刘熹 国海证券计算机分析师: 那下面的话，会议就正式开始。那首先的话，还是想请这个云赛的张总先做一个开场，然 后后面再请武博士，再做，结合这个 PPT 再做一个分享。好，下面有请张总。 云杉智联董秘张欣欣: 好，各位投资者，大家晚上好。今天那个是还是年初八，八，在在中国农历那个新年当中 又是马年的那个开始第一天，那么有幸国海证券刘老师安排，今天晚上利用一点时间跟大 家见面。我觉得非常意义，非常有意义的。那么我是银赛智联董秘张欣欣，那么银赛智联 大家应该各位大概应该都有所了解。那么它现在的那个主要是一一家信息服务业企业，主 要业务有三三块，一个是云服务大数据，第二个是解决方案，第三个是智能产品。然后云 服务大数据板块，里面包括一个是云服务，一个是数据要素，数据治理，那么解决方案主 要做城市安全治理和和那个民生，包括医疗、教育，很多信息化服务。 那么还有一块就是智能产品智能产品。那么目前应该说我们那个银盛智联各块业务，就是 运营比，很正常。那个应该说我们目前还在对 25 年的那个年报的数据还在审计，那么应 该很快吧，大概还有一个月。到 3 月底的时候，我们会公布年报。那么应该说还不错的， 还不错的，因为整个 25 ...

Core Insights - U.S. equities are not alarmingly overvalued, but finding undervalued stocks can be challenging, especially within high-growth ETFs like QQQ and QQQM [1] - Despite fears of AI disruption, certain technology stocks within these ETFs are now considered undervalued, presenting potential investment opportunities [1] Group 1: Undervalued Stocks in QQQ/QQQM - Adobe (ADBE) is viewed as vulnerable to AI but is currently undervalued, with analysts seeing rebound potential due to its addressable market exceeding $200 billion and ongoing M&A efforts [1] - Intuit (INTU), owner of TurboTax, is also undervalued and has significant upside potential through its unique offerings like TurboTax Live and QB Live, which differentiate it from competitors [1] - Microsoft (MSFT), while not a traditional value stock, trades at a 34% discount to Morningstar's fair value estimate and is transitioning to cloud-based SaaS solutions, maintaining a strong position in office productivity software [1]

The Federal Trade Commission (FTC) has been asking several of Microsoft’s competitors about the company’s business practices as part of an ongoing antitrust probe, Bloomberg reported Friday (Feb. 13), citing unnamed sources.By completing this form, you agree to receive marketing communications from PYMNTS and to the sharing of your information with our sponsor, if applicable, in accordance with our Privacy Policy and Terms and Conditions .Complete the form to unlock this article and enjoy unlimited free acc ...

Core Viewpoint - Microsoft has seen a 15% decline in stock price since the earnings report on January 28, primarily due to concerns over increased capital expenditures without a corresponding rise in Azure growth guidance and competition from new AI tools in the Office 365 business [1] Group 1: Capital Expenditure and Azure Growth - Microsoft is adjusting its computing power allocation, increasing the share for internal research and Copilot from approximately 10% to 20% [3][5] - For fiscal year 2026, capital expenditures are projected to reach $148 billion, a 400% increase from fiscal year 2022 [3] - Current computing power allocation is about 70% for Azure and 30% for Copilot and internal research, with new GPU allocations approaching a more balanced 60/40 split [6] Group 2: Iceberg Model of Computing Investment - Goldman Sachs introduced an "iceberg analogy" to explain Microsoft's computing capital expenditure strategy, distinguishing between visible investments that directly impact Azure and Office 365 and invisible investments crucial for long-term strategy [4] - If all new computing capacity is allocated to Azure, its growth rate could exceed 40%, indicating a strategic postponement of some growth potential for long-term investments [4][10] Group 3: Drivers of Increased Internal Allocation - Four structural factors are driving the increased demand for Microsoft's computing power, leading to a significant evolution in capital expenditure structure [7] - Azure AI computing power accounts for the largest share, with a multiplier of 4-5 times for "computing power/quarterly new revenue," translating to $0.20 to $0.25 in revenue for every $1 of capital expenditure [8] Group 4: Copilot Investment Strategy - Microsoft is positioning Copilot as a cost-effective enterprise-level LLM solution, priced at approximately $30 per user per month, compared to ChatGPT Enterprise at $60 [12] - Copilot is beginning to make a substantial contribution to M365 commercial cloud revenue, projected to reach $89 billion in fiscal year 2026 [13] - The adoption rate of Copilot is increasing among both internal employees and external customers, driving up computing power consumption [9]

