Profile - To The Moon

TIDARAT

Posts · 121Posts · 121
Following · 0Following · 0
Followers · 0Followers · 0

- TIDARATTIDARAT
  ·05-05
  ,0 ya
  UOB Kay Hian Downgrades XLSMART Telecom Sejahtera to Hold from Buy; Price Target is IDR3,100
  XLSMART Telecom Sejahtera (IDX:EXCL) has an average rating of overweight and mean price target of 2,867.06 Indonesian rupiah, according to analysts polled by FactSet.(MT Newswires covers equity, commo
  UOB Kay Hian Downgrades XLSMART Telecom Sejahtera to Hold from Buy; Price Target is IDR3,100
  780 Comment
  1
  Share
  Report
- TIDARATTIDARAT
  ·2021-09-23
  Good
  Why exactly does Amazon block Chinese sellers?
  因为封号冻资事件，亚马逊今年一直被热搜。所以导致整个跨境电商笼罩的信息一直是今天这个大卖关了，明天那个大卖裁员了。特别是在一些自媒体流量吞噬下，行业似乎就没有一些积极事件发生。这里面比较关键的信息
  Why exactly does Amazon block Chinese sellers?
  4.91K 1
  3
  Share
  Report
- TIDARATTIDARAT
  ·2021-09-21
  Good
  Sorry, this post has been deleted
  4.59K Comment
  1
  Share
  Report
- TIDARATTIDARAT
  ·2021-09-10
  Good
  Sorry, this post has been deleted
  3.70K Comment
  4
  Share
  Report
- TIDARATTIDARAT
  ·2021-09-09
  Good
  When was "Nvidia" born in China?
  这一事件引发了人们对虚拟现实、元宇宙、AI换脸等技术和概念的激烈讨论，同时也让“英伟达”这家美国芯片霸主从半导体行业“出圈”，走入了大众视野。另外，作为全球芯片销量大国，中国却没有出现一家“英伟达”这样的芯片巨头，大市场并没有产生与之匹配的大公司。
  When was "Nvidia" born in China?
  4.07K Comment
  3
  Share
  Report
- TIDARATTIDARAT
  ·2021-09-09
  Omg
  Multiple Indicators in the United States Alarm!
  美国人曾憧憬，这个夏天经济能回归常态，上班族回到办公室，孩子们重返校园，街角的咖啡店再次开业。但一切事与愿违，随着疫情复燃，复苏在8月踩下了急刹车。上周五，让人大跌眼镜的非农就业数据带来了最直接的报
  Multiple Indicators in the United States Alarm!
  4.47K Comment
  7
  Share
  Report
- TIDARATTIDARAT
  ·2021-08-24
  Good
  Sorry, this post has been deleted
  3.81K Comment
  1
  Share
  Report
- TIDARATTIDARAT
  ·2021-08-23
  Hi
  Sorry, this post has been deleted
  5.77K 1
  4
  Share
  Report
- TIDARATTIDARAT
  ·2021-08-23
  Good
  Sorry, this post has been deleted
  3.93K Comment
  1
  Share
  Report
- TIDARATTIDARAT
  ·2021-08-21
  555
  Sorry, this post has been deleted
  3.86K Comment
  1
  Share
  Report

Most Discussed

{"i18n":{"language":"en_US"},"isCurrentUser":false,"userPageInfo":{"id":"3578895363297107","uuid":"3578895363297107","gmtCreate":1615817074645,"gmtModify":1624678721238,"name":"TIDARAT","pinyin":"tidarat","introduction":"","introductionEn":null,"signature":"","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","hat":null,"hatId":null,"hatName":null,"vip":1,"status":2,"fanSize":7,"headSize":348,"tweetSize":121,"questionSize":0,"limitLevel":999,"accountStatus":4,"level":{"id":1,"name":"萌萌虎","nameTw":"萌萌虎","represent":"呱呱坠地","factor":"评论帖子3次或发布1条主帖（非转发）","iconColor":"3C9E83","bgColor":"A2F1D9"},"themeCounts":0,"badgeCounts":0,"badges":[],"moderator":false,"superModerator":false,"manageSymbols":null,"badgeLevel":null,"boolIsFan":false,"boolIsHead":false,"favoriteSize":0,"symbols":null,"coverImage":null,"realNameVerified":"success","userBadges":[{"badgeId":"1026c425416b44e0aac28c11a0848493-3","templateUuid":"1026c425416b44e0aac28c11a0848493","name":" Tiger Idol","description":"Join the tiger community for 1500 days","bigImgUrl":"https://static.tigerbbs.com/8b40ae7da5bf081a1c84df14bf9e6367","smallImgUrl":"https://static.tigerbbs.com/f160eceddd7c284a8e1136557615cfad","grayImgUrl":"https://static.tigerbbs.com/11792805c468334a9b31c39f95a41c6a","redirectLinkEnabled":0,"redirectLinkType":null,"redirectLink":null,"redirectLinkValidityFrom":null,"redirectLinkValidityTo":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2025.04.24","exceedPercentage":null,"individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":1001,"isScarce":0,"effectConfig":null,"effectEnabled":0,"plateImgUrl":null,"plateColors":null,"validityTo":null,"validityToTimestamp":null,"wearingSort":0},{"badgeId":"44212b71d0be4ec88898348dbe882e03-2","templateUuid":"44212b71d0be4ec88898348dbe882e03","name":"Executive Tiger","description":"The transaction amount of the securities account reaches $300,000","bigImgUrl":"https://static.tigerbbs.com/9d20b23f1b6335407f882bc5c2ad12c0","smallImgUrl":"https://static.tigerbbs.com/ada3b4533518ace8404a3f6dd192bd29","grayImgUrl":"https://static.tigerbbs.com/177f283ba21d1c077054dac07f88f3bd","redirectLinkEnabled":0,"redirectLinkType":null,"redirectLink":null,"redirectLinkValidityFrom":null,"redirectLinkValidityTo":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2023.07.14","exceedPercentage":"80.05%","individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":1101,"isScarce":0,"effectConfig":null,"effectEnabled":0,"plateImgUrl":null,"plateColors":null,"validityTo":null,"validityToTimestamp":null,"wearingSort":0},{"badgeId":"cbe1c45c584340f3bd9ae6c77e0e9981-1","templateUuid":"cbe1c45c584340f3bd9ae6c77e0e9981","name":"Academy Experiencer","description":"5 lessons