Core Viewpoint - Nvidia is set to report its fiscal Q4 2025 earnings on February 25, 2026, with a focus on sustaining AI-driven growth momentum after a volatile start to the year [1] Revenue Expectations - Wall Street consensus estimates for Nvidia's Q4 revenue are approximately $38.3 billion, slightly above the company's prior guidance of around $37.5 billion [1] - Analysts expect adjusted earnings per share to be in the range of $0.84 to $0.85, indicating continued year-over-year growth driven by demand in data centers and artificial intelligence [2] Stock Performance - Nvidia's stock has experienced notable volatility at the beginning of 2026, attempting to break through the $200 resistance zone, with shares valued at $185, up nearly 0.5% in the last session [2] - Over the past year, Nvidia's equity has rallied more than 40% [2] Price Predictions - In a base-case scenario where Nvidia meets or slightly exceeds revenue expectations, the stock could trade between $215 and $225 post-earnings, reflecting a positive market reaction based on earnings durability [4] - A more bullish outcome, with a clear earnings beat and stronger guidance on AI infrastructure demand, could see shares advance to the $235 to $245 range as growth expectations for 2026 are repriced [5] - Conversely, if Nvidia meets earnings expectations but provides cautious guidance, the stock may struggle, trading between $175 and $185, indicating consolidation rather than a breakdown in the long-term thesis [6] - In a less likely bearish scenario with disappointing earnings or guidance, shares could retrace to the $155 to $165 range, requiring evidence of slowing AI demand or profitability pressures [8] Market Sentiment - Overall, Nvidia's post-earnings price action is expected to be driven more by forward guidance than by headline earnings alone, according to AI modeling by ChatGPT [3][8]

贵金属狂欢盛宴，现货白银正迎来史诗级"完美风暴"。 今年以来，白银价格已上涨近25%。 赶超英伟达！ 受美联储降息预期、鲍威尔遭起诉以及地缘局势升温影响，市场避险需求显著提升。 其中，现货白银盘中大涨5%报91.31美元/盎司，纽约期银日内也涨超5%突破91美元/盎司。 周三，银价一度强势突破91美元大关，为历史上首次。 截至发稿，涨幅有所回落，现货白银现涨超3%，交投于90美元附近。 这增强了市场对美联储货币政策转向的押注。 与此同时，美联储主席鲍威尔遭起诉的消息，再次引发了投资者对美联储独立性的担忧。 地缘政治紧张局势加剧，加大了避险需求。 不久前，特朗普威胁对伊朗采取"非常强有力的行动"，美国敦促公民立即离开伊朗。 这也被视为美国可能很快采取行动的最新信号。 当前，白银总市值已经突破5万亿美元大关。 CompaniesMarketCap数据显示，受银价提振，白银已超越英伟达成为全球第二大资产。 黄金以32.162万亿美元的市值位居榜首，芯片巨头英伟达以4.523万亿美元的市值滑落至第三。 白银狂飙背后是多重因素的共振。 首先，昨日美国公布的通胀数据表现疲软，市场加大了对美联储降息的预期。 美国12月核心 ...

Core Viewpoint - The rise of artificial intelligence (AI) has significantly boosted Nvidia's market value, but historical challenges suggest potential risks for the company moving forward [2][10]. Group 1: Nvidia's Market Position - Nvidia has added over $4.1 trillion in market value since the beginning of 2023 and briefly reached a market cap of $5 trillion [2]. - The company's GPUs are the preferred choice for AI-accelerated data centers, with competitors struggling to match Nvidia's compute capabilities [5]. - Nvidia's current stock price is around $185.60, with a market cap of $4.5 trillion and a gross margin of 70.05% [7]. Group 2: Technological Advancements - Nvidia's CEO Jensen Huang is committed to an aggressive innovation timeline, planning to release a new advanced chip annually, with the upcoming Vera Rubin chip requiring fewer GPUs for training large language models [7]. - The CUDA software platform is crucial for maximizing the compute potential of Nvidia's hardware, fostering customer loyalty [8]. Group 3: Partnerships and Collaborations - Nvidia has formed strategic partnerships, including a significant collaboration with OpenAI, which will deploy at least 10 gigawatts of AI data centers using Nvidia's GPUs [9]. Group 4: Historical Challenges - Historical trends indicate that major technological advancements often face early-stage bubble-bursting events, which could pose risks for Nvidia as AI technology matures [12][13]. - Nvidia's price-to-sales (P/S) ratio has recently exceeded 30, a level historically associated with market bubbles, indicating potential overvaluation [16]. - The S&P 500's high valuation could lead to significant declines for companies like Nvidia if a market correction occurs [17]. Group 5: Competitive Landscape - While Nvidia currently leads the market, there is a risk of losing data center space to top customers who are developing their own, cheaper GPUs [18].

周三，国际银价首次突破90美元/盎司，这标志着一个历史性的里程碑。目前该贵金属的总市值已经突破5万亿美元，超越英伟达，成为全球第二大资 产。这凸显出白银在全球经济中的战略重要性日益增强。 周三，现货白银开盘强势拉升，盘中一度涨3.55%，首次站上90美元关口，再创历史新高。一方面，周二公布的美国通胀数据巩固了美联储降息押注， 另一方面，持续的地缘政治和经济不确定性推动了避险需求。 根据市值数据平台CompaniesMarketCap的数据，目前白银市值为5.045万亿美元；黄金以32.251万亿美元的市值位居榜首，芯片巨头英伟达以4.523万亿 美元的市值滑落至第三。前十名榜单如下： | | | Top Assets by Market Cap | | | | | | --- | --- | --- | --- | --- | --- | --- | | | | All assets, including public companies , precious metals , cryptocurrencies , ETFs | | | | | | Rank + | Name | 0 Market Cap | ...

