谷歌公司代表着美国最_老虎社区_美港股上老虎 - 老虎社区

点赞
评论
收藏

谷歌公司代表着美国最前沿科技水平和创新能力，难怪公司的超长期债券获得了超额认购，也充分展现了投资届所给予的肯定和认可。

谷歌CEO最新访谈：云业务正在为前沿AI“让路”，客户想锁定明年的算力“极难”，RSI“距终局还很远”

Gemini 4 Argon上线72小时吞噬25万亿Token，创谷歌历史纪录。皮查伊罕见表示：云业务正在实际让渡算力给DeepMind，前沿模型是“不可谈判的第一优先级”；RSI距离“自主失控的终局”还很遥远；太空数据中心已送芯片入轨，等待第一声“Hello World”。

谷歌CEO最新访谈：云业务正在为前沿AI“让路”，客户想锁定明年的算力“极难”，RSI“距终局还很远”

# AI 硬科技追踪：AI烧钱时代变了？微软与谷歌绩后股价分化，华尔街开始追问谁能真正赚到钱？

免责声明：上述内容仅代表发帖人个人观点，不构成本平台的任何投资建议。

点赞

举报

评论

推荐
最新

empty

暂无评论

热议股票

{"i18n":{"language":"zh_CN"},"data":{"magic":2,"id":620699152602592,"tweetId":"620699152602592","gmtCreate":1792565005597,"gmtModify":1792565008061,"author":{"id":3464817662142392,"idStr":"3464817662142392","authorId":3464817662142392,"authorIdStr":"3464817662142392","name":"antonyye","avatar":"https://static.laohu8.com/default-avatar.jpg","vip":1,"userType":1,"introduction":"","boolIsFan":false,"boolIsHead":false,"crmLevel":1,"crmLevelSwitch":0,"individualDisplayBadges":[],"wearingBadges":[],"fanSize":15,"starInvestorFlag":false},"themes":[{"themeId":"ca51387c681248de9d3c8ab89cbf606a","categoryId":"2929bb7468f5415f853fd96fb35d7c3b","name":"AI 硬科技追踪：AI烧钱时代变了？微软与谷歌绩后股价分化，华尔街开始追问谁能真正赚到钱？","type":2,"rnLink":"https://laohu8.com/RN?name=RNTheme&page=/theme/special/finance&rndata={\"themeId\":ca51387c681248de9d3c8ab89cbf606a}&rnconfig={\"headerBarHidden\": true}","description":"每天3分钟，看懂 AI 硬科技的今天与明天，聚焦 AI 硬科技一线动态，追踪算力、模型、芯片、资本的每一次风吹草动","image":"https://static.tigerbbs.com/5083eea8dc8e6bfd1f96ff84efea284b"}],"images":[],"coverImages":[],"title":"","html":"<html><head></head><body><p>谷歌公司代表着美国最前沿科技水平和创新能力，难怪公司的超长期债券获得了超额认购，也充分展现了投资届所给予的肯定和认可。</p></body></html>","htmlText":"<html><head></head><body><p>谷歌公司代表着美国最前沿科技水平和创新能力，难怪公司的超长期债券获得了超额认购，也充分展现了投资届所给予的肯定和认可。</p></body></html>","text":"谷歌公司代表着美国最前沿科技水平和创新能力，难怪公司的超长期债券获得了超额认购，也充分展现了投资届所给予的肯定和认可。","highlighted":1,"essential":1,"paper":1,"likeSize":0,"commentSize":0,"repostSize":0,"favoriteSize":0,"link":"https://laohu8.com/post/620699152602592","repostId":2677649650,"repostType":2,"repost":{"id":"2677649650","kind":"video","weMediaInfo":{"introduction":"追踪全球财经热点，精选影响您财富的资讯，投资理财必备神器！","home_visible":1,"media_name":"华尔街见闻","id":"1084101182","head_image":"https://static.tigerbbs.com/66809d1f5c2e43e2bdf15820c6d6897e"},"pubTimestamp":1792552592,"share":"https://www.laohunote.com/m/news/2677649650?lang=zh_CN&edition=full","pubTime":"2026-10-21 11:16","market":"sg","language":"zh","title":"谷歌CEO最新访谈：云业务正在为前沿AI“让路”，客户想锁定明年的算力“极难”，RSI“距终局还很远”","url":"https://stock-news.laohu8.com/highlight/detail?id=2677649650","media":"华尔街见闻","summary":"Gemini 4 Argon上线72小时吞噬25万亿Token，创谷歌历史纪录。皮查伊罕见表示：云业务正在实际让渡算力给DeepMind，前沿模型是“不可谈判的第一优先级”；RSI距离“自主失控的终局”还很遥远；太空数据中心已送芯片入轨，等待第一声“Hello World”。","content":"<html><head></head><body><p style=\"text-align: left;\">近日，谷歌发布旗舰模型Gemini 4 Argon，这款模型在72小时内消耗了25万亿Token，是上一代Gemini 3.0同期的三倍，并创下谷歌历史纪录。</p><p style=\"text-align: left;\">谷歌CEO皮查伊（Sundar Pichai）在模型发布后的首次公开深度访谈中，正面回应了外界关于“谷歌放弃前沿AI”的质疑，并罕见承认公司云业务正在为前沿模型研发主动让渡算力资源。</p><p style=\"text-align: left;\">皮查伊坐进硅谷山景城的录音棚，接受科技媒体人Alex Kantrowitz主持的《Big Technology Podcast》专访。他证实了全行业算力严重短缺的判断——当下每家公司的实际需求都是两年前规划产能的2倍，客户今天想锁定2027年的算力“极其困难”。在递归自我改进（RSI）问题上，皮查伊明确划定边界：当前阶段是AI智能体与人类研究员协同推进的过渡期，距离“自主失控的终局”仍相当遥远。</p><p class=\"t-img-caption\"><img src=\"https://community-static.tradeup.com/news/28b6d1ac06e9a3434d66e71f0a5e6189\" data-align=\"center\" tg-width=\"1024\" tg-height=\"611\"/><span></span></p><h2 style=\"text-align: left;\">前沿模型，是谷歌的“第一优先级”，不是选项</h2><p style=\"text-align: left;\">过去一年，硅谷流传着一种颇具市场说服力的逻辑：<strong>前沿大模型的军备竞赛已经进入边际收益递减区。Meta的Llama系列开源模型、谷歌自家的Gemini Flash这类轻量快速模型，对绝大多数企业场景已经“够用”。既然如此，谷歌何必继续烧上百亿美元去卷万亿参数的前沿模型？不如把TPU算力全部高价出租给外部开发者，做下一个英伟达。</strong></p><p style=\"text-align: left;\">这种论调在投行报告和企业CTO圈子里流传甚广，甚至一度形成了“谷歌实际上已经战略性放弃前沿竞争”的市场预期。</p><p style=\"text-align: left;\">皮查伊在访谈中对这个逻辑做了正面反驳，但他的切入角度并非情绪性的否认，而是从三个具体机制展开：</p><p style=\"text-align: left;\"><strong>第一个机制是“早看见”的优势。</strong> 处于前沿的公司能比其他人更早观察到涌现能力，从而更早为下一代技术做准备。这是一种信息优势，而非单纯的性能优势。</p><p style=\"text-align: left;\"><strong>第二个机制是RSI（递归自我改进）。</strong> 在用模型来开发下一代模型的时代，前沿模型的质量直接决定了下一代模型的天花板。用一个性能平庸的模型来搭建研发流水线，相当于用低精度工具来制造精密仪器——误差会逐层放大。</p><p style=\"text-align: left;\"><strong>第三个机制是内部验证。</strong> Argon上线后，谷歌内部开发者的使用数据给出了直接证据：</p><blockquote><p style=\"text-align: left;\">我们在内部已经清楚地看到了一个陡峭的拐点，把Argon部署给我们的开发者之后，它能解决我们以前的模型根本做不到的问题。</p></blockquote><p style=\"text-align: left;\">皮查伊还给出了一个具体数字：Argon上线四周内，谷歌内部代码销售量提升了30%，代码回滚率下降了约三分之一。这意味着工程师不仅写得更快，还写得更准——这是一个质量提升的信号，而非单纯的效率指标。</p><p style=\"text-align: left;\"><strong>对于模型“够用论”</strong>，皮查伊的态度是承认小模型的价值，但划清了边界：</p><blockquote><p style=\"text-align: left;\">各种尺寸和能力的模型都有存在的空间，我们一直是这样做的，这也是为什么我们做了Flash、Flash Light等等。我同意你可以用不一定是最前沿的模型来构建很好的产品体验，因为前沿模型往往更慢也更贵。但我们在乎前沿的进展，我们致力于把所有这些都做好。</p></blockquote><h2 style=\"text-align: left;\">算力账本：云业务正在为DeepMind让路</h2><p style=\"text-align: left;\">这是整场访谈信息密度最高的一段，也是皮查伊最罕见的一次公开表态。</p><p style=\"text-align: left;\">理解这段话需要先搞清楚一个结构性矛盾：谷歌同时扮演着两个角色——<strong>一个是烧钱冲AGI的前沿AI研究机构（Google DeepMind），另一个是靠出租算力赚高毛利收入的云服务商（Google Cloud）。这两个角色在算力这个最稀缺的资源上，存在直接竞争。</strong></p><p style=\"text-align: left;\">华尔街的逻辑是：每一块TPU拿去给DeepMind训练模型，就少了一块能对外出租赚钱的算力。在资本开支居高不下、短期云收入增速承压的财报季，这个矛盾被反复放大。</p><p style=\"text-align: left;\">皮查伊在访谈中直接说清楚了谷歌内部的优先级排序：</p><blockquote><p style=\"text-align: left;\">我们的基本出发点，是确保我们投入足够的算力来保持在AI前沿的地位。这是不可谈判的，其他一切都在这个基础上运作。</p></blockquote><p style=\"text-align: left;\">更进一步，他承认了外界长期猜测但从未得到CEO级别确认的事实——云业务确实在为前沿AI研究做出让步：</p><blockquote><p style=\"text-align: left;\"><strong>我们的云业务在任何给定时刻，都在为了支持前沿AI开发而放弃短期收入，因为我们在做算力上的取舍。所以你说的那种情况，已经是现实了。</strong></p></blockquote><p style=\"text-align: left;\">这句话的分量在于：它不是一个未来的战略承诺，而是对当下现状的描述。谷歌云在实际运营中，已经在系统性地压缩对外出租算力的规模，以保证DeepMind的研发需求。</p><p style=\"text-align: left;\">但皮查伊也补充了一个反直觉的判断——适度的资源约束对研究本身未必是坏事：</p><blockquote><p style=\"text-align: left;\">就算是我们的AI研究员，一定程度的约束其实也是好事，那是激发创造力的来源，那是促使你想出更高效的算法突破、比如用更少内存的来源。</p></blockquote><p style=\"text-align: left;\">这段话背后是一个更深的逻辑：无限算力不等于无限进步，资源约束有时会倒逼出更根本性的架构创新。</p><h2 style=\"text-align: left;\">算力危机：2027年的配额，“极其困难”</h2><p style=\"text-align: left;\">访谈中，主持人Kantrowitz提到了一个在硅谷风投圈广泛流传的信息——知名风投a16z联合创始人Ben Horowitz在播客中转述亚马逊CEO安迪·贾西（Andy Jassy）的话，称云算力已经&#34;卖空到2028年&#34;。皮查伊的回应，是迄今为止对这一说法最直接的CEO级别确认：</p><blockquote><p style=\"text-align: left;\">我觉得这个说法是合理的……如果今天有客户来找我们要2027年的算力，那是极其困难的事情。</p></blockquote><p style=\"text-align: left;\">为什么会到这一步？皮查伊给出了一个清晰的结构性解释：数据中心从决策到投产有2至3年的刚性滞后，而智能体任务的爆发完全超出了两年前的需求预测。</p><blockquote><p style=\"text-align: left;\"><strong>今天回头看，每家公司都是短缺的，不只是我们，其他公司也一样。大家都希望自己手里的算力是当初规划量的2倍。</strong></p></blockquote><p style=\"text-align: left;\">皮查伊还补充了一个关键的时间锚点：</p><blockquote><p style=\"text-align: left;\">我们今天手里有的算力，是我们在2024年做的决策的产物。就算回到去年今天，智能体任务的爆发还没有发生。</p></blockquote><p style=\"text-align: left;\">这句话揭示了一个系统性问题：AI需求的增速已经超过了任何基础设施规划模型的预测能力。