learned","bigImgUrl":"https://community-static.tradeup.com/news/fb5ae275631fb96a92d475cdc85d2302","smallImgUrl":"https://community-static.tradeup.com/news/c2660a1935bd2105e97c9915619936c3","grayImgUrl":null,"redirectLinkEnabled":0,"redirectLinkType":null,"redirectLink":null,"redirectLinkValidityFrom":null,"redirectLinkValidityTo":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2022.08.06","exceedPercentage":null,"individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":2006,"isScarce":0,"effectConfig":null,"effectEnabled":0,"plateImgUrl":null,"plateColors":null,"validityTo":null,"validityToTimestamp":null,"wearingSort":0},{"badgeId":"7a9f168ff73447fe856ed6c938b61789-1","templateUuid":"7a9f168ff73447fe856ed6c938b61789","name":"Knowledgeable Investor","description":"Traded more than 10 stocks","bigImgUrl":"https://static.tigerbbs.com/e74cc24115c4fbae6154ec1b1041bf47","smallImgUrl":"https://static.tigerbbs.com/d48265cbfd97c57f9048db29f22227b0","grayImgUrl":"https://static.tigerbbs.com/76c6d6898b073c77e1c537ebe9ac1c57","redirectLinkEnabled":0,"redirectLinkType":null,"redirectLink":null,"redirectLinkValidityFrom":null,"redirectLinkValidityTo":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2021.12.21","exceedPercentage":null,"individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":1102,"isScarce":0,"effectConfig":null,"effectEnabled":0,"plateImgUrl":null,"plateColors":null,"validityTo":null,"validityToTimestamp":null,"wearingSort":0},{"badgeId":"a83d7582f45846ffbccbce770ce65d84-1","templateUuid":"a83d7582f45846ffbccbce770ce65d84","name":"Real Trader","description":"Completed a transaction","bigImgUrl":"https://static.tigerbbs.com/2e08a1cc2087a1de93402c2c290fa65b","smallImgUrl":"https://static.tigerbbs.com/4504a6397ce1137932d56e5f4ce27166","grayImgUrl":"https://static.tigerbbs.com/4b22c79415b4cd6e3d8ebc4a0fa32604","redirectLinkEnabled":0,"redirectLinkType":null,"redirectLink":null,"redirectLinkValidityFrom":null,"redirectLinkValidityTo":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2021.12.21","exceedPercentage":null,"individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":1100,"isScarce":0,"effectConfig":null,"effectEnabled":0,"plateImgUrl":null,"plateColors":null,"validityTo":null,"validityToTimestamp":null,"wearingSort":0},{"badgeId":"972123088c9646f7b6091ae0662215be-2","templateUuid":"972123088c9646f7b6091ae0662215be","name":"Master Trader","description":"Total number of securities or futures transactions reached 100","bigImgUrl":"https://static.tigerbbs.com/ad22cfbe2d05aa393b18e9226e4b0307","smallImgUrl":"https://static.tigerbbs.com/36702e6ff3ffe46acafee66cc85273ca","grayImgUrl":"https://static.tigerbbs.com/d52eb88fa385cf5abe2616ed63781765","redirectLinkEnabled":0,"redirectLinkType":null,"redirectLink":null,"redirectLinkValidityFrom":null,"redirectLinkValidityTo":null,"hasAllocated":1,"isWearing":0,"stamp":null,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2021.12.21","exceedPercentage":"80.15%","individualDisplayEnabled":0,"backgroundColor":null,"fontColor":null,"individualDisplaySort":0,"categoryType":1100,"isScarce":0,"effectConfig":null,"effectEnabled":0,"plateImgUrl":null,"plateColors":null,"validityTo":null,"validityToTimestamp":null,"wearingSort":0}],"userBadgeCount":6,"currentWearingBadge":null,"individualDisplayBadges":null,"crmLevel":11,"crmLevelSwitch":0,"location":null,"starInvestorFollowerNum":0,"starInvestorFlag":false,"starInvestorOrderShareNum":0,"subscribeStarInvestorNum":2,"ror":null,"winRationPercentage":null,"showRor":false,"investmentPhilosophy":null,"starInvestorSubscribeFlag":false},"page":1,"watchlist":null,"tweetList":[{"id":560827356808184,"gmtCreate":1777952572696,"gmtModify":1777952576494,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":",0 ya","listText":",0 ya","text":",0 ya","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":1,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/560827356808184","repostId":"2588725924","repostType":2,"repost":{"id":"2588725924","kind":"highlight","weMediaInfo":{"introduction":"The most recognized names in North America, Europe and Asia rely on MT Newswires to power their applications. Better news, better service, better price.","home_visible":1,"media_name":"MT Newswires Live","id":"1092851196","head_image":"https://community-static.tradeup.com/news/3002d84abbd5ace3c99397c7f95b8d4e"},"pubTimestamp":1764583136,"share":"https://ttm.financial/m/news/2588725924?lang=en_US&edition=fundamental","pubTime":"2025-12-01 17:58","market":"fut","language":"en","title":"UOB Kay Hian Downgrades XLSMART Telecom Sejahtera to Hold from Buy; Price Target is IDR3,100","url":"https://stock-news.laohu8.com/highlight/detail?id=2588725924","media":"MT Newswires Live","summary":"XLSMART Telecom Sejahtera (IDX:EXCL) has an average rating of overweight and mean price target of 2,867.06 Indonesian rupiah, according to analysts polled by FactSet.(MT Newswires covers equity, commo","content":"<html><body><p> XLSMART Telecom Sejahtera (IDX:EXCL) has an average rating of overweight and mean price target of 2,867.06 Indonesian rupiah, according to analysts polled by FactSet.