Core Insights - The article discusses the competitive landscape of AI agents, highlighting how traditional software companies are racing to develop AI products that can automate tasks previously performed by human workers [1][3][8] Group 1: AI Agent Development - Companies like Microsoft, ServiceNow, and Snowflake are launching AI agent products to help clients create customized AI agents that can interact with various enterprise applications [1][3] - The emergence of AI agent management dashboards raises questions about the necessity of multiple dashboards, suggesting that each client may ultimately only need one [2] Group 2: Key Players and Products - Major players in the AI agent space include Anthropic, OpenAI, and Google, with products designed to automate tasks across different applications [3][5] - OpenAI's Frontier project aims to assist companies like Uber and Thermo Fisher Scientific in developing multiple AI collaborative assistants [9][10] Group 3: Market Dynamics and Challenges - Microsoft CEO Satya Nadella predicts that traditional software applications will collapse in the era of AI agents, as they are merely databases with business logic [8] - Despite the potential of AI agents, significant security concerns and operational challenges remain, such as the risk of credential leaks and the high operational threshold for current products [8][11] Group 4: Industry Sentiment - Executives from traditional software companies express a mix of caution and optimism regarding the integration of AI agents, with some companies already utilizing AI technologies from OpenAI and Anthropic [11][12] - The sentiment in the industry is that software leaders feel they must either achieve a trillion-dollar valuation or face extinction due to the disruptive nature of AI [12]

12 February 2026 | 4:03PM EST Equity Research Microsoft Corp. (MSFT) Microsoft is down ~15% since its 1/28 report, primarily for two reasons: 1) upward revisions to capex without upward revisions to Azure, resurfacing questions on ROI and Azure's competitive positioning relative to peers; and 2) ongoing debate on whether the knowledge worker applications business (Office 365) is disintermediated with new AI competition, such as Claude Cowork. We focus our analysis here on the first concern: capex ROI. Micro ...

视频里分析师指着柱状图讲： 昨天晚上刷YouTube时，正好刷到Bloomberg刚出的一个深度视频，标题是《Big Tech's $650 Billion Gamble》（科技巨头的6500亿豪赌）。 2026年，就亚马逊、谷歌、微软这几家，预计就要砸进去6500亿美金的资本支出（Capex）。 紧接着，他抛出一个特尴尬的结论：投入是指数级涨的，收入是线性涨的；如果不解决这个问题，2026 年的 AI 产业，很有可能撞上一个巨大的气穴。 就跟飞机似的，飞着飞着突然掉进真空里，那种失重的感觉，大家应该都能想象到。所以，看完这个视 频我认为，这不光是华尔街的焦虑，更是整个AI行业的过渡时刻。 咱们看看这6500亿美金是怎么来的，到底能烧出点啥？ Bloomberg视频里说的6500亿美金，是个挺微妙的数。我特意去翻了高盛的原始研报才发现，这数背后 是一种特别罕见的「倒挂」。 怎么理解这个倒挂？ 基建都跑到平流层了，应用还在慢慢爬坡。你看亚马逊、微软、谷歌、Meta这几家，2026年的资本支 出也差不多是这个数；这笔钱都花哪儿了？ 全用来买卡、建数据中心，甚至去抢电力资源了，这种投入力度，已经是「赌国运」级别的基 ...

