential AI chips, indicating a long-term commitment to AI infrastructure development [5][6] - Concerns about potential over-investment in AI and the existence of an AI bubble have been raised, with industry leaders expressing uncertainty about future demand [6] Group 3: Competitive Landscape - TSMC remains a leader in advanced process technology, having begun mass production of 2nm chips, while competitors like Intel are struggling with production yields [6] - The semiconductor industry faces competitive pressures, but TSMC's technological advancements position it favorably [6] - The market is closely watching NVIDIA's upcoming quarterly earnings report as a potential indicator of the AI sector's health [6]

Group 1 - TSMC's stock price rose by 4.25% to $385.75 per share, achieving a market capitalization of $2 trillion, marking a significant milestone just 16 months after reaching $1 trillion [2] - TSMC's market capitalization doubled within a year, driven by strong demand signals from the AI sector, with projected Q1 revenue between $34.6 billion and $35.8 billion [2] - TSMC's downstream AI chip customers, AMD and NVIDIA, are seeking business expansion, with NVIDIA announcing a strategic partnership with Meta for the deployment of millions of chips [2] Group 2 - Amazon plans to invest $12 billion in building new data centers in Louisiana to support AI and cloud computing infrastructure, indicating ongoing high capital expenditures in the AI sector [4] - NVIDIA's CEO highlighted the importance of TSMC in the production of Blackwell and Rubin chips, indicating a potential 100% increase in TSMC's capacity over the next decade [4] - TSMC is leading in advanced process technology, having started mass production of 2nm (N2) process, while competitors like Intel are struggling with production yields [4] Group 3 - Concerns exist regarding potential over-investment in AI by tech giants, with some industry leaders questioning the clarity of future AI infrastructure needs [5] - A recent global fund manager survey indicated that capital expenditures by AI cloud service providers are viewed as a significant systemic credit risk [5] - NVIDIA's upcoming quarterly earnings report is anticipated to serve as a barometer for the strength of AI demand [5]

Core Viewpoint - TSMC's market capitalization has doubled in just over a year, reaching $2 trillion, driven by strong demand in the AI sector and significant partnerships with major tech companies [1][2][3] Group 1: Market Performance - TSMC's stock price increased by 4.25% to close at $385.75 per share, achieving a market capitalization of $2 trillion [2] - The company reached the $1 trillion market cap milestone in October 2024, becoming the seventh tech company in history to do so [2] Group 2: AI Demand and Partnerships - TSMC is experiencing strong demand signals from AI chip customers, with revenue projections for Q1 expected to be between $34.6 billion and $35.8 billion due to the booming AI industry [3] - Major clients like AMD and NVIDIA are expanding their business, with NVIDIA announcing a strategic partnership with Meta for the deployment of millions of chips [3] - Amazon plans to invest $12 billion in building new data centers in Louisiana to support AI and cloud computing infrastructure [3] Group 3: Industry Outlook - NVIDIA's CEO highlighted the importance of TSMC in the production of their Blackwell and Rubin chips, indicating a potential 100% increase in TSMC's capacity over the next decade [4] - TSMC is leading in advanced process technology, with plans to start mass production of 2nm chips by Q4 2025, while competitors like Intel are struggling with production yields [4] Group 4: Market Concerns - There are concerns about potential over-investment in AI by tech giants, with some industry leaders questioning the sustainability of current AI infrastructure plans [5] - A recent survey indicated that capital expenditures by AI cloud service providers are viewed as a significant credit risk [5] - NVIDIA's upcoming quarterly earnings report is anticipated to provide insights into the ongoing AI demand and market expectations [5]