若"50万吨"这一数据准确无误，那么一个1吉瓦的数据中心，就将消耗全球年铜供应量的1.7%。按照人工智能产业发展的合理预期，未来若新建总容量30吉 瓦的数据中心，仅这一个领域，理论上就将消耗全球近一半的铜矿产量。 这一数据听起来堪称大宗商品市场的终极利好催化剂，铜价也随之屡创历史新高。但现实是，这件事恰恰给市场敲响了警钟。 在大宗商品领域，"结构性牛市逻辑"与"现实物理限制"息息相关，至少在媒体报道与部分过度乐观的分析师眼中是如此。 近日，《福布斯》杂志翻出了英伟达（NVDA）一份去年5月首次发布的技术报告，这份最初源于英伟达技术简报的文件，随后在各类研究报告与人工智能 训练数据集中广泛传播。英伟达在报告中称，传统架构中，单个1兆瓦机架在物理上要求高达200千克的铜母线，单个1吉瓦的机架则可能需要多达50万吨的 铜。显然，现有电力分配技术，难以支撑未来吉瓦级数据中心的大规模建设需求。相关表述目前仍可在其官网查询。 记者们跟风报道的行为尚可理解，但连美国铜业发展协会（The Copper Development Association）也在宣扬这一错误数据，而该机构本该具备专业的行业判 断能力。 对于多年来一直 ...

Core Insights - The Trump administration has finalized the requirements for Nvidia to sell its H200 chips in the Chinese market [1] Group 1 - The finalized requirements indicate a regulatory framework that Nvidia must adhere to for its chip sales [1]

提高大模型记忆这块儿，美国大模型开源王者——英伟达也出招了。 联合Astera研究所、斯坦福大学、UC伯克利、加州大学圣地亚哥分校等机构推出了TTT-E2E方法。 在128K超长文本上处理速度比全注意力模型快2.7倍，处理2M上下文时提速达35倍，性能还不打折。 这项技术与前几天大火的DeepSeek条件记忆模块有所不同。 DeepSeek的Engram模块依赖的是"按需查表"的静态学习路径，而英伟达走的是动态学习的路子，关键在于上下文压缩。 通过实时学习将关键内容压缩到自身权重中，让模型在测试阶段依然保持学习状态。 这样既避免了额外缓存的负担，又能精准捕捉长文本中的核心逻辑。 给模型装上记忆压缩包 TTT-E2E并没有依赖复杂特殊架构，反而是基于带滑动窗口注意力的标准Transformer，容易部署。 这个方法的核心思路是将长文本建模从架构设计问题转化为「持续学习」任务。 在测试阶段，模型会基于当前读取的上下文进行下一个词预测。 为了平衡效率与稳定性，TTT-E2E还设计了三项关键优化。 一是采用「迷你批处理+滑动窗口」的组合策略。将测试时的训练数据分成多个迷你批，配合8K大小的滑动窗口注意力，既解决了单t ...

（来源：财闻） | | | | Top Assets by Market Cap | | | | | | | --- | --- | --- | --- | --- | --- | --- | --- | --- | | | | | All assets, including public companies , precious metals , cryptocurrencies | | | ETFs | | | | Rank 4 | Name | | | Market Cap 0 | � Price | Today | Price (30 days) | Country | | 1 | | Gold GOLD | | $32.265 T | $4,640 | 0.90% | | | | 2 | | Silver SILVER | | $5.070 T | $90.08 | 4.33% | | | | 3 | 0 | NVIDIA NVDA | | $4.523 T | $185.81 | 0.47% | | USA | | 4 | | Alphabet (Google) GOOG | | $4.061 T ...

（图片来源：摄图网） AI+医药重磅利好消息来了! 1月12日，英伟达与制药巨头礼来正式宣布，将在未来五年内共同投资10亿美元，在旧金山湾区建设联 合实验室，加速AI在药物研发领域的应用，这是首个由制药公司与AI芯片巨头共建的实体实验室，标 志着AI制药从软件协作走向深度硬件-数据融合。 随着英伟达加速渗透生物技术市场，该公司采取的战略是提供开源人工智能模型和软件，让制药商能够 基于英伟达的硬件，利用这些模型和软件构建专属的药物开发平台。不仅如此，在宣布合作当天，英伟 达还发布了一个可用于确保使用人工智能工具设计的药物在真实世界实验室中合成的更新模型，致力于 打造全球首个药物研发的AI工厂。 礼来对AI制药的投入可谓不遗余力。早在2025年10月，礼来就宣布采用英伟达AI系统搭建AI工厂，并 着手部署搭载逾1000颗英伟达旗舰Grace Blackwell芯片的超级计算机。礼来董事长满怀信心地表 示："将我们海量的数据和科学知识与英伟达的计算能力和模型构建专长相结合，有望彻底革新我们所 知的药物研发模式。"据统计，仅2025年，礼来就进行了十次对外的AI制药领域合作，合作对象除了英 伟达、OpenAI，还涵盖 ...

Jim Cramer highlighted the underappreciated synergy between Nvidia Corp. (NASDAQ:NVDA) and Eli Lilly And Co. (NYSE:LLY), calling their deep collaboration a “monumental effort” to revolutionize drug development that Wall Street is foolishly dismissing as a “gigantic sideshow.”$1 Billion Validation Of AI In HealthcareWhile the broader market remains fixated on daily earnings fluctuations and inflation data, Cramer urged investors to focus on the structural shift happening in San Francisco.Nvidia and Eli Lilly ...