以基本面为基础，专注于趋势交易 还在信"AI是泡沫"？ 三家海外巨头刚甩来的财报，直接把这论调抽成了筛子。 Snowflake 、 MongoDB 、 CrowdStrike 的 FY26Q3 成绩单新鲜出炉，营收、 EPS 双双 炸穿预期。 以下文章来源于格隆汇交易学苑 ，作者格隆汇小编 格隆汇交易学苑 . AI 业务更狠，从" PPT 概念"焊死在"赚钱实景"里。 前阵子有人喊 " AI 只烧钱不造血"，现在看，数据处理、平台适配、安全防护三条赛道，全跑 成了"印钞机"。 别猜泡沫了，真相是： AI 应用早起飞，关键是抓对赛道 。 01 数据基础设施AI化： AI落地的"地基工程" AI 要赚钱，先得把数据"伺候"好。 这是行业铁律，就像盖楼先打地基，少了这步，再牛的 AI 也是空中楼阁。 Snowflake 这季度的表现，把"地基刚需"焊得明明白白。 总营收 12.1 亿美元，同比暴涨 29% ；核心产品营收 11.6 亿美元，增速丝毫不差。 双双踩碎市场预期，最狠的是 AI 业务——年化收入破 1 亿美元，比原计划早了整整一个季 度。 这不是 "实验室胜利"，是客户真金白银堆出来的。 福布斯 2000 ...

2026-2027年度英伟达研究生奖学金名单新鲜出炉，10位获得最高6万美金资助的「天选之子」中，华人学者直接霸榜8席！其中更是不乏名师 高徒，比如李飞飞的学生也赫然在列，正在教机器人怎么像人一样干活。从让游戏画质起飞的神经渲染，到防不胜防的AI安全攻防，这些博 士生研究的都是硬核中的硬核。 刚刚，英伟达研究生奖学金（Nvidia Graduate Fellowship）名单出炉！ 该项目致力于挖掘并支持全球顶尖的博士研究生，助力他们在计算创新的前沿领域攻克难题。 作为一项久负盛名的学术荣誉，该项目已走过25年历程，不仅为获奖者提供高达60,000美元的科研经费，还提供在英伟达实验室进行暑期实习的宝贵机 会，使学生能够直接接触最先进的技术资源与行业专家。 以下是本年度10位最高奖学金获得者以及5位最终入围者的详细履历。 奖学金获得者(Recipients) Jiageng Mao 学校：南加州大学(University of Southern California) 研究方向：物理人工智能、具身智能、自动驾驶 获奖理由：通过利用互联网规模数据中的多样化先验知识，解决复杂的物理人工智能问题，为现实世界中的具身智 ...

Core Viewpoint - The report from Dongfang Securities highlights that Green Alliance Technology focuses on AI and data security, cultivating new future scenarios through a "3+3+X+Y" strategy, driving differentiated competitiveness with dual engines of AI security and data security [1] Group 1: AI Security Developments - The company has iterated its Fengyunwei AI security capability platform, establishing a multi-agent security application center to address challenges in AI model assessment [1] - The launch of the AI-Scan large model risk assessment product enhances vulnerability detection and threat assessment efficiency, with successful implementations in government and financial sectors [1] Group 2: Emerging Security Scenarios - The company is actively developing new security scenarios, achieving technological breakthroughs in areas such as APT tracking and forensics, cloud offensive and defensive strategies, new drone countermeasures, and vehicle networking test fields [1] - The vehicle networking test field business has completed key technology validation, laying the groundwork for future large-scale applications [1] Group 3: Cloud Security and Privacy Computing - The company is deepening research in cloud-native security and privacy computing, enhancing its product matrix to meet diverse cloud security needs [1] - Continuous breakthroughs in these emerging fields are expected to open new growth opportunities for the company, forming long-term development potential [1] Group 4: Financial Outlook - Considering the current market conditions and comparable companies, a target price of 9.42 yuan is set for 2026, maintaining a "buy" rating with a projected 3x PS ratio [1]