</p><p>(MT Newswires covers equity, commodity and economic research from major banks and research firms in North America, Asia and Europe. Research providers may contact us here: https://www.mtnewswires.com/contact-us)</p></body></html>","source":"mtnewswires_news","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>UOB Kay Hian Downgrades XLSMART Telecom Sejahtera to Hold from Buy; Price Target is IDR3,100</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\">\nUOB Kay Hian Downgrades XLSMART Telecom Sejahtera to Hold from Buy; Price Target is IDR3,100\n</h2>\n\n<h4 class=\"meta\">\n\n\n<a class=\"head\" href=\"https://laohu8.com/wemedia/1092851196\">\n\n\n<div class=\"h-thumb\" style=\"background-image:url(https://community-static.tradeup.com/news/3002d84abbd5ace3c99397c7f95b8d4e);background-size:cover;\"></div>\n\n<div class=\"h-content\">\n<p class=\"h-name\">MT Newswires Live </p>\n<p class=\"h-time\">2025-12-01 17:58</p>\n</div>\n\n</a>\n\n\n</h4>\n\n</header>\n<article>\n<html><body><p> XLSMART Telecom Sejahtera (IDX:EXCL) has an average rating of overweight and mean price target of 2,867.06 Indonesian rupiah, according to analysts polled by FactSet.</p><p>(MT Newswires covers equity, commodity and economic research from major banks and research firms in North America, Asia and Europe. Research providers may contact us here: https://www.mtnewswires.com/contact-us)</p></body></html>\n\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"","relate_stocks":{},"source_url":"https://www.mtnewswires.com/","is_english":true,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"2588725924","content_text":"XLSMART Telecom Sejahtera (IDX:EXCL) has an average rating of overweight and mean price target of 2,867.06 Indonesian rupiah, according to analysts polled by FactSet.(MT Newswires covers equity, commodity and economic research from major banks and research firms in North America, Asia and Europe. Research providers may contact us here: https://www.mtnewswires.com/contact-us)","news_type":1,"symbols_score_info":{}},"isVote":1,"tweetType":1,"viewCount":780,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":863524636,"gmtCreate":1632407213825,"gmtModify":1676530775482,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Good","listText":"Good","text":"Good","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":3,"commentSize":1,"repostSize":0,"link":"https://ttm.financial/post/863524636","repostId":"2169565516","repostType":4,"repost":{"id":"2169565516","kind":"news","pubTimestamp":1632364656,"share":"https://ttm.financial/m/news/2169565516?lang=en_US&edition=fundamental","pubTime":"2021-09-23 10:37","market":"us","language":"zh","title":"Why exactly does Amazon block Chinese sellers?","url":"https://stock-news.laohu8.com/highlight/detail?id=2169565516","media":"36氪","summary":"因为封号冻资事件，亚马逊今年一直被热搜。\n所以导致整个跨境电商笼罩的信息一直是今天这个大卖关了，明天那个大卖裁员了。特别是在一些自媒体流量吞噬下，行业似乎就没有一些积极事件发生。\n这里面比较关键的信息","content":"<p>Because of the account ban and capital freezing incident,<a href=\"https://laohu8.com/S/AMZN\">Amazon</a>It has been hot searched this year.</p><p>Therefore, the information that leads to the whole cross-border e-commerce has always been that today's big sale is closed, and tomorrow's big sale is laid off. Especially when some We Media traffic is swallowed up, it seems that there are no positive events in the industry.</p><p>The key information asymmetry here is how many Chinese stores has Amazon sealed? At first, the most mainstream rumor was that there were 50,000. As Amazon officials kept silent, this number gradually defaulted to fact.</p><p>Even at Amazon's event of settling the world's first comprehensive seller training center in Hangzhou held last Friday, it was still tight-lipped about the number of closed stores, and didn't plan to disclose it to the outside world. It was hoped that the cold treatment would pass.</p><p>The invited media will certainly not miss this rare opportunity, so they turned a Q&A session of settlement activity into a press conference of store closure tide. So much so that Cindy Tai, global vice president of Amazon and executive president of Amazon Global Shop Asia Pacific, said frankly, \"Today's release was originally for the training center to settle in Hangzhou, but I didn't expect so many media to pay attention to the account being blocked, so I made the first response together.\"</p><p>According to her, in the past five months, Amazon has closed the sales rights of about 600 Chinese sellers, involving about 3,000 brands, including some big sellers. \"These sellers have repeatedly, repeatedly and seriously abused reviews, as well as many other violations. In the past period, Amazon has warned these sellers many times, and they have had many opportunities to appeal, and even restored some deactivated accounts. However, these sellers continue to violate regulations, so this time it is decided to terminate the cooperation with these sellers.\"</p><p>In translation, the blocked seller violated the rules seriously, and Amazon was forced to take the last step after giving many opportunities.</p><p>\"Amazon only released the data for the first five months, but the data for June, July and August were not disclosed. Many stores were also affected in these three months. So the final number must be much larger than this.\" A senior practitioner told<a href=\"https://laohu8.com/S/KRKR\">36 Krypton</a>。