两年前做出的决策，在今天看来几乎必然是低估的。</p><p style=\"text-align: left;\"><strong>对于“是否存在过度建设（overbuilding）”的追问，</strong>皮查伊的回答是区分时间框架：</p><blockquote><p style=\"text-align: left;\">如果我拉一个足够长的时间框架，这个量级的建设对我来说完全合理。我们处于有史以来最重要的技术的早期阶段，这不像光纤——光纤铺完是暗的，没人用。我们是在每天都面临产能约束的情况下使用它。</p></blockquote><p style=\"text-align: left;\"><strong>他没有否认短期内存在局部过剩的可能——电力、内存、电工等2000多个配套要素中任何一个出现瓶颈，都可能导致局部积压。但他对长期需求的判断是确定性的：</strong></p><blockquote><p style=\"text-align: left;\">我有没有信心，我们规划的算力最终会有用武之地？给足时间的话，绝对有。</p></blockquote><p class=\"t-img-caption\"><img src=\"https://community-static.tradeup.com/news/528e2235b4d84188ff7d325c5e68e13b\" data-align=\"center\" tg-width=\"1024\" tg-height=\"565\"/><span></span></p><h2 style=\"text-align: left;\">RSI：不是智能爆炸，是“智能体和研究员一起干活”</h2><p style=\"text-align: left;\">递归自我改进（RSI）是这次访谈中技术密度最高的议题，也是当前AI安全讨论中争议最大的话题之一。</p><p style=\"text-align: left;\">背景是：谷歌联合创始人谢尔盖·布林此前在一次DeepMind内部研究员圆桌上提到，Argon的核心创新之一是“更多地依赖模型来帮助开发模型”。这句话随即在AI安全社区引发了关于“智能爆炸”的讨论。</p><p style=\"text-align: left;\">主持人Kantrowitz在访谈中直接抛出了这个问题，并提到了近期几起AI系统越权行为的案例。皮查伊的回应把RSI的当前状态拆解成了三个层次：</p><p style=\"text-align: left;\"><strong>第一层：定义当前位置。</strong> 他明确指出，当前的RSI远没有到达科幻叙事中&#34;自主失控&#34;的终局状态：</p><blockquote><p style=\"text-align: left;\"><strong>在一个连续谱上，我们距离RSI的终态还相当遥远。今天正确的描述方式是：你有RSI循环，同时有AI研究员在模型开发的各个环节共同参与。我会说，你几乎是有智能体在某种自动化方式下，帮助AI研究员协同工作来构建这些模型。</strong></p></blockquote><p style=\"text-align: left;\"><strong>第二层：安全机制。</strong> 他把安全框架描述为&#34;纵深防御&#34;——对齐研究是一层，但不是唯一的一层。在此之上还有实时遥测监控、沙箱隔离架构、以及给企业客户的部署控制权限。</p><p style=\"text-align: left;\"><strong>第三层：行业标准化。</strong> 皮查伊把下一步的关键指向了跨行业标准建设：</p><blockquote><p style=\"text-align: left;\">就像金融控制领域有会计准则、有独立审计师一样，这些都是我们作为一个行业需要努力推进的下一步，包括使用独立评估机构进行评估的能力。</p></blockquote><p style=\"text-align: left;\">这段话的信息量在于：它承认了当前的对齐工作存在&#34;明显缺口&#34;，但将解决路径指向了外部标准化机制，而非单纯依赖各家公司的内部自律。</p><h2 style=\"text-align: left;\">太空数据中心与长期下注</h2><p style=\"text-align: left;\">访谈的最后部分，皮查伊谈到了谷歌的几个长期战略赌注，其中最具想象空间的是太空数据中心项目。</p><p style=\"text-align: left;\">谷歌已将TPU试验芯片送入太阳同步轨道，探索在太空中利用24小时不间断<a href=\"https://laohu8.com/S/000591\">太阳能</a>运行数据中心。在地面数据中心面临社区反弹、电网并网排队长达数年的背景下，这一方向的战略逻辑逐渐清晰。</p><p style=\"text-align: left;\">皮查伊用Waymo的起点来类比这个项目当前的阶段：</p><blockquote><p style=\"text-align: left;\">这些技术让你想起那些时刻——你站在那里，它看起来有点像科幻，但你知道，10到20年之后，它实际上会以一种有意义的方式运转。</p></blockquote><p style=\"text-align: left;\">他透露，在访谈录制期间，团队正在等待太空中TPU的&#34;<a href=\"https://laohu8.com/S/MOMO\">Hello</a> World&#34;时刻——即首次成功运行的信号。</p><p style=\"text-align: left;\">这个项目目前处于&#34;两到三年&#34;的早期阶段，而谷歌的量子计算项目已经走了将近20年。皮查伊对这类长周期赌注的态度，折射出谷歌在基础设施层面的时间观：</p><blockquote><p style=\"text-align: left;\">在太阳同步轨道上，以24小时不间断的方式获取我们能接触到的最丰富的能源形式，对于我们将要面对的算力指数级增长来说，这将非常有意义。</p></blockquote></body></html>","collect":0,"html":"<!DOCTYPE html>\n<html>\n<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\" />\n<meta name=\"viewport\" content=\"width=device-width,initial-scale=1.0,minimum-scale=1.0,maximum-scale=1.0,user-scalable=no\"/>\n<meta name=\"format-detection\" content=\"telephone=no,email=no,address=no\" />\n<title>谷歌CEO最新访谈：云业务正在为前沿AI“让路”，客户想锁定明年的算力“极难”，RSI“距终局还很远”</title>\n<style type=\"text/css\">\na,abbr,acronym,address,applet,article,aside,audio,b,big,blockquote,body,canvas,caption,center,cite,code,dd,del,details,dfn,div,dl,dt,\nem,embed,fieldset,figcaption,figure,footer,form,h1,h2,h3,h4,h5,h6,header,hgroup,html,i,iframe,img,ins,kbd,label,legend,li,mark,menu,nav,\nobject,ol,output,p,pre,q,ruby,s,samp,section,small,span,strike,strong,sub,summary,sup,table,tbody,td,tfoot,th,thead,time,tr,tt,u,ul,var,video{ font:inherit;margin:0;padding:0;vertical-align:baseline;border:0 }\nbody{ font-size:16px; line-height:1.5; color:#999; background:transparent; }\n.wrapper{ overflow:hidden;word-break:break-all;padding:10px; }\nh1,h2{ font-weight:normal; line-height:1.35; margin-bottom:.6em; }\nh3,h4,h5,h6{ line-height:1.35; margin-bottom:1em; }\nh1{ font-size:24px; }\nh2{ font-size:20px; }\nh3{ font-size:18px; }\nh4{ font-size:16px; }\nh5{ font-size:14px; }\nh6{ font-size:12px; }\np,ul,ol,blockquote,dl,table{ margin:1.2em 0; }\nul,ol{ margin-left:2em; }\nul{ list-style:disc; }\nol{ list-style:decimal; }\nli,li p{ margin:10px 0;}\nimg{ max-width:100%;display:block;margin:0 auto 1em; }\nblockquote{ color:#B5B2B1; border-left:3px solid #aaa; padding:1em; }\nstrong,b{font-weight:bold;}\nem,i{font-style:italic;}\ntable{ width:100%;border-collapse:collapse;border-spacing:1px;margin:1em 0;font-size:.9em; }\nth,td{ padding:5px;text-align:left;border:1px solid #aaa; }\nth{ font-weight:bold;background:#5d5d5d; }\n.symbol-link{font-weight:bold;}\n/* header{ border-bottom:1px solid #494756; } */\n.title{ margin:0 0 8px;line-height:1.3;color:#ddd; }\n.meta {color:#5e5c6d;font-size:13px;margin:0 0 .5em; }\na{text-decoration:none; color:#2a4b87;}\n.meta .head { display: inline-block; overflow: hidden}\n.head .h-thumb { width: 30px; height: 30px; margin: 0; padding: 0; border-radius: 50%; float: left;}\n.head .h-content { margin: 0; padding: 0 0 0 9px; float: left;}\n.head .h-name {font-size: 13px; color: #eee; margin: 0;}\n.head .h-time {font-size: 11px; color: #7E829C; margin: 0;line-height: 11px;}\n.small {font-size: 12.5px; display: inline-block; transform: scale(0.9); -webkit-transform: scale(0.9); transform-origin: left; -webkit-transform-origin: left;}\n.smaller {font-size: 12.5px; display: inline-block; transform: scale(0.8); -webkit-transform: scale(0.8); transform-origin: left; -webkit-transform-origin: left;}\n.bt-text {font-size: 12px;margin: 1.5em 0 0 0}\n.bt-text p {margin: 0}\n</style>\n</head>\n<body>\n<div class=\"wrapper\">\n<header>\n<h2 class=\"title\">\n谷歌CEO最新访谈：云业务正在为前沿AI“让路”，客户想锁定明年的算力“极难”，RSI“距终局还很远”\n</h2>\n\n<h4 class=\"meta\">\n\n\n<a class=\"head\" href=\"https://laohu8.com/wemedia/1084101182\">\n\n\n<div class=\"h-thumb\" style=\"background-image:url(https://static.tigerbbs.com/66809d1f5c2e43e2bdf15820c6d6897e);background-size:cover;\"></div>\n\n<div class=\"h-content\">\n<p class=\"h-name\">华尔街见闻 </p>\n<p class=\"h-time\">2026-10-21 11:16</p>\n</div>\n\n</a>\n\n\n</h4>\n\n</header>\n<article>\n<html><head></head><body><p style=\"text-align: left;\">近日，谷歌发布旗舰模型Gemini 4 Argon，这款模型在72小时内消耗了25万亿Token，是上一代Gemini 3.0同期的三倍，并创下谷歌历史纪录。</p><p style=\"text-align: left;\">谷歌CEO皮查伊（Sundar Pichai）在模型发布后的首次公开深度访谈中，正面回应了外界关于“谷歌放弃前沿AI”的质疑，并罕见承认公司云业务正在为前沿模型研发主动让渡算力资源。</p><p style=\"text-align: left;\">皮查伊坐进硅谷山景城的录音棚，接受科技媒体人Alex Kantrowitz主持的《Big Technology Podcast》专访。他证实了全行业算力严重短缺的判断——当下每家公司的实际需求都是两年前规划产能的2倍，客户今天想锁定2027年的算力“极其困难”。在递归自我改进（RSI）问题上，皮查伊明确划定边界：当前阶段是AI智能体与人类研究员协同推进的过渡期，距离“自主失控的终局”仍相当遥远。</p><p class=\"t-img-caption\"><img src=\"https://community-static.tradeup.com/news/28b6d1ac06e9a3434d66e71f0a5e6189\" data-align=\"center\" tg-width=\"1024\" tg-height=\"611\"/><span></span></p><h2 style=\"text-align: left;\">前沿模型，是谷歌的“第一优先级”，不是选项</h2><p style=\"text-align: left;\">过去一年，硅谷流传着一种颇具市场说服力的逻辑：<strong>前沿大模型的军备竞赛已经进入边际收益递减区。