国际要闻： 20、南下资金净买入港股51.6亿港元，大幅加仓快手、中芯国际和美团； 21、公告精选︱长江电力：控股股东拟增持40亿元-80亿元公司股份；景旺电子：拟50亿元对景旺电子 珠海金湾基地进行扩产投资； 大中华区要闻： 1、鲍威尔暗示或在9月降息，美股大涨！道指再创新高，中概指数涨2.73%； 2、以太坊突破4900美元/枚，创历史新高； 3、惠誉：确认美国"AA+"评级，展望稳定； 4、鲍威尔"加入"鸽派阵营 称经济风险为降息提供更足理由； 5、特朗普宣布对家具进口展开关税调查； 6、美国小额关税豁免将到期 多国邮政公司暂停接收往美包裹； 7、加拿大做出让步！宣布取消多项针对美国商品的报复性关税； 8、英特尔：获美国政府按20.47美元/股认购10%股份； 9、苹果拟考虑使用谷歌的Gemini模型来支持新版Siri； 10、世界钢铁协会：7月全球粗钢产量为1.501亿吨，同比减少1.3%； 1、沪指续创十年新高！"市值/GDP"指标仍在历史中枢，大幅低于其他全球主要经济体； 2、沪指时隔10年重上3800点 高盛：对冲基金以7周来最快速度净买入中国股票； 3、中信证券："存款搬家"信号出现，流出规模 ...

今天分享的是：2025年从"千人一面"到"一人千面"：人工智能引领广告行业智能化转型报告 报告共计：47页 2025年《从"千人一面"到"一人千面"：人工智能引领广告行业智能化转型报告》（腾讯研究院与腾讯广告联合出品）指出， 广告业作为国民经济关键支柱，2023年数字广告渗透率达83.4%，其发展始终由技术驱动，当前生成式AI推动广告从计算广告 迈入智能广告时代。AI深度赋能数字广告全链路，在素材生产上，呈现"自动化"与"精品化"两极，效果导向的"跑量型"广告 依赖AIGC实现批量生成，品牌导向的"精品型"广告仍需人类创意；在投放优化上，AI通过生成式召回、数据增强、多模态理 解突破传统广告引擎困境，推动匹配范式从"千人千面"向"一人千面"升级，实现供需智能匹配；在形态升维上，广告入口从 用户界面转向智能代理，形态从展示位转向对话式嵌入，催生无感嵌入式原生广告，如游戏影视中的动态品牌元素生成、社 交平台的内容植入等。全球实践中，谷歌以Gemini模型构建生态协同，Meta依托Llama4模型推进广告沉浸化，腾讯整合混元 大模型与广告系统提升投放效果，快手通过可灵AI降低中小企业广告成本；广告服务商如微盟以多层 ...

马斯克曾邀扎克伯格助其收购OpenAI。 ③杭州：新学期将在中小学全面开展AI通识课程，每学年不少于10课时。 每字年不少污了10張的 1 ■ W 载 s a well his - 来源： = 财 快 讯 ①马斯克曾邀扎克伯格助其收购OpenAI； ②苹果拟考虑使用谷歌的Gemini模型来支持新版Siri； ...

【原油】纽约商品交易所10月交货的轻质原油期货价格上涨14美分，收于每桶63.66美元，涨幅为 0.22%；10月交货的伦敦布伦特原油期货价格上涨6美分，收于每桶67.73美元，涨幅为0.09%。 【外汇】衡量美元对六种主要货币的美元指数当天下跌0.92%，在汇市尾市收于97.716。截至纽约汇市 尾市，1欧元兑换1.1722美元，高于前一交易日的1.1610美元；1英镑兑换1.3526美元，高于前一交易日 的1.3419美元。1美元兑换146.87日元，低于前一交易日的148.35日元；1美元兑换0.8011瑞士法郎，低于 前一交易日的0.8089瑞士法郎；1美元兑换1.3825加元，低于前一交易日的1.3899加元；1美元兑换9.4971 瑞典克朗，低于前一交易日的9.6200瑞典克朗。 【美股】截至收盘，道指涨846.24点，涨幅为1.89%，报45631.74点；纳指涨396.22点，涨幅为1.88%， 报21496.53点；标普500指数涨96.74点，涨幅为1.52%，报6466.91点。周五盘中，道指最高上涨至 45787.84点，再创盘中新高。谷歌(GOOG.US)收涨3.17%，英特尔(IN ...