</p><p>For sellers who rely heavily on Amazon, the loss is really not small. For example, once<a href=\"https://laohu8.com/S/01668\">China South City</a>There is a tree among the four young people, 340 shops have been closed, and 130 million yuan of funds have been frozen. It directly led to half of the employees leaving their jobs, half of their income, and a net loss of over 700 million yuan; Another big seller, Tongtuo Technology, was banned from selling and closing 54 stores, and frozen funds of 41.43 million yuan.</p><p>So, why is Amazon's store closure so strong this time? What are the things behind it that make us really reflect?</p><p>Advances in algorithms</p><p>In fact, there is an unspoken rule known in the industry. That is, Amazon has two fixed account sweeps every year, one on Prime Day around July. Amazon usually closes some accounts two months in advance for its large-scale promotion, that is, from the end of April to the beginning of May; The other is in the first 2 months of Christmas, i.e. mid-October or early November.</p><p>Because they see this huge loophole, many sellers have their own countermeasures. For example, if a brand authorizes 50 stores, it will have no impact at all if a few are sealed. \"In fact, as long as the number of shops closed does not exceed a certain percentage, even the boss doesn't need to know about this.\" A seller told 36Kr.</p><p>After last year's industry explosion, all sellers are more optimistic this year, and they don't see the real risks behind this incident at all. In fact, as early as 3 years ago, Amazon was using algorithms<a href=\"https://laohu8.com/S/300024\">Robot</a>In detecting whether product reviews are true or false. Once, many self-operated listings (product pages) have been blocked. For example, the Listing of a seller in Shenzhen has been reduced from 17,000 to 1,000, and the circle of friends is also a wail. Fortunately, because the error was too serious, Amazon stopped closing on a large scale.</p><p>But technology kept improving until it began to mature a lot this year. In addition, Amazon has changed the rules, and it is no longer judged according to the store logic, but according to the brand dimension logic. As long as the algorithm identifies the problematic brand, and it exceeds a certain proportion, all its stores will be closed. That's why a large number of store groups have been blocked.</p><p>36Kr also noticed that Andy Jassy, the new CEO of Amazon, was previously the CEO of Amazon Cloud Computing, not the head of e-commerce. This also proves that Amazon, as a data-driven company, pays more attention to the protection of the authenticity of platform data.</p><p>\"I believe that after this incident, many sellers will start to adopt the white hat game. And Amazon also encourages on-site reviews, and invites users to evaluate through on-site letters. If the product and experience are not good, the invitation will definitely be bad reviews. Therefore, this matter is very good for the whole ecology. If bad money drives out good money all the time, good money will not be possible.\" The above-mentioned seller said.</p><p>Supplier logic</p><p>In addition to the continuous iteration of the algorithm, Amazon's definition of sellers is also different from that of domestic Tmall Taobao. The difference is that Amazon regards all sellers on the platform as suppliers, and it is equivalent to a retailer.</p><p>That is, all of the seller's digital assets on Amazon belong to Amazon. Even if you are<a href=\"https://laohu8.com/S/AAPL\">Apple</a>, Yes<a href=\"https://laohu8.com/S/NKE\">Nike</a>, are all suppliers of Amazon.</p><p>\"It can be understood that Tmall is similar to a shopping center, and Amazon is a department store. So this has laid the foreshadowing of some punishment mechanisms of Amazon. If Tmall sellers violate the regulations, will the platform move your goods and funds? Amazon talks more about sales scale, while Tmall talks about GMV.\" Zhou Jun, vice president of sales of Zhiyun Tiangong, told 36Kr.</p><p>It's different for Amazon. Once the customer's payment is received, it is responsible for the user. As long as it is found that the seller may cause harm to the customer, the funds will be frozen first to cope with whether the subsequent prosecution will bear joint and several liability.</p><p>This was also confirmed by Cindy Tai. The Amazon funds freeze period is 90 days, which is mainly used to bear the seller's refund return compensation costs of past customers, as well as other unpaid expenses. After 90 days, if there is no violation, you can apply for retrieval.</p><p>Therefore, when you open a shop in Amazon, you are actually not selling goods yourself, but the seller supplies goods to the platform, and the platform then helps sell goods. Amazon is not an administrative and judicial body, and has no power to freeze funds and goods. But because of Amazon's supplier logic, as long as it violates the supplier rules, it has to be kicked out of the directory.