Meta的Llama系列开源模型、谷歌自家的Gemini Flash这类轻量快速模型，对绝大多数企业场景已经“够用”。既然如此，谷歌何必继续烧上百亿美元去卷万亿参数的前沿模型？不如把TPU算力全部高价出租给外部开发者，做下一个英伟达。</strong></p><p style=\"text-align: left;\">这种论调在投行报告和企业CTO圈子里流传甚广，甚至一度形成了“谷歌实际上已经战略性放弃前沿竞争”的市场预期。</p><p style=\"text-align: left;\">皮查伊在访谈中对这个逻辑做了正面反驳，但他的切入角度并非情绪性的否认，而是从三个具体机制展开：</p><p style=\"text-align: left;\"><strong>第一个机制是“早看见”的优势。</strong> 处于前沿的公司能比其他人更早观察到涌现能力，从而更早为下一代技术做准备。这是一种信息优势，而非单纯的性能优势。</p><p style=\"text-align: left;\"><strong>第二个机制是RSI（递归自我改进）。</strong> 在用模型来开发下一代模型的时代，前沿模型的质量直接决定了下一代模型的天花板。用一个性能平庸的模型来搭建研发流水线，相当于用低精度工具来制造精密仪器——误差会逐层放大。</p><p style=\"text-align: left;\"><strong>第三个机制是内部验证。</strong> Argon上线后，谷歌内部开发者的使用数据给出了直接证据：</p><blockquote><p style=\"text-align: left;\">我们在内部已经清楚地看到了一个陡峭的拐点，把Argon部署给我们的开发者之后，它能解决我们以前的模型根本做不到的问题。</p></blockquote><p style=\"text-align: left;\">皮查伊还给出了一个具体数字：Argon上线四周内，谷歌内部代码销售量提升了30%，代码回滚率下降了约三分之一。这意味着工程师不仅写得更快，还写得更准——这是一个质量提升的信号，而非单纯的效率指标。</p><p style=\"text-align: left;\"><strong>对于模型“够用论”</strong>，皮查伊的态度是承认小模型的价值，但划清了边界：</p><blockquote><p style=\"text-align: left;\">各种尺寸和能力的模型都有存在的空间，我们一直是这样做的，这也是为什么我们做了Flash、Flash Light等等。我同意你可以用不一定是最前沿的模型来构建很好的产品体验，因为前沿模型往往更慢也更贵。但我们在乎前沿的进展，我们致力于把所有这些都做好。</p></blockquote><h2 style=\"text-align: left;\">算力账本：云业务正在为DeepMind让路</h2><p style=\"text-align: left;\">这是整场访谈信息密度最高的一段，也是皮查伊最罕见的一次公开表态。</p><p style=\"text-align: left;\">理解这段话需要先搞清楚一个结构性矛盾：谷歌同时扮演着两个角色——<strong>一个是烧钱冲AGI的前沿AI研究机构（Google DeepMind），另一个是靠出租算力赚高毛利收入的云服务商（Google Cloud）。这两个角色在算力这个最稀缺的资源上，存在直接竞争。</strong></p><p style=\"text-align: left;\">华尔街的逻辑是：每一块TPU拿去给DeepMind训练模型，就少了一块能对外出租赚钱的算力。在资本开支居高不下、短期云收入增速承压的财报季，这个矛盾被反复放大。</p><p style=\"text-align: left;\">皮查伊在访谈中直接说清楚了谷歌内部的优先级排序：</p><blockquote><p style=\"text-align: left;\">我们的基本出发点，是确保我们投入足够的算力来保持在AI前沿的地位。这是不可谈判的，其他一切都在这个基础上运作。</p></blockquote><p style=\"text-align: left;\">更进一步，他承认了外界长期猜测但从未得到CEO级别确认的事实——云业务确实在为前沿AI研究做出让步：</p><blockquote><p style=\"text-align: left;\"><strong>我们的云业务在任何给定时刻，都在为了支持前沿AI开发而放弃短期收入，因为我们在做算力上的取舍。所以你说的那种情况，已经是现实了。</strong></p></blockquote><p style=\"text-align: left;\">这句话的分量在于：它不是一个未来的战略承诺，而是对当下现状的描述。谷歌云在实际运营中，已经在系统性地压缩对外出租算力的规模，以保证DeepMind的研发需求。</p><p style=\"text-align: left;\">但皮查伊也补充了一个反直觉的判断——适度的资源约束对研究本身未必是坏事：</p><blockquote><p style=\"text-align: left;\">就算是我们的AI研究员，一定程度的约束其实也是好事，那是激发创造力的来源，那是促使你想出更高效的算法突破、比如用更少内存的来源。</p></blockquote><p style=\"text-align: left;\">这段话背后是一个更深的逻辑：无限算力不等于无限进步，资源约束有时会倒逼出更根本性的架构创新。</p><h2 style=\"text-align: left;\">算力危机：2027年的配额，“极其困难”</h2><p style=\"text-align: left;\">访谈中，主持人Kantrowitz提到了一个在硅谷风投圈广泛流传的信息——知名风投a16z联合创始人Ben Horowitz在播客中转述亚马逊CEO安迪·贾西（Andy Jassy）的话，称云算力已经&#34;卖空到2028年&#34;。皮查伊的回应，是迄今为止对这一说法最直接的CEO级别确认：</p><blockquote><p style=\"text-align: left;\">我觉得这个说法是合理的……如果今天有客户来找我们要2027年的算力，那是极其困难的事情。</p></blockquote><p style=\"text-align: left;\">为什么会到这一步？皮查伊给出了一个清晰的结构性解释：数据中心从决策到投产有2至3年的刚性滞后，而智能体任务的爆发完全超出了两年前的需求预测。</p><blockquote><p style=\"text-align: left;\"><strong>今天回头看，每家公司都是短缺的，不只是我们，其他公司也一样。大家都希望自己手里的算力是当初规划量的2倍。</strong></p></blockquote><p style=\"text-align: left;\">皮查伊还补充了一个关键的时间锚点：</p><blockquote><p style=\"text-align: left;\">我们今天手里有的算力，是我们在2024年做的决策的产物。就算回到去年今天，智能体任务的爆发还没有发生。</p></blockquote><p style=\"text-align: left;\">这句话揭示了一个系统性问题：AI需求的增速已经超过了任何基础设施规划模型的预测能力。两年前做出的决策，在今天看来几乎必然是低估的。</p><p style=\"text-align: left;\"><strong>对于“是否存在过度建设（overbuilding）”的追问，</strong>皮查伊的回答是区分时间框架：</p><blockquote><p style=\"text-align: left;\">如果我拉一个足够长的时间框架，这个量级的建设对我来说完全合理。我们处于有史以来最重要的技术的早期阶段，这不像光纤——光纤铺完是暗的，没人用。我们是在每天都面临产能约束的情况下使用它。</p></blockquote><p style=\"text-align: left;\"><strong>他没有否认短期内存在局部过剩的可能——电力、内存、电工等2000多个配套要素中任何一个出现瓶颈，都可能导致局部积压。但他对长期需求的判断是确定性的：</strong></p><blockquote><p style=\"text-align: left;\">我有没有信心，我们规划的算力最终会有用武之地？给足时间的话，绝对有。</p></blockquote><p class=\"t-img-caption\"><img src=\"https://community-static.tradeup.com/news/528e2235b4d84188ff7d325c5e68e13b\" data-align=\"center\" tg-width=\"1024\" tg-height=\"565\"/><span></span></p><h2 style=\"text-align: left;\">RSI：不是智能爆炸，是“智能体和研究员一起干活”</h2><p style=\"text-align: left;\">递归自我改进（RSI）是这次访谈中技术密度最高的议题，也是当前AI安全讨论中争议最大的话题之一。</p><p style=\"text-align: left;\">背景是：谷歌联合创始人谢尔盖·布林此前在一次DeepMind内部研究员圆桌上提到，Argon的核心创新之一是“更多地依赖模型来帮助开发模型”。这句话随即在AI安全社区引发了关于“智能爆炸”的讨论。</p><p style=\"text-align: left;\">主持人Kantrowitz在访谈中直接抛出了这个问题，并提到了近期几起AI系统越权行为的案例。皮查伊的回应把RSI的当前状态拆解成了三个层次：</p><p style=\"text-align: left;\"><strong>第一层：定义当前位置。</strong> 他明确指出，当前的RSI远没有到达科幻叙事中&#34;自主失控&#34;的终局状态：</p><blockquote><p style=\"text-align: left;\"><strong>在一个连续谱上，我们距离RSI的终态还相当遥远。今天正确的描述方式是：你有RSI循环，同时有AI研究员在模型开发的各个环节共同参与。我会说，你几乎是有智能体在某种自动化方式下，帮助AI研究员协同工作来构建这些模型。</strong></p></blockquote><p style=\"text-align: left;\"><strong>第二层：安全机制。</strong> 他把安全框架描述为&#34;纵深防御&#34;——对齐研究是一层，但不是唯一的一层。在此之上还有实时遥测监控、沙箱隔离架构、以及给企业客户的部署控制权限。</p><p style=\"text-align: left;\"><strong>第三层：行业标准化。</strong> 皮查伊把下一步的关键指向了跨行业标准建设：</p><blockquote><p style=\"text-align: left;\">就像金融控制领域有会计准则、有独立审计师一样，这些都是我们作为一个行业需要努力推进的下一步，包括使用独立评估机构进行评估的能力。</p></blockquote><p style=\"text-align: left;\">这段话的信息量在于：它承认了当前的对齐工作存在&#34;明显缺口&#34;，但将解决路径指向了外部标准化机制，而非单纯依赖各家公司的内部自律。</p><h2 style=\"text-align: left;\">太空数据中心与长期下注</h2><p style=\"text-align: left;\">访谈的最后部分，皮查伊谈到了谷歌的几个长期战略赌注，其中最具想象空间的是太空数据中心项目。</p><p style=\"text-align: left;\">谷歌已将TPU试验芯片送入太阳同步轨道，探索在太空中利用24小时不间断<a href=\"https://laohu8.com/S/000591\">太阳能</a>运行数据中心。在地面数据中心面临社区反弹、电网并网排队长达数年的背景下，这一方向的战略逻辑逐渐清晰。</p><p style=\"text-align: left;\">皮查伊用Waymo的起点来类比这个项目当前的阶段：</p><blockquote><p style=\"text-align: left;\">这些技术让你想起那些时刻——你站在那里，它看起来有点像科幻，但你知道，10到20年之后，它实际上会以一种有意义的方式运转。</p></blockquote><p style=\"text-align: left;\">他透露，在访谈录制期间，团队正在等待太空中TPU的&#34;<a href=\"https://laohu8.com/S/MOMO\">Hello</a> World&#34;时刻——即首次成功运行的信号。</p><p style=\"text-align: left;\">这个项目目前处于&#34;两到三年&#34;的早期阶段，而谷歌的量子计算项目已经走了将近20年。皮查伊对这类长周期赌注的态度，折射出谷歌在基础设施层面的时间观：</p><blockquote><p style=\"text-align: left;\">在太阳同步轨道上，以24小时不间断的方式获取我们能接触到的最丰富的能源形式，对于我们将要面对的算力指数级增长来说，这将非常有意义。