新智元报道 【新智元导读】2026-2027年度英伟达研究生奖学金名单新鲜出炉，10位获得最高6万美金资助的'天选 之子'中，华人学者直接霸榜8席！其中更是不乏名师高徒，比如李飞飞的学生也赫然在列，正在教机器 人怎么像人一样干活。从让游戏画质起飞的神经渲染，到防不胜防的AI安全攻防，这些博士生研究的 都是硬核中的硬核。 刚刚，英伟达研究生奖学金（Nvidia Graduate Fellowship）名单出炉！ 该项目致力于挖掘并支持全球顶尖的博士研究生，助力他们在计算创新的前沿领域攻克难题。 作为一项久负盛名的学术荣誉，该项目已走过25年历程，不仅为获奖者提供高达60，000美元的科研经 费，还提供在英伟达实验室进行暑期实习的宝贵机会，使学生能够直接接触最先进的技术资源与行业专 家。 以下是本年度10位最高奖学金获得者以及5位最终入围者的详细履历。 奖学金获得者（Recipients） Jiageng Mao 学校：南加州大学（University of Southern California） 学校：加州大学圣地亚哥分校（University of California， SanDiego） 研究方向：物理 ...

Investment Rating - The report maintains a "Buy" rating for the company with a target price of 9.42 CNY, based on a projected price-to-sales (PS) ratio of 3 times for 2026 [3][6]. Core Insights - The company is focusing on AI and data security, aiming to enhance its competitive edge through a strategic framework termed "3+3+X+Y" [2][10]. - Revenue forecasts for 2025-2027 are adjusted to 2.402 billion CNY, 2.546 billion CNY, and 2.857 billion CNY respectively, reflecting a growth trajectory despite previous estimates being lower [3][11]. - The company has made significant improvements in cost control, leading to a reduction in expense ratios and an overall enhancement in operational efficiency [10]. Financial Performance Summary - The company reported a revenue of 1.681 billion CNY in 2023, with a projected growth of 40.3% to 2.358 billion CNY in 2024, followed by modest growth rates in subsequent years [5][13]. - The gross margin is expected to stabilize around 53.6% to 54.0% from 2025 to 2027, indicating a strong pricing power and cost management [5][10]. - The net profit margin is projected to improve significantly, moving from -58.1% in 2023 to a positive 8.1% by 2027, showcasing a turnaround in profitability [5][10]. Strategic Focus - The company is actively exploring new security scenarios, including APT tracking, cloud security, and vehicle networking, which are expected to open new revenue streams [10]. - The introduction of innovative products like the AI-Scan model risk assessment tool is aimed at enhancing operational efficiency and effectiveness in threat detection [10].

Investment Rating - The report maintains a "Buy" rating for the company, with a target price of 59.87 CNY based on a 2.65x PS multiple for 2026 [3][6]. Core Insights - The company is experiencing improved profitability primarily due to effective cost control, with a significant reduction in expense ratios [9]. - AI security is identified as the strategic core of the company, with the upcoming launch of the "Heng Nao 3.0" platform, which is expected to enhance product competitiveness [9]. - New business segments, particularly data security and MSS services, are showing rapid growth, with data security contracts increasing by nearly 25% year-on-year [9]. Financial Performance Summary - Revenue projections for 2025-2027 are adjusted to 2.126 billion, 2.306 billion, and 2.538 billion CNY, respectively, reflecting a growth rate of 4.1%, 8.5%, and 10.1% [3][5]. - The company is expected to achieve a net profit of 111 million CNY by 2027, with a significant turnaround from a net loss of 360 million CNY in 2023 [5][9]. - Gross margin is projected to decline slightly from 61.8% in 2023 to 58.3% by 2026, while net margin is expected to improve from -16.6% in 2023 to 4.4% in 2027 [5][9].

由新华社品牌工作办公室、新华网、新华社海南分社主办的2025企业家博鳌论坛系列活动于12月2日至5 日在海南博鳌举办。作为论坛系列活动之一，2025AI生态共建论坛暨《睿德数字经济创新发展大数据 平台》发布会于3日在博鳌举办。 贾昊表示，当前AI发展正从通用普惠阶段深化至垂直行业应用阶段，但在规模化落地过程中，仍面临 数据治理、算力成本、安全可信及复合型人才短缺等挑战。人工智能的未来既需要开放拥抱，也离不开 理性建构。他认为，只有在数据可信、安全可控、人才充沛的前提下，持续推动技术与行业的深度融 合，才能真正赋能千行百业，助力数字时代高质量发展。 在人才跨界融合方面，覃卓认为，当前兼具专业技术与行业洞察能力的复合型人才较为稀缺。为系统化 培养适应时代需求的未来人才，北大法宝与多所高校合作，共同开设法律大数据、法律AI等课程，着 力推进产学研协同育人。在提升垂类AI可信安全方面，他介绍，北大法宝基于二十余年积累，已建立 严谨的数据采集、标注与审核流程，并通过为AI回答提供法条原文链接等方式，确保结果可溯源、可 验证，有效维护信息准确性，增强用户信任。 本次圆桌论坛聚焦垂直领域AI落地的真问题、真场景及安全问题， ...