</p><p>\"In terms of values, brushing bills is a serious act of lying and fraud, and we all hate it. But we can't blindly stigmatize Chinese sellers. Everyone in the industry is working day and night. What Amazon doesn't do well is that there is at least a hearing to give sellers a self-discrimination link. For sellers, the legitimate rights and interests that should be fought for must be fought for. If this time passes in silence, the future will probably be a nightmare. \" Zhou Jun said.</p><p>In the complaint process, Cindy Tai, vice president of Amazon Global and CEO of Amazon Global Store Asia Pacific, said that if a seller's account is deactivated because of violations, Amazon will provide sellers with an opportunity to appeal; As long as the seller proves that Amazon is wrong in judgment, or that the violation is only temporary or unintentional, and provides a solution to avoid another violation, these accounts will resume normal operation.</p><p>However, according to 36Kr from many sellers, the probability of successful appeal from Amazon is almost 0. At the same time, it is not very accurate to grasp the variety of changes in Amazon's rules. The main problem lies in the inefficiency of email communication.</p><p>Where's the way out?</p><p>For the whole cross-border e-commerce industry, it is still in a very early stage, and it is not even an industry. 36 Krypton found from a number of investment institutions visited that investors' optimistic judgment on the industry has not been affected in any way. What has been looking for is which enterprises can cross the cycle and have the consciousness and ability to come out of a new brand.</p><p>It is difficult to build a brand. What is needed is the comprehensive ability of enterprises to meet user needs, marketing methods, supply chain impact, etc. However, it is precisely because of the high threshold that it is even more scarce. Especially<a href=\"https://laohu8.com/S/300866\">Anker Innovation</a>And the success of Shein, which makes all investors very excited, look forward to finding the next such investment target.</p><p>In the future, the first thing the industry has to reconstruct is consciousness. The past distribution path is already a dead end, and the earlier it is, the more opportunities for transformation. A big seller in the industry lost a lot of first-mover dividends because he transformed his brand two years late. When he turned back and chased after him, he realized that the probability of success was extremely small. This is also the dividend path dependence left by the early stage of the industry. Because it was easier to make fast money in the past, they were reluctant to take a difficult and correct road with accumulation.</p><p>Then, find a subdivided vertical category to cultivate deeply. If it was a rough operation before, it must be refined now. Moreover, the threshold for category requirements has become higher. Why won't the previous simple rough mode work? Because consumers are awakening and have higher requirements for product quality and service, third parties like Amazon are gradually changing their strategies to encourage brand determination to be stronger. In such an environment, without the ability to refine, there may be no profit at all in the future.</p><p>Secondly, introduce some high-quality technical talents. The epidemic is the beginning of cross-border e-commerce out of the circle, and the title is accelerated. The direct impact is that more entrepreneurs enter across the border. Industries that seem to have no threshold are actually not low, especially in the practical stage. Perhaps some pure business logic can still be passed through until now, but creating competitive barriers requires more high-quality talents to join. Especially on the technical side, as far as the current industry practitioners are concerned, there is still a lot of room for improvement.</p><p>Finally, be good at using the power of capital. All along, because cross-border e-commerce is easy to make money, there is even no shortage of cash flow. Therefore, the desire for capital is relatively general, but if you sort out Anker Innovation and Shein, it will be obvious that the reason why these two companies have such a high valuation in the capital market, especially the high growth of the latter. It is inseparable from the help of capital. Capital is not good or bad, and the core lies in how to use it. Providing capital is only the underlying role, and greater empowerment will greatly help to become stronger and bigger.</p><p>From the perspective of any industry development cycle, the large-scale entry of capital is after the penetration rate has increased to a certain extent. At this time, some talents such as big factories, high education and rich industry experience are invested. Driven by technology, these innovative forces began to take a new path and reconstruct the business pattern. In other words, you don't accept investment and capital invests in your competitors.</p>","source":"tencent","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>Why exactly does Amazon block Chinese sellers?</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: 12.5px; color: #7E829C; margin: 0;}\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\">\nWhy exactly does Amazon block Chinese sellers?