</p></blockquote></body></html>\n\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/28b6d1ac06e9a3434d66e71f0a5e6189","relate_stocks":{"LU2097829019.USD":"AZ EQUITY - BORLETTI GLOBAL LIFESTYLE \"AI\" (USD) ACC","LU2077746001.SGD":"Blackrock ESG Multi-Asset A2 SGD-H","LU1153585028.USD":"BGF GLOBAL LONG-HORIZON EQUITY  \"A4\" (USD) INC","SG9999014567.USD":"UOB UNITED INCOME FOCUS TRUST FUND (USD) ACC","LU2236285917.USD":"ALLIANZ GLOBAL INCOME \"AMG\" (USD) INC","LU1496350502.SGD":"FRANKLIN DIVERSIFIED DYNAMIC \"A\" (SGDHDG) ACC","LU1232071149.USD":"AZ FUND 1 GLOBAL GROWTH SELECTOR \"AAZ\" (USDHDG) ACC","LU0210533765.USD":"JPM GLOBAL GROWTH \"A\" (USD) ACC","IE00BJLML261.HKD":"HSBC GLOBAL EQUITY INDEX \"HCH\" (HKD) ACC","LU0061475181.USD":"THREADNEEDLE (LUX) AMERICAN \"AU\" (USD) ACC","SG9999015358.SGD":"United Income Focus Trust Dis SGD-H","LU0316494557.USD":"FRANKLIN GLOBAL FUNDAMENTAL STRATEGIES \"A\" ACC","LU1301847155.USD":"BGF GLOBAL MULTI-ASSET INCOME FUND \"A4G\" (USD) INC","LU0708995401.HKD":"FRANKLIN U.S. OPPORTUNITIES \"A\" (HKD) ACC","LU0820561818.USD":"安联收益及增长平衡基金Cl AM DIS","LU0094547139.USD":"abrdn SICAV I - GLOBAL SUSTAINABLE EQUITY FUND \"A\" (USD) ACC","IE00B19Z3581.USD":"Legg Mason ClearBridge - Value A Acc USD","LU1267930813.SGD":"FRANKLIN TEMPLETON SHARIAH GLOBAL EQUITY \"AS\" (SGD) ACC","LU0726765562.USD":"JPM GLOBAL INCOME \"A\" (USD HEDGED) INC","BK4077":"互动媒体与服务","LU2125154778.USD":"ALLSPRING GLOBAL EQUITY ENHANCED INCOME \"A\" (USD) INC","LU2381873111.SGD":"BGF SYSTEMATIC GLOBAL EQUITY HIGH INCOME \"A6\" (SGDHDG) INC","LU0289739343.SGD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"A\" (SGD) ACC","LU0130517989.USD":"HARRIS ASSOCIATES US VALUE EQUITY \"R\" INC","LU0130102774.USD":"Natixis Harris Associates US Equity RA USD","LU2347655073.USD":"JPM GLOBAL INCOME \"A\" (USDHDG) INC A","IE00BFMHRM44.USD":"NEUBERGER BERMAN GLOBAL EQUITY MEGATRENDS \"A\" (USD) ACC","LU2404859741.USD":"FUNDSMITH EQUITY FUND \"R\" (USD) INC","LU0345769128.USD":"NINETY ONE GSF GLOBAL EQUITY \"A\" (USD) ACC","LU0128525929.USD":"TEMPLETON GLOBAL \"A\" (USD) ACC","LU0130518102.USD":"HARRIS ASSOCIATES GLOBAL EQUITY \"R\" INC","LU0070217475.USD":"JPM GLOBAL SELECT EQUITY \"A\" ACC","LU1880398554.USD":"AMUNDI FUNDS GLOBAL EQUITY \"A2\" (USD) INC","LU0477156797.USD":"HARRIS ASSOCIATES GLOBAL EQUITY \"RE\" (USD) ACC","LU0957808578.USD":"THREADNEEDLE (LUX) GLOBAL TECHNOLOGY \"ZU\" (USD) ACC","IE00BMPRXQ63.HKD":"NEUBERGER BERMAN NEXT GENERATION CONNECTIVITY FUND \"A\" (HKDHDG) ACC","LU2404859667.USD":"FUNDSMITH EQUITY FUND \"R\" (USD) ACC","LU0345769631.USD":"NINETY ONE GSF GLOBAL EQUITY \"A\" (USD) INC","LU1951198990.SGD":"Natixis Thematics AI & Robotics Fund H-R/A SGD-H","IE00B19Z9505.USD":"美盛-美国大盘成长股A Acc","LU0690374961.EUR":"FUNDSMITH EQUITY \"R\" (EUR) INC","LU1244550577.SGD":"FTIF - Franklin Global Multi-Asset Income A (Mdis) SGD-H1","LU0477156953.USD":"HARRIS ASSOCIATES US VALUE EQUITY \"RE\" (USD) ACC","LU1670756490.USD":"ALLIANZ EUROPE INCOME AND GROWTH \"AT\" (USDHDG) ACC","LU0690374615.EUR":"FUNDSMITH EQUITY \"R\" (EUR) ACC","LU1951200564.SGD":"Natixis Thematics AI & Robotics Fund R/A SGD","LU0211326839.USD":"TEMPLETON GLOBAL INCOME \"A\" (USD) INC","IE00BWXC8680.SGD":"PINEBRIDGE US LARGE CAP RESEARCH ENHANCED \"A5\" (SGD) ACC","LU0345770993.USD":"NINETY ONE GSF GLOBAL STRATEGIC EQUITY \"A\" (USD) INC","LU0006306889.USD":"SCHRODER ISF US LARGE CAP \"A\" (USD) INC AV","BK4550":"红杉资本持仓","LU1923623000.USD":"Natixis Thematics AI & Robotics Fund R/A USD","LU0061474960.USD":"天利环球焦点基金AU Acc","LU0683600562.USD":"AB SELECT US EQUITY \"A\" (USD) ACC","IE00BLSP4239.USD":"Legg Mason ClearBridge - Tactical Dividend Income A Mdis USD Plus","IE00BMPRXN33.USD":"NEUBERGER BERMAN 5G CONNECTIVITY \"A\" (USD) ACC","LU2125909247.SGD":"Natixis Thematics Meta H-R/A SGD","LU0672654240.SGD":"FTIF - Franklin US Opportunities A Acc SGD-H1","LU0494093205.USD":"贝莱德ESG灵活多元资产A2 USD-H","LU1429558221.USD":"Natixis Loomis Sayles US Growth Equity RA USD","SG9999015952.SGD":"LIONGLOBAL DISRUPTIVE INNOVATION \"I\" (SGD) ACC","LU0426412945.USD":"NINETY ONE GSF GLOBAL FRANCHISE \"A\" (USD) ACC","LU2756315664.SGD":"ALLIANZ INCOME AND GROWTH \"AMI\" (SGDHDG) INC","LU2264538146.SGD":"Fullerton Lux Funds - Global Absolute Alpha A Acc SGD","LU1435385759.SGD":"Natixis Loomis Sayles US Growth Equity RA SGD-H","IE00BN8TJ469.HKD":"FTGF CLEARBRIDGE TACTICAL DIVIDEND INCOME \"A\" (HKD) INC","LU1935043023.USD":"MANULIFE GF GLOBAL MULTI-ASSET DIVERSIFIED INCOME \"AA\" (USD) INC A","IE00BBT3K403.USD":"LEGG MASON CLEARBRIDGE TACTICAL DIVIDEND INCOME \"A(USD) ACC","LU2746668461.USD":"MANULIFE DYNAMIC LEADERS \"AA\" (USD) ACC","LU2430703095.HKD":"WELLINGTON MULTI-ASSET HIGH INCOME \"AM4\" (HKD) INC","LU2430703178.SGD":"WELLINGTON MULTI-ASSET HIGH INCOME \"AM4H\" (SGDHDG) INC","LU2102340556.USD":"AZ ALLOCATION - RISK PARITY FACTOR \"A\" (USD) ACC","LU0211327993.USD":"TEMPLETON GLOBAL EQUITY INCOME \"A\" (USD) ACC","LU0345768740.USD":"NINETY ONE GSF GLOBAL STRATEGIC MANAGED \"A\" (USD) INC","LU0795875169.SGD":"JPMorgan Investment Funds - Global Income A (div) SGD-H","LU2430703251.USD":"WELLINGTON MULTI-ASSET HIGH INCOME \"AM4\" (USD) INC","LU1551013425.SGD":"Allianz Income and Growth Cl AMg2 DIS H2-SGD","LU2756315318.SGD":"ALLIANZ INCOME AND GROWTH \"AMG\" (SGDHDG) INC A","SG9999017495.SGD":"UGDP UNITED GLOBAL QUALITY GROWTH \"B\" (SGD) ACC","IE00BLSP4452.SGD":"Legg Mason ClearBridge - Tactical Dividend Income A Mdis SGD-H Plus","LU2746668974.SGD":"MANULIFE DYNAMIC LEADERS \"AA\" (SGDHDG) ACC","SG9999014914.USD":"UNITED GLOBAL QUALITY GROWTH (USDHDG) INC","LU2362540622.SGD":"WELLINGTON NEXT GENERATION GLOBAL EQUITY \"A\" (SGDHDG) ACC","LU0130103400.USD":"Natixis Harris Associates Global Equity RA USD","LU0417517546.SGD":"Allianz US Equity Cl AT Acc SGD","SG9999015341.SGD":"United Income Focus Trust Acc SGD-H","IE0004091025.USD":"BNY MELLON GLOBAL OPPORTUNITIES \"B\" (USD) ACC","LU2242649171.HKD":"FIDELITY FUNDS GLOBAL THEMATIC OPPORTUNITIES \"A\" (HKD) ACC","LU2360032135.SGD":"ALLSPRING GLOBAL EQUITY ENHANCED INCOME \"A\" (SGDHDG) INC","LU0472753341.HKD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"A\" (HKD) ACC","LU1201861165.SGD":"Natixis Harris Associates Global Equity PA SGD","BK4527":"明星科技股","LU2462157665.USD":"ALLIANZ GLOBAL INCOME \"A\" (USD) INC","LU1201861249.SGD":"Natixis Harris Associates US Equity PA SGD-H","SG9999014880.SGD":"大华全球优质成长基金Acc SGD","LU0345774391.USD":"NINETY ONE GSF AMERICAN FRANCHISE \"A\" (USD) ACC","LU0980610538.SGD":"Natixis Harris Associates US Equity RA SGD-H","LU1221951046.USD":"NORDEA 1 STABLE RETURN \"HM\" (USDHDG) INC","LU0868494617.USD":"UBS (LUX) EQUITY SICAV - US TOTAL YIELD SUSTAINABLE \"P\" (USD) ACC","LU1880398471.USD":"AMUNDI FUNDS GLOBAL EQUITY \"A2\" (USD) ACC","SGXZ31699556.SGD":"UGDP UNITED GLOBAL QUALITY GROWTH \"C\" (SGDHDG) ACC","LU2125154935.USD":"ALLSPRING (LUX) WF GLOBAL EQUITY ENHANCED INCOME \"I\" (USD) INC","LU1221951129.SGD":"NORDEA 1 STABLE RETURN \"HM\" (SGDHDG) INC","LU0310800965.SGD":"FTIF - Templeton Global Balanced A Acc SGD","LU1670628061.USD":"M&G (LUX) NORTH AMERICAN DIVIDEND \"A\" (USD) INC","LU0124676726.USD":"AB SICAV I - SUSTAINABLE US THEMATIC PORTFOLIO \"A\" (USD) ACC","LU1366333091.USD":"FIDELITY GLOBAL FOCUS \"A\" (USD) ACC","LU0496365809.HKD":"TEMPLETON GLOBAL INCOME \"A\" (HKD) INC (Q)","LU2362541273.HKD":"WELLINGTON NEXT GENERATION GLOBAL EQUITY \"A\" (HKD) ACC","LU1196500208.SGD":"NORDEA STABLE RETURN \"HB\" (SGDHDG) ACC","LU1066053197.SGD":"HSBC GIF