智通财经APP获悉，周五，三大指数大涨。美联储主席鲍威尔表示，不断"变化"的经济风险让美联储有 了更充分的降息理由。这番言论表明鲍威尔站到了负责制定利率的联邦公开市场委员会中"鸽派"阵营， 也释放出他可能支持美联储在9月的下次会议上降息25个基点的信号。 尽管鲍威尔承认政府贸易战对消费者物价的影响如今"清晰可见"，但他暗示这种影响不太可能持续，反 而可能只是一次性冲击，央行可以不予理会。他表示："鉴于劳动力市场并非特别紧张，且面临越来越 大的下行风险，（通胀持续上升的）这种结果似乎不太可能出现。"他还补充道，"通胀面临上行风险， 而就业则面临下行风险，这是个颇具挑战的局面"。交易员增加了对美联储下月降息的押注。交易员现 在认为美联储9月份降息的可能性接近90%，而鲍威尔发表讲话前的可能性约为75%。 【美股】截至收盘，道指涨846.24点，涨幅为1.89%，报45631.74点；纳指涨396.22点，涨幅为1.88%， 报21496.53点；标普500指数涨96.74点，涨幅为1.52%，报6466.91点。周五盘中，道指最高上涨至 45787.84点，再创盘中新高。谷歌(GOOG.US)收涨3.17%，英特 ...

公众号记得加星标⭐️，第一时间看推送不会错过。 来源 ：内容 编译自 nextplatform 。 每两年晶体管数量增加一倍，从而将晶体管价格降低一半，因为可以在给定面积上塞入两倍的晶体 管，这在 CMOS 芯片时代改变了计算方式并推动了它的发展。 但摩尔定律的数学意义还在于此。每次略微提高功率预算，就能将设备性能每年提升 40%。对于计 算引擎而言，这意味着每年的升级频率足以满足大多数公司事务性和分析性工作负载的增长需求。而 对于那些需要更多性能的公司，对称多处理 (SMP) 和非均匀内存访问 (NUMA) 集群技术可以使多个 设备看起来像一个大设备——我们称之为"纵向扩展"。 随着 20 世纪 90 年代末 Web 2.0 互联网时代的到来，一台机器即使配备了 SMP 或 NUMA，这种所 谓的通过共享内存扩展来提高计算能力的做法也显得力不从心。（有趣的是，巧合的是，当时超级计 算中奇特的联邦 NUMA 配置（如今我们可以使用现成的 PCI-Express 交换机轻松构建）也不足以支 持传统的 HPC 仿真和建模工作负载，但其规模要大得多，对于企业和像谷歌这样尚未进入超大规模 的网络公司来说成本太高。）因 ...

Investment Rating - The report maintains a "Positive" investment rating for the computer industry [5] Core Insights - The AI application sector is entering a commercialization phase, presenting numerous investment opportunities in both AI application and computing power industries [3][8] - Major tech companies are increasingly integrating AI into their core business operations, driving significant revenue growth [8][47] Summary by Sections Section 1: Rapid Model Iteration and AI Business Realization - The competition among global tech companies in large model development is intensifying, with 11 companies releasing or iterating 32 versions of large models within approximately 220 days [12][18] - The introduction of reinforcement learning has significantly enhanced model inference capabilities, leading to the rapid development of various applications in programming, law, sales, and design [17][20] Section 2: Increased Computing Power Consumption and Capital Expenditure - Microsoft leads the AI infrastructure investment, with capital expenditures reaching $88.2 billion in the 2025 fiscal year, followed by Google and Meta with $85 billion and $66-72 billion respectively [8][71] - The rapid growth in AI token consumption is driving tech giants to increase their capital expenditures to support AI infrastructure [71] Section 3: Investment Recommendations and Targets - Investors are advised to focus on the overall opportunities in the AI application sector, with specific companies highlighted for potential investment, including: - AI Application Sector: Zhuoyi Information, Keda Xunfei, Tax Friend, Kingsoft Office, Dingjie Smart, Hand Information, Focus Technology, Fanwei Network, Caixun Co., Wanjing Technology, Foxit Software, Dongtu Technology, New Zhisoft [3] - AI Computing Power Sector: Cambrian-U, Haiguang Information, Runjian Co., Xiechuang Data, Aofei Data, Capital Online, Yike De-W, Guanghuan New Network, Runze Technology [3]