\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">36氪</strong><span class=\"h-time small\">2021-09-23 10:37</span>\n</p>\n</h4>\n</header>\n<article>\n<p>Because of the account ban and capital freezing incident,<a href=\"https://laohu8.com/S/AMZN\">Amazon</a>It has been hot searched this year.</p><p>Therefore, the information that leads to the whole cross-border e-commerce has always been that today's big sale is closed, and tomorrow's big sale is laid off. Especially when some We Media traffic is swallowed up, it seems that there are no positive events in the industry.</p><p>The key information asymmetry here is how many Chinese stores has Amazon sealed? At first, the most mainstream rumor was that there were 50,000. As Amazon officials kept silent, this number gradually defaulted to fact.</p><p>Even at Amazon's event of settling the world's first comprehensive seller training center in Hangzhou held last Friday, it was still tight-lipped about the number of closed stores, and didn't plan to disclose it to the outside world. It was hoped that the cold treatment would pass.</p><p>The invited media will certainly not miss this rare opportunity, so they turned a Q&A session of settlement activity into a press conference of store closure tide. So much so that Cindy Tai, global vice president of Amazon and executive president of Amazon Global Shop Asia Pacific, said frankly, \"Today's release was originally for the training center to settle in Hangzhou, but I didn't expect so many media to pay attention to the account being blocked, so I made the first response together.\"</p><p>According to her, in the past five months, Amazon has closed the sales rights of about 600 Chinese sellers, involving about 3,000 brands, including some big sellers. \"These sellers have repeatedly, repeatedly and seriously abused reviews, as well as many other violations. In the past period, Amazon has warned these sellers many times, and they have had many opportunities to appeal, and even restored some deactivated accounts. However, these sellers continue to violate regulations, so this time it is decided to terminate the cooperation with these sellers.\"</p><p>In translation, the blocked seller violated the rules seriously, and Amazon was forced to take the last step after giving many opportunities.</p><p>\"Amazon only released the data for the first five months, but the data for June, July and August were not disclosed. Many stores were also affected in these three months. So the final number must be much larger than this.\" A senior practitioner told<a href=\"https://laohu8.com/S/KRKR\">36 Krypton</a>。</p><p>For sellers who rely heavily on Amazon, the loss is really not small. For example, once<a href=\"https://laohu8.com/S/01668\">China South City</a>There is a tree among the four young people, 340 shops have been closed, and 130 million yuan of funds have been frozen. It directly led to half of the employees leaving their jobs, half of their income, and a net loss of over 700 million yuan; Another big seller, Tongtuo Technology, was banned from selling and closing 54 stores, and frozen funds of 41.43 million yuan.</p><p>So, why is Amazon's store closure so strong this time? What are the things behind it that make us really reflect?</p><p>Advances in algorithms</p><p>In fact, there is an unspoken rule known in the industry. That is, Amazon has two fixed account sweeps every year, one on Prime Day around July. Amazon usually closes some accounts two months in advance for its large-scale promotion, that is, from the end of April to the beginning of May; The other is in the first 2 months of Christmas, i.e. mid-October or early November.</p><p>Because they see this huge loophole, many sellers have their own countermeasures. For example, if a brand authorizes 50 stores, it will have no impact at all if a few are sealed. \"In fact, as long as the number of shops closed does not exceed a certain percentage, even the boss doesn't need to know about this.\" A seller told 36Kr.</p><p>After last year's industry explosion, all sellers are more optimistic this year, and they don't see the real risks behind this incident at all. In fact, as early as 3 years ago, Amazon was using algorithms<a href=\"https://laohu8.com/S/300024\">Robot</a>In detecting whether product reviews are true or false. Once, many self-operated listings (product pages) have been blocked. For example, the Listing of a seller in Shenzhen has been reduced from 17,000 to 1,000, and the circle of friends is also a wail. Fortunately, because the error was too serious, Amazon stopped closing on a large scale.</p><p>But technology kept improving until it began to mature a lot this year. In addition, Amazon has changed the rules, and it is no longer judged according to the store logic, but according to the brand dimension logic. As long as the algorithm identifies the problematic brand, and it exceeds a certain proportion, all its stores will be closed. That's why a large number of store groups have been blocked.