Core Insights - The integration of AI and data security has reached a critical stage, with major tech companies releasing core solutions to shift the industry from passive defense to proactive governance [1][2] - High-quality data is essential for the effective functioning of AI models, as poor data leads to suboptimal outcomes [1][2] - The relationship between large models and data is likened to "rocket engines and fuel," emphasizing the need for deep collaboration [2] Industry Trends - By 2027, 60% of large enterprises in China are expected to adopt exposure management technology in their security operations centers to enhance incident response efficiency [1] - Currently, only 5% of Chinese enterprises deploying AI technologies are using collaborative AI defense strategies, but this is projected to rise significantly by 2028 [1] Technological Advancements - AI technology has improved the efficiency of data classification and grading by over three times, achieving an accuracy rate of over 95% [2] - In the API security domain, AI agents can filter non-API assets and create high-purity asset lists through advanced analysis techniques [2][3] Practical Applications - The advancements in AI and data security are not just theoretical; they have been transformed into mature solutions for industries such as finance and government [4] - AI's role in data security is twofold: enhancing security intelligence and ensuring the safety of intelligent systems [2] Regulatory Landscape - The balance between security and data utilization is a key focus, with industry leaders advocating for robust security measures to facilitate data flow [6] - Recent global regulatory changes, such as the relaxation of GDPR for AI training data, reflect a shift towards balancing technological competition and regulatory oversight [8] Collaborative Ecosystem - The establishment of a collaborative ecosystem involving government, enterprises, and research institutions is crucial for effective security measures [9] - The concentration of numerous small security firms in specific regions creates a unique competitive advantage, fostering a complete ecosystem of "security services + digital applications" [9]

Core Insights - ServiceNow is in advanced talks to acquire security startup Veza for over $1 billion, aiming to enhance its AI Agent-related products and services [1][2] - Veza's subscription software helps enterprises manage access permissions for employees and AI Agents, addressing new security challenges posed by AI [2] - The acquisition could differentiate ServiceNow's offerings in a crowded market of similar AI products from various software vendors [2] Group 1: Acquisition Details - ServiceNow is negotiating a deal that may be announced as soon as next week, although it is not finalized yet [2] - The acquisition is part of ServiceNow's strategy to sell more AI tools to its clients, enhancing its market position [2] Group 2: Veza's Business Model - Veza's annual service contracts, sold through Amazon Web Services, are priced at $100,000 [3] - The company has notable clients including Blackstone, Instacart, and Snowflake, indicating its market relevance [3] - Veza has experienced significant growth, with employee numbers increasing by over 50% in the past year [3] Group 3: Market Context - The demand for data control is rising as AI Agents may lead to new types of cyberattacks, making Veza's solutions increasingly relevant [2] - ServiceNow's previous acquisition of Moveworks for $2.85 billion is currently under antitrust review, highlighting the competitive landscape in the IT services sector [4]

立体密算体系融合国产芯片、国密算法与可信计算，构建覆盖算力、数据、AI全链路的内生安全防 护，将性能损耗降至传统方案1/3以下。 当数据泄露事件频发、AI算法黑箱难破、跨域算力信任缺失逐渐演变为数字经济发展的三大"拦路虎"， 一场围绕安全体系的范式革命，正悄然拉开序幕。 在此背景下，11月26日，由数据安全关键技术与产业应用评价工业和信息化部重点实验室、曙光云计算 集团、中国科学院计算技术研究所、北京大学大数据分析与应用技术国家工程实验室等多家单位联合发 起的"立体密算体系"正式亮相。 该体系以国产自主芯片为可信根基，融合商用密码、云计算与人工智能技术，不仅突破了传统安全方 案"头痛医头"的局限，更构建起覆盖算力、网络、存储、数据与AI全流程的立体化密态防护体系。 为深入理解该体系的技术路径与产业前景，21世纪经济报道记者近日与曙光云多位核心负责人展开对 话，围绕其技术内核、行业落地与生态共建进行了探讨。 范式演进：从单点防护到体系化防御 在数字安全的发展历程中，机密计算、隐私计算、可信计算等技术曾如孤岛般各自为战，虽在特定场景 有所建树，却始终难以形成贯穿全局的防护合力。 对此，曙光云总裁助理李明达直言：" ...