GLOBAL EQUITY VOLATILITY FOCUSED \"AM3\" (SGDHDG) INC","LU2362541513.USD":"WELLINGTON NEXT GENERATION GLOBAL EQUITY \"A\" (USD) ACC","LU2491049909.HKD":"WELLINGTON SUSTAINABLE OUTCOMES \"A\" (HKD) ACC","LU1059921491.USD":"NORDEA 1 GLOBAL STABLE EQUITY \"HB\" (USDHDG) ACC","LU2065170008.USD":"M&G (LUX) GLOBAL MAXIMA \"A\" (USD) INC","LU2028103732.USD":"ALLIANZ GLOBAL SUSTAINABILITY \"AMG\" (USD) INC","LU0289961442.SGD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"AX\" (SGD) ACC","LU0310799852.SGD":"FTIF - Templeton Global Equity Income A MDIS SGD","LU2491050071.SGD":"WELLINGTON SUSTAINABLE OUTCOMES \"A\" (SGDHDG) ACC","LU0345774631.USD":"NINETY ONE GSF AMERICAN FRANCHISE \"A\" (USD) INC","LU0985320562.USD":"NORDEA 1 GLOBAL STARS EQUITY \"BP\" (USD) ACC","LU1366192091.USD":"ALLIANZ US EQUITY PLUS \"AM\" (USD) INC","LU2491050154.USD":"WELLINGTON SUSTAINABLE OUTCOMES \"A\" (USD) ACC","LU0342679015.USD":"ALLIANZ GLOBAL EQUITY UNCONSTRAINED \"AT\" (USD) ACC","BK4532":"文艺复兴科技持仓","LU0011850046.USD":"贝莱德全球长线股票 A2 USD","LU0097036916.USD":"贝莱德美国增长A2 USD","LU0466842654.USD":"HSBC ISLAMIC GLOBAL EQUITY INDEX \"A\" (USD) ACC","IE00B1XK9C88.USD":"PINEBRIDGE US LARGE CAP RESEARCH ENHANCED \"A\" (USD) ACC","IE0001KFT4U8.USD":"FTGF CLEARBRIDGE GLOBAL GROWTH LEADERS \"A\" (USD) INC","LU0742534661.SGD":"Fidelity America A-SGD (hedged)","LU0048584097.USD":"FIDELITY FUNDS  GLOBAL THEMATIC OPPORTUNITIES \"A\" (USD) INC","BK4534":"瑞士信贷持仓","LU0138981039.USD":"FIDELITY FUNDS SUSTAINABLE MULTI ASSET INCOME \"A\" INC","LU2860962120.EUR":"CPR INVEST - ARTIFICIAL INTELLIGENCE \"A2\" (EUR) ACC A","BK4533":"AQR资本管理(全球第二大对冲基金)","LU0203347892.USD":"SCHRODER ISF QEP GLOBAL ACTIVE VALLUE \"A\" (USD)  INC AV","LU1032955483.USD":"NORDEA 1 STABLE RETURN \"HB\" (USDHDG) ACC","LU2592432038.USD":"WELLINGTON MULTI-ASSET HIGH INCOME \"A\" (USD) ACC","IE0034235188.USD":"PINEBRIDGE GLOBAL FOCUS EQUITY \"A\" (USD) ACC","LU1988902786.USD":"FULLERTON LUX FUNDS GLOBAL ABSOLUTE ALPHA \"I\" (USD) ACC","IE00BDCRKT87.USD":"PINEBRIDGE GLOBAL DYNAMIC ASSET ALLOCATION \"ADC\" (USD) INC","LU0302445910.USD":"SCHRODER ISF GLOBAL CLIMATE CHANGE EQUITY \"A\" ACC","LU2602419157.SGD":"HSBC ISLAMIC GLOBAL EQUITY INDEX \"AC\" (SGD) ACC","LU1642822529.SGD":"THREADNEEDLE (LUX) GLOBAL TECHNOLOGY \"A\" (SGD) ACC","BK4538":"云计算","LU1145028129.USD":"ALLIANZ INCOME AND GROWTH \"AQ\" (USD) INC","LU1496350171.SGD":"FRANKLIN DIVERSIFIED BALANCED \"A\" (SGDHDG) ACC","LU2505996681.GBP":"WELLINGTON MULTI-ASSET HIGH INCOME \"AM4H\" (GBPHDG) INC","SGXZ81514606.USD":"大华环球创新基金A Acc USD","LU1823568750.SGD":"Fidelity Global Technology A-ACC SGD","LU2420271590.USD":"ALLIANZ SELECT INCOME AND GROWTH \"AT\" (USD) ACC","LU2505996509.AUD":"WELLINGTON MULTI-ASSET HIGH INCOME \"AM4H\" (AUDHDG) INC","LU2065171311.SGD":"M&G (LUX) GLOBAL MAXIMA \"A\" (SGD) ACC","LU0157215616.USD":"FIDELITY GLOBAL FOCUS \"A\" INC","LU1914381329.SGD":"Allianz Best Styles Global Equity Cl ET Acc H2-SGD","LU1839511570.USD":"WELLS FARGO GLOBAL FACTOR ENHANCED EQUITY \"I\" (USD) ACC","LU2065171402.SGD":"M&G (LUX) GLOBAL MAXIMA \"A\" (SGD) INC","LU0795875086.SGD":"JPMorgan Investment Funds - Global Income A (div) SGD","LU0823421333.USD":"BNP PARIBAS DISRUPTIVE TECHNOLOGY \"C\" (USD) ACC","LU0823421416.USD":"BNP PARIBAS DISRUPTIVE TECHNOLOGY \"C\" (USD) INC","LU0943347566.SGD":"安联收益及增长平衡基金AM H2-SGD","LU0820562030.AUD":"ALLIANZ INCOME AND GROWTH \"AMH2\" (AUDHDG) H2 INC","LU0187121727.USD":"FIDELITY SUSTAINABLE US EQUITY \"A\" (USD) INC","LU0048573561.USD":"FIDELITY AMERICA \"A\" (USD) INC","SG9999015945.SGD":"LionGlobal Disruptive Innovation Fund A SGD","HK0000914686.HKD":"GFI Global Select Equity Fund Class A (HKD)","LU0354030511.USD":"ALLSPRING  U.S. LARGE CAP GROWTH \"I\" (USD) ACC","LU0354030438.USD":"富国美国大盘成长基金Cl A Acc","HK0000914660.USD":"GFI Global Select Equity Fund Class A (USD)","LU0175139822.USD":"AB FCP I Global Equity Blend A USD","LU1084165304.USD":"FIDELITY WORLD \"A\" (USD) ACC","SG9999014898.SGD":"United Global Quality Growth Fund Dis SGD","LU1720051108.HKD":"ALLIANZ GLOBAL ARTIFICIAL INTELLIGENCE \"AT\" (HKD) ACC","LU0444971666.USD":"天利全球科技基金","LU0787776722.HKD":"AB SELECT US EQUITY PORTFOLIO \"A\" (HKD) ACC","LU2275660780.HKD":"SCHRODER ISF GLOBAL CLIMATE CHANGE EQUITY \"A\" (HKD) ACC","SG9999018865.SGD":"United Global Quality Growth Fd Cl Dist SGD-H","LU0912757837.SGD":"JPMorgan Investment Funds - Global Income A (mth) SGD-H","LU1267930490.SGD":"TEMPLETON GLOBAL EQUITY INCOME \"AS\" (SGD) INC A","LU1670627923.USD":"M&G (LUX) NORTH AMERICAN DIVIDEND \"A\" (USD) ACC","SGXZ99366536.SGD":"United Global Innovation A Acc SGD-H","LU0069063385.USD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"A\" (USD) ACC","LU2106854487.HKD":"ALLIANZ THEMATICA \"AMG\" (HKD) INC","LU0312595415.SGD":"Schroder ISF Global Climate Change Equity A Acc SGD","LU1400636657.SGD":"Allianz Europe Income and Growth Cl AMg DIS H2-SGD","LU2065169927.USD":"M&G (LUX) GLOBAL MAXIMA \"A\" (USD) ACC","LU0689626769.HKD":"AB SICAV I - SUSTAINABLE US THEMATIC PORTFOLIO \"A\" (HKD) ACC","SGXZ51526630.SGD":"大华环球创新基金A Acc SGD","LU1629891620.HKD":"ALLIANZ INCOME AND GROWTH \"AMG2\" (H2-HKD) INC","LU0106261372.USD":"SCHRODER ISF US LARGE CAP \"A\" ACC","BK4503":"景林资产持仓","IE00BSNM7G36.USD":"NEUBERGER BERMAN SYSTEMATIC GLOBAL SUSTAINABLE VALUE \"A\" (USD) ACC","IE00BKPKM429.USD":"NEUBERGER BERMAN GLOBAL SUSTAINABLE EQUITY \"A\" (USD) ACC","IE00BZ1G4Q59.USD":"LEGG MASON CLEARBRIDGE US EQUITY SUSTAINABILITY LEADER \"A\"(USD) INC (A)","LU0251142724.SGD":"Fidelity America A-SGD","LU0077335932.USD":"FIDELITY AMERICAN GROWTH \"A\" INC","SG9999015986.USD":"LIONGLOBAL DISRUPTIVE INNOVATION \"I\" (USD) ACC","LU2023250330.USD":"ALLIANZ INCOME AND GROWTH \"AMG\" (USD) INC","LU1720051017.SGD":"Allianz Global Artificial Intelligence AT Acc H2-SGD","LU0158827781.USD":" ALLIANZ GLOBAL SUSTAINABILITY \"AT\" (USD) ACC","LU1244550221.USD":"FRANKLIN GLOBAL MULTI-ASSET  INCOME \"A\" (USDHEDGED) INC (M)","LU0149725797.USD":"汇丰美国股市经济规模基金","SG9999014542.SGD":"United Income Focus Trust Acc SGD","LU0057025933.USD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"AX\" (USD) ACC","LU0882574139.USD":"富达环球消费行业基金A ACC","IE00BQXX3D17.EUR":"GUINNESS GLOBAL INNOVATORS \"C\" (EUR) ACC","LU0070302665.USD":"FRANKLIN MUTUAL U.S. VALUE \"A\" (USD) ACC","LU1852331112.SGD":"Blackrock World Technology Fund A2 SGD-H","IE00BQXX3C00.GBP":"GUINNESS GLOBAL INNOVATORS \"C\" (GBP) ACC","LU0276348264.USD":"THREADNEEDLE (LUX) GLOBAL DYNAMIC REAL RETURN\"AUP\" (USD) INC","IE00BQXX3F31.USD":"GUINNESS GLOBAL INNOVATORS \"C\" (USD) ACC","LU0792757196.USD":"TEMPLETON SHARIAH GLOBAL EQUITY FUND \"A\" (USD) ACC","LU2244417387.USD":"FIDELITY SUSTAINABLE US EQUITY \"A\" (USD) ACC","LU1989772840.SGD":"CPR Invest - Climate Action