</p><p>36Kr also noticed that Andy Jassy, the new CEO of Amazon, was previously the CEO of Amazon Cloud Computing, not the head of e-commerce. This also proves that Amazon, as a data-driven company, pays more attention to the protection of the authenticity of platform data.</p><p>\"I believe that after this incident, many sellers will start to adopt the white hat game. And Amazon also encourages on-site reviews, and invites users to evaluate through on-site letters. If the product and experience are not good, the invitation will definitely be bad reviews. Therefore, this matter is very good for the whole ecology. If bad money drives out good money all the time, good money will not be possible.\" The above-mentioned seller said.</p><p>Supplier logic</p><p>In addition to the continuous iteration of the algorithm, Amazon's definition of sellers is also different from that of domestic Tmall Taobao. The difference is that Amazon regards all sellers on the platform as suppliers, and it is equivalent to a retailer.</p><p>That is, all of the seller's digital assets on Amazon belong to Amazon. Even if you are<a href=\"https://laohu8.com/S/AAPL\">Apple</a>, Yes<a href=\"https://laohu8.com/S/NKE\">Nike</a>, are all suppliers of Amazon.</p><p>\"It can be understood that Tmall is similar to a shopping center, and Amazon is a department store. So this has laid the foreshadowing of some punishment mechanisms of Amazon. If Tmall sellers violate the regulations, will the platform move your goods and funds? Amazon talks more about sales scale, while Tmall talks about GMV.\" Zhou Jun, vice president of sales of Zhiyun Tiangong, told 36Kr.</p><p>It's different for Amazon. Once the customer's payment is received, it is responsible for the user. As long as it is found that the seller may cause harm to the customer, the funds will be frozen first to cope with whether the subsequent prosecution will bear joint and several liability.</p><p>This was also confirmed by Cindy Tai. The Amazon funds freeze period is 90 days, which is mainly used to bear the seller's refund return compensation costs of past customers, as well as other unpaid expenses. After 90 days, if there is no violation, you can apply for retrieval.</p><p>Therefore, when you open a shop in Amazon, you are actually not selling goods yourself, but the seller supplies goods to the platform, and the platform then helps sell goods. Amazon is not an administrative and judicial body, and has no power to freeze funds and goods. But because of Amazon's supplier logic, as long as it violates the supplier rules, it has to be kicked out of the directory.</p><p>\"In terms of values, brushing bills is a serious act of lying and fraud, and we all hate it. But we can't blindly stigmatize Chinese sellers. Everyone in the industry is working day and night. What Amazon doesn't do well is that there is at least a hearing to give sellers a self-discrimination link. For sellers, the legitimate rights and interests that should be fought for must be fought for. If this time passes in silence, the future will probably be a nightmare. \" Zhou Jun said.</p><p>In the complaint process, Cindy Tai, vice president of Amazon Global and CEO of Amazon Global Store Asia Pacific, said that if a seller's account is deactivated because of violations, Amazon will provide sellers with an opportunity to appeal; As long as the seller proves that Amazon is wrong in judgment, or that the violation is only temporary or unintentional, and provides a solution to avoid another violation, these accounts will resume normal operation.</p><p>However, according to 36Kr from many sellers, the probability of successful appeal from Amazon is almost 0. At the same time, it is not very accurate to grasp the variety of changes in Amazon's rules. The main problem lies in the inefficiency of email communication.</p><p>Where's the way out?</p><p>For the whole cross-border e-commerce industry, it is still in a very early stage, and it is not even an industry. 36 Krypton found from a number of investment institutions visited that investors' optimistic judgment on the industry has not been affected in any way. What has been looking for is which enterprises can cross the cycle and have the consciousness and ability to come out of a new brand.</p><p>It is difficult to build a brand. What is needed is the comprehensive ability of enterprises to meet user needs, marketing methods, supply chain impact, etc. However, it is precisely because of the high threshold that it is even more scarce. Especially<a href=\"https://laohu8.com/S/300866\">Anker Innovation</a>And the success of Shein, which makes all investors very excited, look forward to finding the next such investment target.</p><p>In the future, the first thing the industry has to reconstruct is consciousness. The past distribution path is already a dead end, and the earlier it is, the more opportunities for transformation. A big seller in the industry lost a lot of first-mover dividends because he transformed his brand two years late. When he turned back and chased after him, he realized that the probability of success was extremely small. This is also the dividend path dependence left by the early stage of the industry. Because it was easier to make fast money in the past, they were reluctant to take a difficult and correct road with accumulation.