Core Insights - The article discusses the evolving landscape of AI investments and the implications for major tech companies, highlighting a shift in market evaluation criteria from mere technological advancement to actual revenue contributions and profitability [4] - It emphasizes the transformative impact of AI on traditional software models and the competitive dynamics within the industry, particularly focusing on the challenges faced by companies like Microsoft [11][8] Group 1: AI Investment Trends - Major North American tech companies, including Amazon, Google, Meta, and Microsoft, plan to invest approximately $660 billion in capital expenditures by 2026 [1] - The market's response to aggressive capital spending has changed, with a focus on companies that can demonstrate sustainable profitability from AI investments [4] Group 2: AI Model Development - Claude Code represents a pivotal shift in AI development, moving from passive response models to proactive execution, fundamentally altering human-computer interaction [7] - The widespread adoption of AI agents is expected to disrupt traditional software industries, reducing marginal costs and undermining existing business models [8] Group 3: Storage Industry Dynamics - The storage industry is characterized by cyclical supply-demand mismatches, with significant capital investments required for chip manufacturing leading to low supply elasticity [12] - The current AI-driven storage supercycle is unprecedented, with structural demand surges and supply constraints leading to significant shortages in both HBM and general DRAM [14][15] Group 4: Future Projections - The AI-driven supercycle is anticipated to last until 2027, with ongoing supply shortages and high prices expected to persist in the short term [20] - Long-term changes in the industry may include a shift towards long-term supply contracts with cloud providers, reducing inherent cyclical volatility [21]

Core Insights - The market capitalization of SaaS, data, and software companies has evaporated by approximately $300 billion due to the release of an AI product, rather than poor earnings or macroeconomic shocks [1] - The IGV software index has dropped about 30% from its peak in late September, with significant declines in stock prices for major companies like Salesforce, ServiceNow, Adobe, and Intuit [2] - The average expected price-to-earnings ratio for software companies has plummeted from around 39 times to approximately 21 times in just a few months [2] Group 1: Market Dynamics - The crisis in the SaaS sector has been ongoing for months, with a recent acceleration in the speed of market reactions [2] - Short sellers have profited over $20 billion by betting against traditional SaaS businesses, indicating a loss of confidence in the sustainability of their growth models [2] - The core assumption that has been challenged is the sustainability of traditional SaaS growth models in the face of AI advancements [4] Group 2: AI Impact - AI is fundamentally testing the logic behind traditional SaaS models, as modern AI systems can replace many human workflows across various applications [6] - Investors are increasingly concerned that the growth of many SaaS companies may be rapidly supplanted by lower-cost, AI-driven solutions [8] - The shift towards AI-driven workflows is seen as a significant threat to the traditional SaaS business model, which relied on high growth with low or no profitability [7] Group 3: Industry Perspectives - Notable figures like Chamath Palihapitiya and Microsoft CEO Satya Nadella have expressed that the era of SaaS is over, emphasizing a shift towards AI-driven platforms [11][7] - The software industry's profit pool is expected to shift towards AI agents, with predictions that by 2030, over 60% of software economic benefits may come from agent systems rather than traditional SaaS services [15] - The transition is not indicative of a shrinking market but rather a reallocation of economic benefits from static applications to adaptive systems [18] Group 4: Future Outlook - Companies must embrace AI agents and integrate them into their business models to remain competitive in the evolving landscape [14] - The historical reliance on predictable revenue and low customer churn in the software industry is being reassessed as AI changes the dynamics of customer engagement and product value [20] - The future of software will likely focus on platforms rather than individual products, as platforms can offer greater integration and customer stickiness [22][27]

SemiAnalysis分析指出，Claude Code已占据GitHub代码提交的4%，预计2026年底超20%，标志着AI智能体技术进入规模化商用拐点。它正重构全 球15万亿美元的信息工作市场，并推动Anthropic季度营收增量超越OpenAI。传统软件商业模式面临根本冲击，而微软等巨头陷入平衡云服务增长与保 护核心产品的战略两难。智能体竞争已从生成响应转向实现可交付成果，整个行业正转向效率为王的实战阶段。 该报告强调， Claude Code不仅改变了编程形态，更预示着AI智能体将重塑全球约15万亿美元的信息工作市场。 目前，Anthropic依托该技术实现营收 快速增长，其季度年度经常性收入（ARR）增量已超越OpenAI，反映出AI智能体赛道的竞争格局正发生结构性变化。 这一趋势已引发产业链的深度响应。埃森哲已签署协议，计划在金融、医疗及公共部门培训3万名专业人员使用Claude，成为目前规模最大的企业级部 署案例。与此同时， 传统软件与SaaS商业模式面临根本性挑战。 分析指出，相当多的领域都将受到影响。 随着Claude Code（和Cowork）的兴起，智能体的总市场规模远大于大型语言模型本 ...