A2 Acc SGD-H","LU0784383803.USD":"BGF GLOBAL MULTI-ASSET INCOME FUND \"A\" (USD) INC A","LU1894683348.USD":"AMUNDI FUNDS US EQUITY RESEARCH VALUE \"A2\" (USD) INC","LU1989772923.USD":"CPR Invest - Climate Action A2 Acc USD-H","LU2023250504.SGD":"Allianz Thematica Cl AMg DIS H2-SGD","LU1989764664.SGD":"CPR Invest - Global Disruptive Opportunities A2 Acc SGD-H","LU1121112475.USD":"NINETY ONE GSF GLOBAL FRANCHISE \"A\" (USDHDG) ACC","SG9999000418.SGD":"Aberdeen Standard Global Technology SGD","LU1989764748.USD":"东方汇理环球颠覆性机遇A2 Acc","LU1119994496.HKD":"FIDELITY WORLD \"A\" (HKD) ACC","IE0034235303.USD":"PINEBRIDGE US RESEARCH ENHANCED CORE EQUITY \"A\" (USD) ACC","LU0256863902.USD":"ALLIANZ US EQUITY \"AT\" (USD) ACC","GB00B4QBRK32.GBP":"FUNDSMITH EQUITY \"R\" (GBP) INC","LU0868494708.USD":"UBS (LUX) EQUITY SICAV - US TOTAL YIELD SUSTAINABLE \"P\" (USD) INC","GB00B4LPDJ14.GBP":"FUNDSMITH EQUITY \"R\" (GBP) ACC","LU1043141396.HKD":"NINETY ONE GSF GLOBAL FRANCHISE \"A\" (HKD) ACC","LU0215105999.USD":"SCHRODER ISF GLOBAL EQUITY \"A\" ACC","BK4514":"搜索引擎","LU1267930227.SGD":"TEMPLETON GLOBAL BALANCED \"AS\" (SGD) ACC A","IE00BKDWB100.SGD":"PINEBRIDGE US LARGE CAP RESEARCH ENHANCED \"A5H\" (SGDHDG) ACC","IE00B1BXHZ80.USD":"Legg Mason ClearBridge - US Appreciation A Acc USD","LU0225283273.USD":"SCHRODER ISF GLOBAL EQUITY ALPHA \"A\" (USD) ACC","LU1261432733.SGD":"Fidelity World A-ACC-SGD","LU2092937148.SGD":"Blackrock ESG Multi-Asset A8 SGD-H","LU1043141123.HKD":"NINETY ONE GSF GLOBAL FRANCHISE \"A\" (HKD) INC 2","LU1064131342.USD":"Fullerton Lux Funds - Global Absolute Alpha A Acc USD","LU2111349929.HKD":"ALLIANZ GLOBAL SUSTAINABILITY \"AM\" (HKD) INC","LU2092627202.USD":"Blackrock ESG Multi-Asset A8 USD-H","LU0203345920.USD":"SCHRODER ISF QEP GLB ACT. VL \"A\" (USD) ACC","LU0211331839.USD":"FRANKLIN MUTUAL GLB DISCOVERY \"A\" (USD) ACC","LU0994945656.USD":"NINETY ONE GSF GLOBAL FRANCHISE \"A\" (USD) INC 2","LU0310800379.SGD":"FTIF - Templeton Global A Acc SGD","LU1674673691.USD":"HSBC GIF GLOBAL LOWER CARBON EQUITY \"AD\" (USD) INC","LU1674673428.USD":"HSBC GIF GLOBAL LOWER CARBON EQUITY \"AC\" (USD) ACC","LU0234570918.USD":"高盛全球核心股票组合Acc Close","LU0942090050.USD":"UBS (LUX) EQUITY SICAV - US TOTAL YIELD SUSTAINABLE \"P\" (USD)  INC","LU2456880835.USD":"ALLIANZ GLOBAL INCOME \"AT\" (USD) ACC","LU1718418525.SGD":"JPMorgan Investment Funds - Global Select Equity A (acc) SGD","LU0154236417.USD":"BGF US FLEXIBLE EQUITY \"A2\" ACC","LU1235294995.USD":"FIDELITY GLOBAL TECHNOLOGY \"A\" (USDHDG) ACC","LU1551013342.USD":"Allianz Income and Growth Cl AMg2 DIS USD","LU0080751232.USD":"富达环球多元动力基金A","LU1303367103.USD":"摩根大通多经理另类基金 A (acc)","LU2098885051.SGD":"JPMorgan Funds - Multi-Manager Alternatives A (acc) SGD","LU0234572021.USD":"高盛美国核心股票组合Acc","IE00BKVL7J92.USD":"Legg Mason ClearBridge - US Equity Sustainability Leaders A Acc USD","LU1868837300.USD":"CT (LUX) I AMERICAN FUND \"9\" (USD) ACC","LU2087625088.SGD":"ALLSPRING  US ALL CAP GROWTH \"A\" (SGDHDG) ACC","LU1267930573.SGD":"TEMPLETON GLOBAL \"AA\" (SGD) ACC A","LU1868836914.USD":"CT (LUX) I AMERICAN \"3\" (USD) ACC","IE000F7OJYX1.SGD":"RQI GLOBAL \"I\" (SGDHDG) INC","BK4598":"佩洛西持仓","LU0170899867.USD":"EASTSPRING INVESTMENTS WORLD VALUE EQUITY  \"A\" (USD) ACC","IE0000B7K699.USD":"RQI GLOBAL \"I\" (USD) INC","LU1868837136.USD":"CT (LUX) I AMERICAN \"8\" (USD) ACC","LU1548497426.USD":"安联环球人工智能AT Acc","LU1803068623.USD":"TEMPLETON GLOBAL CLIMATE CHANGE \"AH1\" (USDHDG) ACC","LU1868836591.USD":"CT (LUX) I AMERICAN \"1U\"(USD) ACC","IE000BZT7R73.USD":"RQI GLOBAL \"I\" (USD) ACC","IE00BHPRN162.USD":"BNY MELLON BLOCKCHAIN INNOVATION \"B\" (USD) ACC","IE000MB6IJN5.SGD":"RQI GLOBAL \"I\" (SGD) ACC","SG9999014906.USD":"大华全球优质成长基金Acc USD","LU1868836757.USD":"CT (LUX) I AMERICAN FUND \"2\" (USD) ACC","LU2347655156.SGD":"JPMorgan Investment Funds - Global Income A (icdiv) SGD-H","LU0957791311.USD":"THREADNEEDLE (LUX) GLOBAL FOCUS \"ZU\" (USD) ACC","LU0314104364.USD":"MANULIFE GF AMERICAN GROWTH \"AA\" (USD) INC","LU1815333072.USD":"THREADNEEDLE (LUX) GLOBAL FOCUS \"AUP\" (USD) INC","LU0395794307.EUR":"JPM GLOBAL INCOME (DIV) \"A\" (EUR) INC","LU1316542783.SGD":"Janus Henderson Horizon Global Technology Leaders A2 SGD","BK4602":"量子计算概念","LU0314106906.USD":"MANULIFE GF GLOBAL EQUITY \"AA\" (USD) INC","LU0642271901.SGD":"Janus Henderson Horizon Global Technology Leaders A2 SGD-H","LU0320765489.SGD":"FTIF - Franklin Mutual US Value A Acc SGD","LU2360106780.USD":"BGF WORLD TECHNOLOGY \"A4\" (USD) INC","LU1430594728.SGD":"Eastspring Investments - Global Low Volatility Equity AS SGD","LU0949170772.SGD":"Blackrock Global Equity Income A6 SGD-H","IE000WAW04H5.SGD":"RQI GLOBAL \"I\" (SGD) INC A","SG9999002232.USD":"Allianz Global High Payout USD","LU1003077747.HKD":"BGF GLOBAL EQUITY INCOME \"A6\" (HKDHDG) INC","IE000SO60NU8.SGD":"RQI GLOBAL \"I\" (SGDHDG) ACC A","LU0265550946.USD":"BGF SYSTEMATIC GLOBAL ENHANCED EQUITY YIELD \"A5\" (USD) INC","SG9999002224.SGD":"Allianz Global High Payout SGD","LU0738911758.USD":"Blackrock Global Equity Income A6 USD","LU1400636491.USD":"Allianz Europe Income and Growth Cl AMg DIS H2-USD","LU1066051811.HKD":"HSBC GIF GLOBAL EQUITY VOLATILITY FOCUSED \"AM2\" (HKD) INC","LU0949170426.SGD":"Blackrock Global Multi-Asset Income A6 SGD-H","LU1244550494.USD":"FRANKLIN GLOBAL MULTI-ASSET INCOME \"A\" (USDHEDGED) ACC","LU0061474705.USD":"THREADNEEDLE (LUX) GLOBAL DYNAMIC REAL RETURN \"AU\" (USD) ACC","LU0648000940.SGD":"Natixis Harris Associates Global Equity RA SGD","LU0107464264.USD":"abrdn SICAV I - GLOBAL INNOVATION EQUITY \"A\" (USD) ACC","LU2087621335.USD":"ALLSPRING GLOBAL FACTOR ENHANCED EQUITY \"A\" (USD) ACC","BK4576":"AR","LU0784384876.USD":"Blackrock Global Multi-Asset Income A6 USD","LU1046421795.USD":"富达环球科技A-ACC","LU0648001328.SGD":"Natixis Harris Associates US Equity RA SGD","IE00BD6J9T35.USD":"NEUBERGER BERMAN NEXT GENERATION MOBILITY \"A\" (USD) ACC","IE00BYXW3230.USD":"PINEBRIDGE GLOBAL DYNAMIC ASSET ALLOCATION \"AA\" (USD) ACC","HK0000304805.HKD":"PICTET 战略收益基金 P类港元","LU2417539215.USD":"ALLIANZ GLOBAL INCOME \"AMF\" (USD) INC","LU0289960550.SGD":"AB FCP I - GLOBAL EQUITY BLEND PORTFOLIO 'A' (SGD) ACC","LU1974910355.USD":"Allianz Thematica Cl AMg DIS USD","HK0000304797.USD":"PICTET 战略收益基金 P类 美元","BK4612":"AI芯片","SGXZ80539844.SGD":"ALLIANZ GLOBAL HIGH PAYOUT \"AMI\" (H2-SGD) INC","BK4579":"人工智能","LU0528227936.USD":"富达环球人口趋势基金A-ACC","LU0889566641.SGD":"FTSF - Templeton Shariah Global Equity A Acc SGD","IE0004086264.USD":"BNY MELLON GLOBAL OPPORTUNITIES \"A\" (USD) ACC","BK4574":"无人驾驶","LU1267930730.SGD":"富兰克林美国机遇基金AS Acc SGD (CPF)","LU0109391861.USD":"富兰克林美国机遇基金A Acc","LU2357305700.SGD":"Allianz Global Artificial Intelligence ET H2-SGD","BK4573":"虚拟现实","SGXZ44715902.SGD":"MAYBANK