Core Viewpoint - Google's "AI Overview" tool has become a key defense against AI competition, leading to significant growth in search ad exposure and revenue [1][3]. Group 1: Financial Performance - Alphabet's search revenue reached a record $54.2 billion in Q2, a 12% year-over-year increase, surpassing analyst expectations of $52.9 billion [2]. - The monthly active users of the "AI Overview" tool exceeded 2 billion, a substantial increase from 1.5 billion in the previous quarter [2]. Group 2: Impact of AI Overview Tool - The introduction of the AI Overview tool has resulted in a 49% increase in search ad exposure over the past year [3]. - CEO Sundar Pichai noted that AI is expanding how people search and access information, leading to more searches as users realize their needs can be met [3]. Group 3: Market Environment - Despite fluctuations in the advertising market due to tariffs and macroeconomic uncertainties, the second quarter saw a year-over-year growth in advertising budgets, supporting Google's strong search revenue performance [5]. Group 4: Future Challenges - Although current performance is strong, Google's search engine resilience faces future challenges, including a decline in the number of user clicks on revenue-generating links due to the AI Overview providing direct answers [6]. - Emerging AI-driven web browsers from startups like Perplexity and similar products being developed by OpenAI could change how users access information, posing a challenge to Google's Chrome browser [6]. Group 5: Defensive Measures - In response to these threats, Google plans to increase its capital expenditure from $75 billion to $85 billion by 2025, with further spending increases anticipated for the next year [8]. - Google aims to maintain its competitive edge by modifying the Chrome browser, integrating Gemini into more products, and developing unique AI features like "circle search" on Android devices [8]. Group 6: Historical Context - Historically, Google has taken defensive actions in response to emerging threats, such as acquiring Android when search shifted to mobile and paying Apple billions to make its search the default on Safari [9].

Group 1 - The core viewpoint of the articles highlights that despite concerns about AI startups threatening Google's search engine dominance, Alphabet's latest financial report shows record search revenue of $54.2 billion in Q2, a 12% year-over-year increase, surpassing analyst expectations of $52.9 billion [1] - Alphabet's CEO Sundar Pichai announced that the monthly active users of the "AI Overview" tool have exceeded 2 billion, a significant increase from 1.5 billion in the previous quarter [1] - The introduction of the AI Overview tool has led to a 49% increase in search ad impressions over the past year, indicating a positive impact on advertising performance [1] Group 2 - Despite strong current performance, the resilience of Google's search engine faces future challenges, including a decline in the number of user clicks on revenue-generating links due to the AI Overview providing direct answers [2] - Emerging AI-driven web browsers from startups like Perplexity and similar products being developed by OpenAI may alter how users access information, posing a potential threat to Google's Chrome browser [2] Group 3 - In response to these threats, Google plans to increase its capital expenditure from $75 billion to $85 billion by 2025, with further spending increases anticipated for the following year [3] - Google aims to maintain its competitive edge by modifying the Chrome browser, integrating Gemini into more products, and developing unique AI features like "circle search" on Android devices [3] Group 4 - Historically, Google has taken defensive actions in response to emerging threats, such as acquiring the Android operating system when search shifted to mobile and paying Apple billions annually to make its search the default option on Safari [4] - During the AI surge initiated by Microsoft's Bing, Google increased its investment in AI computing, which ultimately prevented Microsoft from significantly eroding its market share [4]

Core Viewpoint - Google recently announced that its DeepMind team won an IMO gold medal, but shortly after, three key team members were reported to have left for Meta, highlighting a talent drain in the AI sector [1][19]. Group 1: Key Personnel Changes - Three critical figures involved in the training of the Gemini model, Du Yu, Tianhe Yu, and Wang Weiyue, have left Google for Meta [2][3]. - Du Yu has been a significant contributor to the Gemini series models and has worked on Google's conversational AI products [9]. - Tianhe Yu, a research scientist at Google DeepMind, was responsible for the reinforcement learning and training of Gemini, playing a key role in the release of Gemini 2.5 [10]. - Wang Weiyue, a principal research engineer at Google DeepMind, contributed to Gemini 2.5 Pro and has a background in computer vision [13][14]. Group 2: Competitive Landscape - Mark Zuckerberg's recruitment of talent from Google is part of a broader trend, as Microsoft has also been reported to have poached over 20 talents from Google DeepMind [19]. - Amar Subramanya, the former engineering lead for Gemini, has joined Microsoft AI as a vice president, indicating a shift in talent dynamics within the AI industry [19]. - The talent acquisition efforts by Microsoft have been ongoing for six months, led by Mustafa Suleyman, a co-founder of DeepMind, adding a layer of complexity to the competitive landscape [21].