智通财经2月24日讯，今日93股涨停，42股炸板，封板率为69%，豫能控股、美邦股份4连板，汉缆股份 3连板，协鑫集成、韩建河山9天5板，风语筑6天4板，金牛化工7天4板，顶固集创20cm 6天3板。市场冲 高回落，创业板指盘中一度涨超2%。沪深两市成交额2.2万亿，较上一个交易日放量2194亿。盘面上， 市场热点快速轮动，全市场超4000只个股上涨。板块方面，油气、化工、培育钻石方向领涨；影视、 AI应用、算力租赁方向领跌。截至收盘，沪指涨0.87%，深成指涨1.36%，创业板指涨0.99%。 人气及连板股分析 连板晋级率升至60%，但3板及以上连板股仍仅有3只。节前受到Seedance2.0发布以及博弈春节档票房的 双重提振表现火爆的影视传媒、AI应用端双双遭遇退潮，此前5连板的掌阅科技开盘后仅2分钟即封死 跌停，进而带动光线传媒在内的多只影视类个股封跌停。随着春节假期间大宗商品价格大面积走高，涨 价概念得到资金大举抢筹，贵金属为代表的有色板块卷土重来，化工板块涨势持续火爆，十余股封涨 停。而春节长假期间高热度的算力出海以及机器人等方向遭遇资金兑现，整体呈现高开低走的巨震分 歧，可见短线资金整体风偏仍相对 ...

Investment Rating - The report maintains a "Buy" rating for the company [7] Core Views - The company is expected to benefit from the growing demand for AI power supply products, leveraging its product, technology, and channel advantages to secure orders and enhance performance [1][2] - The recent completion of a private placement, raising 2.63 billion RMB, is anticipated to support the company's long-term development and competitiveness in the AI power supply market [2] - Despite a significant decline in net profit for 2025, the company is projected to recover starting in 2026, driven by scale effects and the delivery of AI power supply orders [2] Company Overview - The company is deeply embedded in the power supply iteration cycle of NVIDIA, continuously matching the power supply product development needs for new generation chips like GB300 and Rubin [4] - The company has established a strong customer base and is enhancing its competitive edge through technological advancements and partnerships across the entire supply chain [4] Industry Insights - The demand for AI power supplies is expected to accelerate in the first half of 2026, driven by the ongoing iteration of AI chips and the increasing power requirements for new products [3] - The global market for AI server power supplies compatible with NVIDIA GPUs is projected to reach 47.4 to 63.5 billion RMB in 2026, reflecting a year-on-year growth of approximately 87% [15][16] Financial Forecasts and Valuation - The company's net profit for 2025 is revised down to 141 million RMB, while projections for 2026 and 2027 are increased to 1.1 billion RMB and 1.84 billion RMB, respectively, indicating a compound annual growth rate (CAGR) of 61.65% from 2025 to 2027 [5][19] - The target price for the company is raised to 151.38 RMB, based on an 80x price-to-earnings (PE) ratio for 2026, reflecting the company's competitive advantages in the AI power supply sector [5][21]

Core Insights - Nvidia is facing significant supply chain challenges as it approaches the production timeline for its next-generation AI accelerators, potentially leading to a shift in its procurement strategy to prioritize supply stability over extreme performance [1][5] - The overall market supply dynamics are heavily influenced by Nvidia's final procurement decisions, with expectations that it will adopt a more pragmatic approach by procuring both top-spec and slightly lower-spec versions of HBM4 [1][2] Group 1: Supply Chain Challenges - Major storage chip suppliers are encountering dual obstacles in capacity expansion and yield improvement, which may force Nvidia to relax its technical specifications for HBM4 [1] - Samsung Electronics, despite being ahead in the qualification process for HBM4, may struggle to meet Nvidia's demand for large-scale production due to yield and capacity limitations [3] - SK Hynix is also facing challenges in meeting performance standards for HBM4, with its products struggling to reach the required 11Gbps performance level [4] Group 2: Procurement Strategy - Nvidia is likely to implement a dual-track procurement strategy, purchasing both the highest specification HBM4 and slightly lower-spec versions to ensure stable supply from major suppliers like Samsung, SK Hynix, and Micron [5] - This strategy is seen as a necessary measure to ensure the timely delivery of the next generation of AI infrastructure amid increasing shortages in the storage chip market [5]