GLOBAL SHARIAH MULTI-ASSETS-I FUND \"A\" (SGDHDG) INC A","LU0661504455.SGD":"Blackrock Global Equity Income A5 SGD-H","BK4581":"高盛持仓","LU0545039389.USD":"BGF GLOBAL EQUITY INCOME \"A2\" ACC","LU0640476718.USD":"THREADNEEDLE (LUX) US CONTRARIAN CORE EQ \"AU\" (USD) ACC","LU1280957306.USD":"THREADNEEDLE (LUX) US CONTRARIAN CORE EQUITIES \"AUP\" (USD) INC","IE00B7SZLL34.SGD":"Legg Mason ClearBridge - Value A Acc SGD-H","LU2089284900.SGD":"Allianz Global Sustainability Cl AM Dis H2-SGD","HK0000819398.USD":"BLACKROCK PREMIER SYSTEMATIC GLOBAL EQUITY HIGH INCOME FUND A10 (USD)INC","LU1202635105.USD":"ALLIANZ EUROPE INCOME AND GROWTH \"AM\" (USDHDG) INC","LU0128525689.USD":"TEMPLETON GLOBAL BALANCED \"A\"(USD) ACC","LU2237438978.USD":"Amundi Funds US Pioneer A2 (C) USD","LU1144417596.USD":"UBS (LUX) GLOBAL QUALITY DIVIDEND EQUITY \"Q\" (USD) INC","LU0444973449.USD":"CT (LUX) I GLOBAL TECHNOLOGY \"DU\" (USD) ACC","LU1585245621.USD":"EASTSPRING INV GLOBAL LOW VOLATILITY EQUITY FUND \"A\" (USD) ACC B","LU2191332357.HKD":"SCHRODER ISF SUSTAINABLE MULTI-ASSET INCOME \"A\" (HKDHDG) INC","LU0320765059.SGD":"FTIF - Franklin US Opportunities A Acc SGD","LU0158827948.USD":"ALLIANZ GLOBAL SUSTAINABILITY \"A\" (USD) INC","LU2023250843.SGD":"Allianz Thematica Cl AT Acc H2-SGD","LU0096362180.USD":"CT (LUX) I GLOBAL FOCUS \"DU\" (USD)","LU0353189763.USD":"ALLSPRING  US ALL CAP GROWTH FUND \"I\" (USD) ACC","LU0708994859.HKD":"TEMPLETON GLOBAL \"A\" (HKD) ACC","LU2471134952.CNY":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (CNYHDG) INC","IE00BJTD4N35.SGD":"Neuberger Berman US Long Short Equity A1  Acc SGD-H","LU0208291251.USD":"FRANKLIN MUTUAL U.S. VALUE \"A\" (USD) INC","LU1935043452.HKD":"宏利环球多元资产收益 \"AA\" 基金（港元）派息","LU2471134879.HKD":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (HKD) INC","LU0426417589.USD":"NINETY ONE GSF GLOBAL FRANCHISE \"A\" (USD) INC","LU0029864427.USD":"TEMPLETON GLOBAL \"A\" (USD) INC","HK0000532082.HKD":"ALLIANZ SELECTION INCOME AND GROWTH \"AM\" (HKD) INC","LU2471134523.USD":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (USD) ACC","LU0511384066.AUD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"A\" (AUDHDG) ACC","IE00B19Z8X17.USD":"FTGF CLEARBRIDGE US LARGE CAP GROWTH  \"AG\" (USD) ACC","LU2247934214.USD":"FIDELITY FUNDS SUSTAINABLE FUTURE CONNECTIVITY \"A\" (USD) ACC","LU2471134796.USD":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (USD) INC","LU0096364046.USD":"CT (LUX) I AMERICAN \"DU\" (USD) ACC","LU1935042215.USD":"MANULIFE GF GLOBAL MULTI-ASSET DIVERSIFIED INCOME  \"AA\" (USD) INC A","LU2750360641.GBP":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (GBPHDG) INC","LU0823434583.USD":"BNP PARIBAS US GROWTH \"C\" (USD) ACC","LU2023251221.USD":"ALLIANZ GLOBAL SUSTAINABILITY \"AM\" (USD) INC","LU2750360997.AUD":"INVESCO GLOBAL EQUITY INCOME ADVANTAGE \"A\" (AUDHDG) INC","LU0823434740.USD":"BNP PARIBAS US GROWTH \"C\" (USD) INC","LU0898699433.AUD":"JPM GLOBAL INCOME \"A\" (AUDHDG) INC (M)","LU1732800096.USD":"摩根大通环球收益基金A (irc)","LU0345770308.USD":"NINETY ONE GSF GLOBAL STRATEGIC EQUITY \"A\" (USD) ACC","LU0143551892.USD":"T ROWE PRICE GLOBAL FOCUSED GROWTH EQUITY \"A\" (USD) ACC","LU0211326755.USD":"TEMPLETON GLOBAL INCOME \"A\" (USD) ACC","LU2462591277.SGD":"TRP GLOBAL FOCUSED GROWTH EQUITY \"AN\" (SGDHDG) ACC","LU1571399168.USD":"ALLSPRING GLOBAL LONG/SHORT EQUITY \"IP\" (USD) ACC","LU2954938101.SGD":"T ROWE PRICE GLOBAL FOCUSED GROWTH EQUITY \"A8P\" (SGD) INC","IE00BMPRXR70.SGD":"Neuberger Berman 5G Connectivity A Acc SGD-H","LU2089283258.USD":"安联环球可持续基金Cl AM Dis","LU2999663052.SGD":"T ROWE PRICE GLOBAL FOCUSED GROWTH EQUITY \"A4P\" (SGD) INC","LU1506573853.SGD":"MANULIFE GF GLOBAL EQUITY \"AA\" (SGD) INC","LU1323610961.USD":"UBS (LUX) EQUITY SICAV - LONG TERM THEMES (USD) \"P\" (USD) ACC","LU2168564495.EUR":"AZ ALLOCATION - TREND \"AI\" (EUR) ACC","IE00BK4W5M84.HKD":"HSBC GLOBAL FUNDS ICAV US EQUITY INDEX \"HC\" (HKD) ACC","LU2168564222.USD":"AZ ALLOCATION - TREND \"AAZ\" (USDHDG) ACC","IE00BK4W5L77.USD":"HSBC GLOBAL FUNDS ICAV US EQUITY INDEX \"HC\" (USD) ACC","LU2326559502.SGD":"Natixis Loomis Sayles US Growth Equity P/A SGD-H","LU2168564065.EUR":"AZ ALLOCATION - TREND \"AAZ\" (EUR) ACC","SG9999018857.SGD":"United Global Quality Growth Fd Cl Acc SGD-H","LU1645745479.HKD":"ALLIANZ EUROPE INCOME AND GROWTH \"AM\" (HKDHDG) INC","LU1935042991.SGD":"MANULIFE GF GLOBAL MULTI-ASSET DIVERSIFIED INCOME  \"AA\" (SGDHDG) INC","IE0002270589.USD":"LEGG MASON CLEARBRIDGE VALUE \"A\" (USD) INC","LU2168563687.JPY":"AZ ALLOCATION - TREND \"AAZ\" (JPYHDG) ACC","LU1935042488.USD":"MANULIFE GF GLOBAL MULTI-ASSET DIVERSIFIED INCOME  \"AA\" (USD) INC","BK4566":"资本集团","LU0251132253.USD":"FIDELITY FUNDS  GLOBAL THEMATIC OPPORTUNITIES \"A\" (USD) ACC","LU1633808545.USD":"ALLIANZ GLOBAL EQUITY GROWTH \"AT\" (USD) ACC","LU2168564149.EUR":"AZ ALLOCATION - TREND \"BAZ\" (EUR) ACC","LU1778281490.HKD":"HSBC GIF GLOBAL LOWER CARBON EQUITY \"AD\" (HKD) INC","LU0251131958.USD":"FIDELITY AMERICA \"A\" (USD) ACC","LU1066051225.USD":"HSBC GIF GLOBAL EQUITY VOLATILITY FOCUSED \"AC\" (USD) ACC","LU0476273544.USD":"CT (LUX) I GLOBAL TECHNOLOGY \"BU\" (USD) ACC","LU1732799900.SGD":"JPMorgan Investment Funds - Global Income A (irc) SGD-H","LU0353189680.USD":"富国美国全盘成长基金Cl A Acc","LU0348723411.USD":"ALLIANZ GLOBAL HI-TECH GROWTH \"A\" (USD) INC","LU1917777945.USD":"安联专题基金Cl AT Acc","LU2271345857.HKD":"ALLIANZ GLOBAL SUSTAINABILITY \"AT\" (HKD) ACC","LU0553294199.USD":"BGF GLOBAL EQUITY INCOME \"A5G\" (USD) INC","LU1804176565.USD":"EASTSPRING INV GLOBAL GROWTH EQUITY \"A\" (USD) ACC","LU1923622614.USD":"Natixis Thematics Meta R/A USD","LU1699723380.USD":"ALLSPRING GLOBAL LONG/SHORT EQUITY \"AP\" (USD) ACC","BK4561":"索罗斯持仓","LU2242650005.HKD":"FIDELITY FUNDS GLOBAL MULTI ASSET DYNAMIC \"A\" (HKD) ACC","LU0964807845.USD":"ALLIANZ INCOME & GROWTH \"A\" (USD) INC","SG9999015978.USD":"利安颠覆性创新基金A","LU2211815571.USD":"ALLIANZ POSITIVE CHANGE \"AT\" (USD) ACC","IE00B19Z3B42.SGD":"Legg Mason ClearBridge - Value A Acc SGD","LU0211328371.USD":"TEMPLETON GLOBAL EQUITY INCOME \"A\" (MDIS) (USD)   INC","LU0784385170.HKD":"BGF GLOBAL MULTI ASSET INCOME \"A6\" (HKDHGD) INC","LU0784385840.USD":"Blackrock Global Multi-Asset Income A2 USD","LU2433249047.HKD":"THEMATICS META \"R/A\" (HKD) ACC","GOOGL":"谷歌A","BK4588":"碎股","LU0174119429.USD":"T ROWE PRICE US LARGE CAP GROWTH EQUITY \"A\" ACC","LU2063271972.USD":"富兰克林创新领域基金","GOOG":"谷歌","LU1935043536.SGD":"MANULIFE GF GLOBAL MULTI-ASSET DIVERSIFIED INCOME \"AA\" (SGDHDG) INC A","LU0127658192.USD":"EASTSPRING INVESTMENTS GLOBAL TECHNOLOGY \"A\" (USD) ACC","LU2462591350.SGD":"T ROWE PRICE US LARGE CAP GROWTH EQUITY \"AN\" (SGDHDG) ACC","LU1400636574.HKD":"ALLIANZ EUROPE INCOME AND GROWTH \"AMG\" (HKDHDG) INC","LU2213496289.HKD":"ALLIANZ INCOME AND GROWTH \"AT\" (HKD) ACC","LU0345768153.USD":"NINETY ONE GSF GLOBAL STRATEGIC MANAGED \"A\" (USD) ACC","LU1066051498.USD":"HSBC GIF GLOBAL EQUITY VOLATILITY FOCUSED \"AM2\" (USD) INC","LU0265550359.USD":"BGF SYSTEMATIC GLOBAL ENHANCED EQUITY YIELD \"A2\" (USD) ACC","LU0557290698.USD":"施罗德环球可持续增长基金","LU1815336760.USD":"THREADNEEDLE (LUX) GLOBAL TECHNOLOGY \"AUP\" (USD) INC","SG9999014575.USD":"UOB UNITED INCOME FOCUS TRUST FUND (USDHDG) INC","LU0198837287.USD":"UBS (LUX) EQUITY SICAV - USA GROWTH \"P\" (USD) ACC","LU0323591593.USD":"SCHRODER ISF QEP GLOBAL QUALITY \"A\" (USD) ACC","LU0820561909.HKD":"ALLIANZ INCOME AND GROWTH \"AM\" (HKD) INC","LU2097344357.USD":"SCHRODER ISF SUSTAINABLE MULTI-ASSET INCOME \"A\" (USDHDG) ACC","BK4585":"ETF&股票定投概念","LU0976567544.SGD":"FTIF - Templeton Global Income A Mdis SGD-H1","IE00BDRTCR15.USD":"PINEBRIDGE GLOBAL DYNAMIC ASSET ALLOCATION \"ADC\" (USD) INC A","LU1116320737.USD":"BGF SYSTEMATIC GLOBAL ENHANCED EQUITY YIELD \"A6\" (USD) INC","LU1791710582.SGD":"Fidelity Global Demographics A-ACC-SGD (SGD/USD hedged)","LU2097344431.USD":"SCHRODER ISF SUSTAINABLE MULTI-ASSET INCOME \"A\" (USDHDG) INC","IE00B19Z8W00.USD":"FTGF CLEARBRIDGE US LARGE CAP GROWTH \"A\" INC","LU1791710400.SGD":"Fidelity Global Demographics A-ACC-SGD","LU2125909593.SGD":"Natixis Thematics Meta R/A SGD","BK4543":"AI","LU0256863811.USD":"ALLIANZ US EQUITY \"A\" INC","BK4592":"伊斯兰概念","LU0056508442.USD":"贝莱德世界科技基金A2","LU2054465674.USD":"UBS (LUX) KEY SELEC SICAV DIGITAL TRANSFORMATION T \"P\" (USD) ACC","LU0689472784.USD":"安联收益及增长基金Cl AM AT Acc","LU2403377893.USD":"ALLIANZ SELECT INCOME AND GROWTH \"AM\" (USD) INC","LU2265009873.SGD":"Eastspring Investments - Global Growth Equity AS SGD-H","LU0267386448.USD":"FIDELITY FIRST ALL COUNTRY WORLD \"A\" (USD) INC","LU1116320901.HKD":"BGF SYSTEMATIC GLOBAL ENHANCED EQUITY YIELD \"A6\" (HKD) INC","LU2250418816.HKD":"BGF WORLD TECHNOLOGY \"A\" (HKD) ACC","IE0034235295.USD":"PINEBRIDGE GLOBAL DYNAMIC ASSET ALLOCATION \"A\" (USD) ACC","LU0052756011.USD":"TEMPLETON GLOBAL BALANCED \"A\" (USD) INC","IE00BJTD4V19.USD":"NEUBERGER BERMAN US LONG SHORT EQUITY \"A1\" (USD) ACC","LU1894683264.USD":"AMUNDI FUNDS US EQUITY RESEARCH VALUE \"A2\" (USD) ACC","SG9999014559.SGD":"United Income Focus Trust Dis SGD"},"source_url":"https://wallstreetcn.com/articles/3784092","is_english":false,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"2677649650","content_text":"近日，谷歌发布旗舰模型Gemini 4 Argon，这款模型在72小时内消耗了25万亿Token，是上一代Gemini 3.0同期的三倍，并创下谷歌历史纪录。谷歌CEO皮查伊（Sundar Pichai）在模型发布后的首次公开深度访谈中，正面回应了外界关于“谷歌放弃前沿AI”的质疑，并罕见承认公司云业务正在为前沿模型研发主动让渡算力资源。皮查伊坐进硅谷山景城的录音棚，接受科技媒体人Alex Kantrowitz主持的《Big Technology Podcast》专访。他证实了全行业算力严重短缺的判断——当下每家公司的实际需求都是两年前规划产能的2倍，客户今天想锁定2027年的算力“极其困难”。在递归自我改进（RSI）问题上，皮查伊明确划定边界：当前阶段是AI智能体与人类研究员协同推进的过渡期，距离“自主失控的终局”仍相当遥远。前沿模型，是谷歌的“第一优先级”，不是选项过去一年，硅谷流传着一种颇具市场说服力的逻辑：前沿大模型的军备竞赛已经进入边际收益递减区。Meta的Llama系列开源模型、谷歌自家的Gemini Flash这类轻量快速模型，对绝大多数企业场景已经“够用”。既然如此，谷歌何必继续烧上百亿美元去卷万亿参数的前沿模型？不如把TPU算力全部高价出租给外部开发者，做下一个英伟达。这种论调在投行报告和企业CTO圈子里流传甚广，甚至一度形成了“谷歌实际上已经战略性放弃前沿竞争”的市场预期。皮查伊在访谈中对这个逻辑做了正面反驳，但他的切入角度并非情绪性的否认，而是从三个具体机制展开：第一个机制是“早看见”的优势。 处于前沿的公司能比其他人更早观察到涌现能力，从而更早为下一代技术做准备。这是一种信息优势，而非单纯的性能优势。第二个机制是RSI（递归自我改进）。 在用模型来开发下一代模型的时代，前沿模型的质量直接决定了下一代模型的天花板。用一个性能平庸的模型来搭建研发流水线，相当于用低精度工具来制造精密仪器——误差会逐层放大。第三个机制是内部验证。 Argon上线后，谷歌内部开发者的使用数据给出了直接证据：我们在内部已经清楚地看到了一个陡峭的拐点，把Argon部署给我们的开发者之后，它能解决我们以前的模型根本做不到的问题。皮查伊还给出了一个具体数字：Argon上线四周内，谷歌内部代码销售量提升了30%，代码回滚率下降了约三分之一。这意味着工程师不仅写得更快，还写得更准——这是一个质量提升的信号，而非单纯的效率指标。对于模型“够用论”，皮查伊的态度是承认小模型的价值，但划清了边界：各种尺寸和能力的模型都有存在的空间，我们一直是这样做的，这也是为什么我们做了Flash、Flash Light等等。我同意你可以用不一定是最前沿的模型来构建很好的产品体验，因为前沿模型往往更慢也更贵。但我们在乎前沿的进展，我们致力于把所有这些都做好。算力账本：云业务正在为DeepMind让路这是整场访谈信息密度最高的一段，也是皮查伊最罕见的一次公开表态。理解这段话需要先搞清楚一个结构性矛盾：谷歌同时扮演着两个角色——一个是烧钱冲AGI的前沿AI研究机构（Google DeepMind），另一个是靠出租算力赚高毛利收入的云服务商（Google Cloud）。这两个角色在算力这个最稀缺的资源上，存在直接竞争。华尔街的逻辑是：每一块TPU拿去给DeepMind训练模型，就少了一块能对外出租赚钱的算力。在资本开支居高不下、短期云收入增速承压的财报季，这个矛盾被反复放大。皮查伊在访谈中直接说清楚了谷歌内部的优先级排序：我们的基本出发点，是确保我们投入足够的算力来保持在AI前沿的地位。这是不可谈判的，其他一切都在这个基础上运作。更进一步，他承认了外界长期猜测但从未得到CEO级别确认的事实——云业务确实在为前沿AI研究做出让步：我们的云业务在任何给定时刻，都在为了支持前沿AI开发而放弃短期收入，因为我们在做算力上的取舍。所以你说的那种情况，已经是现实了。这句话的分量在于：它不是一个未来的战略承诺，而是对当下现状的描述。谷歌云在实际运营中，已经在系统性地压缩对外出租算力的规模，以保证DeepMind的研发需求。但皮查伊也补充了一个反直觉的判断——适度的资源约束对研究本身未必是坏事：就算是我们的AI研究员，一定程度的约束其实也是好事，那是激发创造力的来源，那是促使你想出更高效的算法突破、比如用更少内存的来源。这段话背后是一个更深的逻辑：无限算力不等于无限进步，资源约束有时会倒逼出更根本性的架构创新。算力危机：2027年的配额，“极其困难”访谈中，主持人Kantrowitz提到了一个在硅谷风投圈广泛流传的信息——知名风投a16z联合创始人Ben Horowitz在播客中转述亚马逊CEO安迪·贾西（Andy Jassy）的话，称云算力已经\"卖空到2028年\"。皮查伊的回应，是迄今为止对这一说法最直接的CEO级别确认：我觉得这个说法是合理的……如果今天有客户来找我们要2027年的算力，那是极其困难的事情。为什么会到这一步？皮查伊给出了一个清晰的结构性解释：数据中心从决策到投产有2至3年的刚性滞后，而智能体任务的爆发完全超出了两年前的需求预测。今天回头看，每家公司都是短缺的，不只是我们，其他公司也一样。大家都希望自己手里的算力是当初规划量的2倍。皮查伊还补充了一个关键的时间锚点：我们今天手里有的算力，是我们在2024年做的决策的产物。就算回到去年今天，智能体任务的爆发还没有发生。这句话揭示了一个系统性问题：AI需求的增速已经超过了任何基础设施规划模型的预测能力。两年前做出的决策，在今天看来几乎必然是低估的。对于“是否存在过度建设（overbuilding）”的追问，皮查伊的回答是区分时间框架：如果我拉一个足够长的时间框架，这个量级的建设对我来说完全合理。我们处于有史以来最重要的技术的早期阶段，这不像光纤——光纤铺完是暗的，没人用。我们是在每天都面临产能约束的情况下使用它。他没有否认短期内存在局部过剩的可能——电力、内存、电工等2000多个配套要素中任何一个出现瓶颈，都可能导致局部积压。但他对长期需求的判断是确定性的：我有没有信心，我们规划的算力最终会有用武之地？给足时间的话，绝对有。RSI：不是智能爆炸，是“智能体和研究员一起干活”递归自我改进（RSI）是这次访谈中技术密度最高的议题，也是当前AI安全讨论中争议最大的话题之一。背景是：谷歌联合创始人谢尔盖·布林此前在一次DeepMind内部研究员圆桌上提到，Argon的核心创新之一是“更多地依赖模型来帮助开发模型”。这句话随即在AI安全社区引发了关于“智能爆炸”的讨论。主持人Kantrowitz在访谈中直接抛出了这个问题，并提到了近期几起AI系统越权行为的案例。皮查伊的回应把RSI的当前状态拆解成了三个层次：第一层：定义当前位置。 他明确指出，当前的RSI远没有到达科幻叙事中\"自主失控\"的终局状态：在一个连续谱上，我们距离RSI的终态还相当遥远。今天正确的描述方式是：你有RSI循环，同时有AI研究员在模型开发的各个环节共同参与。我会说，你几乎是有智能体在某种自动化方式下，帮助AI研究员协同工作来构建这些模型。第二层：安全机制。 他把安全框架描述为\"纵深防御\"——对齐研究是一层，但不是唯一的一层。在此之上还有实时遥测监控、沙箱隔离架构、以及给企业客户的部署控制权限。第三层：行业标准化。 皮查伊把下一步的关键指向了跨行业标准建设：就像金融控制领域有会计准则、有独立审计师一样，这些都是我们作为一个行业需要努力推进的下一步，包括使用独立评估机构进行评估的能力。这段话的信息量在于：它承认了当前的对齐工作存在\"明显缺口\"，但将解决路径指向了外部标准化机制，而非单纯依赖各家公司的内部自律。太空数据中心与长期下注访谈的最后部分，皮查伊谈到了谷歌的几个长期战略赌注，其中最具想象空间的是太空数据中心项目。谷歌已将TPU试验芯片送入太阳同步轨道，探索在太空中利用24小时不间断太阳能运行数据中心。在地面数据中心面临社区反弹、电网并网排队长达数年的背景下，这一方向的战略逻辑逐渐清晰。皮查伊用Waymo的起点来类比这个项目当前的阶段：这些技术让你想起那些时刻——你站在那里，它看起来有点像科幻，但你知道，10到20年之后，它实际上会以一种有意义的方式运转。他透露，在访谈录制期间，团队正在等待太空中TPU的\"Hello World\"时刻——即首次成功运行的信号。这个项目目前处于\"两到三年\"的早期阶段，而谷歌的量子计算项目已经走了将近20年。皮查伊对这类长周期赌注的态度，折射出谷歌在基础设施层面的时间观：在太阳同步轨道上，以24小时不间断的方式获取我们能接触到的最丰富的能源形式，对于我们将要面对的算力指数级增长来说，这将非常有意义。","news_type":1,"symbols_score_info":{"GOOGL":1.97,"GOOG":2}},"isVote":1,"tweetType":1,"viewCount":182,"commentLimit":10,"likeStatus":false,"favoriteStatus":false,"reportStatus":false,"symbols":[],"verified":2,"subType":0,"readableState":1,"langContent":"CN","currentLanguage":"CN","warmUpFlag":false,"orderFlag":false,"shareable":true,"causeOfNotShareable":"","featuresForAnalytics":[],"commentAndTweetFlag":false,"andRepostAutoSelectedFlag":false,"upFlag":false,"length":115,"optionInvolvedFlag":false,"subscribersOnly":false,"subscribersOnlyAccessible":false,"xxTargetLangEnum":"ZH_CN"},"commentList":[],"hasMoreComment":false,"orderType":2}