</p><p>Then, find a subdivided vertical category to cultivate deeply. If it was a rough operation before, it must be refined now. Moreover, the threshold for category requirements has become higher. Why won't the previous simple rough mode work? Because consumers are awakening and have higher requirements for product quality and service, third parties like Amazon are gradually changing their strategies to encourage brand determination to be stronger. In such an environment, without the ability to refine, there may be no profit at all in the future.</p><p>Secondly, introduce some high-quality technical talents. The epidemic is the beginning of cross-border e-commerce out of the circle, and the title is accelerated. The direct impact is that more entrepreneurs enter across the border. Industries that seem to have no threshold are actually not low, especially in the practical stage. Perhaps some pure business logic can still be passed through until now, but creating competitive barriers requires more high-quality talents to join. Especially on the technical side, as far as the current industry practitioners are concerned, there is still a lot of room for improvement.</p><p>Finally, be good at using the power of capital. All along, because cross-border e-commerce is easy to make money, there is even no shortage of cash flow. Therefore, the desire for capital is relatively general, but if you sort out Anker Innovation and Shein, it will be obvious that the reason why these two companies have such a high valuation in the capital market, especially the high growth of the latter. It is inseparable from the help of capital. Capital is not good or bad, and the core lies in how to use it. Providing capital is only the underlying role, and greater empowerment will greatly help to become stronger and bigger.</p><p>From the perspective of any industry development cycle, the large-scale entry of capital is after the penetration rate has increased to a certain extent. At this time, some talents such as big factories, high education and rich industry experience are invested. Driven by technology, these innovative forces began to take a new path and reconstruct the business pattern. In other words, you don't accept investment and capital invests in your competitors.</p>\n<div class=\"bt-text\">\n\n\n<p> source：<a href=\"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202109221714177c3f3de2&s=b\">36氪</a></p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/9dffe01d72356572c6df4fa7726ddb0c","relate_stocks":{"AMZN":"亚马逊"},"source_url":"http://gu.qq.com/resources/shy/news/detail-v2/index.html#/?id=nesSN202109221714177c3f3de2&s=b","is_english":false,"share_image_url":"https://static.laohu8.com/9a95c1376e76363c1401fee7d3717173","article_id":"2169565516","content_text":"因为封号冻资事件，亚马逊今年一直被热搜。\n所以导致整个跨境电商笼罩的信息一直是今天这个大卖关了，明天那个大卖裁员了。特别是在一些自媒体流量吞噬下，行业似乎就没有一些积极事件发生。\n这里面比较关键的信息不对称是亚马逊到底封了多少个中国店铺？开始最主流传言是有5万家，由于亚马逊官方一直保持不对外发声，这个数字就逐渐默认成事实。\n即使在亚马逊上周五召开的全球首个综合性卖家培训中心落户杭州活动上，对于封店数量还是三缄其口，不计划对外披露，希望冷处理就过去。\n被邀请的媒体们肯定不会放过这个难得机会，所以活生生把一场落户活动问答环节变成了封店潮发布会。以至于亚马逊全球副总裁、亚马逊全球开店亚太区执行总裁 Cindy Tai都坦言，“今天这个发布本来是针对培训中心落户杭州的，只是没想到这么多媒体关注账号被封之事，所以就一起进行了首次回应。”\n据她透露，过去5个月，亚马逊关闭了约600个中国卖家品牌销售权限，涉及品牌数约3000个，包括一些大卖家。“这些卖家都有多次、反复、严重滥用评论行为，以及有许多其他违规行为。在过去一段时间，亚马逊对这些卖家进行过多次警告，他们也有多次申诉机会，甚至恢复了一些停用过的账号。但是这些卖家持续违规，所以这次决定终止和这些卖家的合作关系。”\n翻译过来就是，被封掉的卖家违规行为严重，亚马逊给了很多次机会仍屡教不改，所以被迫走了最后一步。\n“亚马逊只是公布了前五个月的数据，6、7、8月数据没有披露，这三个月也有很多店铺受影响。所以最终的数字一定比这个大很多。”一位资深从业者告诉36氪。\n对于严重依赖亚马逊的卖家来说，损失确实不小。比如曾经的华南城四少之有棵树，被封掉店铺340个，冻结资金1.3亿元。直接导致员工离职一半，收入下滑一半，净亏损超7亿元；另一大卖通拓科技被禁售关闭店铺数 54 个， 冻结资金 4143 万元。\n那么，亚马逊这次封店力度和强度为什么这么大？背后让我们真正反思的又有哪些？\n算法的进步\n事实上，行业内都知道一个潜规则。就是亚马逊每年固定有两次账号扫荡，一次是在7月左右的会员日（Prime Day），亚马逊为了自己这个大型促销活动通常会提前2个月关掉一些账号，也就是在4月底到5月初；另一次是在圣诞节的前2个月，即10月中旬或者11月初。\n因为看到了这个巨大的漏洞，所以很多卖家都有自己的应对之策。比如一个品牌授权50个店铺，封掉几个完全没有影响。“其实只要关掉的店铺没有超过一定比例，连老板都不需要知道这事。”某卖家告诉36氪。\n经过去年行业大爆炸后，今年所有卖家也都更乐观，完全没看到这件事背后真正的风险。其实早在3年前，亚马逊就在利用算法机器人在检测商品评论是真实还是虚假。曾经也封掉了很多自营Listing（产品页面），比如深圳某卖家的Listing就从1.7万缩减至1000，朋友圈也是一篇哀嚎。好在后来因为有误差太严重，亚马逊就没再大规模继续封停。\n但技术一直在进步，直到今年开始成熟许多。加上亚马逊改变了规则，不再是按照店铺逻辑评判，而是按照品牌维度逻辑。只要算法识别出有问题的品牌，也超过了一定比例，旗下店铺一律关停。所以才有了大量店铺群被封。\n36氪也注意到，亚马逊新任CEO安迪·杰西(Andy Jassy)，此前是亚马逊云计算CEO，不是电商负责人。这也证明亚马逊作为一家数据驱动公司，对于平台数据真实度的保护更加重视。\n“相信这次事件以后，很多卖家会开始采用白帽玩法。以及亚马逊也鼓励站内邀评，通过站内信邀请用户评价。如果产品和体验不好，邀回来的肯定是差评。所以这件事对整个生态是很有好处的。如果一直劣币驱逐良币，良币就不可能做起来。”上述卖家说道。\n供应商逻辑\n除了算法的持续迭代外，亚马逊对卖家定义也和国内天猫淘宝不一样。不一样地方在于，亚马逊是把平台上所有卖家视为供应商，自己相当于一个零售商。\n也就是说，卖家在亚马逊上的所有数字资产都归属亚马逊。即使你是苹果，是耐克，都是亚马逊的供应商。\n“可以这么理解，天猫类似于一个购物中心，亚马逊是一个百货公司。所以这就埋下了亚马逊的一些处罚机制伏笔。如果是天猫卖家违规，平台会动你的货和资金吗？亚马逊谈的更多是销售规模，天猫谈才GMV。”智云天工销售副总周骏告诉36氪。\n对于亚马逊就不同，一旦收了顾客的货款，就要为用户负责。只要发现卖家可能对顾客造成伤害，就会先把资金冻结，以应付后续会不会因为起诉承担连带责任。\n这也得到了Cindy Tai的证实，亚马逊资金冻结周期为90天，主要用来承担卖家过去客户的退款退货赔偿费用，以及其他未支付费用。90天后，如果没有违规会可申请取回。\n所以，在亚马逊开店，其实不是自己在卖货，而是卖家给平台供货，平台再帮忙卖货。亚马逊不是行政和司法机构，没权力冻结资金和货。但就因为亚马逊的供应商逻辑，因此只要违反了供应商规则，就得从名录里踢掉。\n“从价值观来讲，刷单是严重的撒谎、造假行为，我们都是深恶痛绝的。但也不能一味污名化中国卖家，行业里大家都在没日没夜干活。亚马逊做的不好的地方是，起码有个听证会，给卖家一个自辨环节。对于卖家来说，该争取的合法权益必须争取。如果这次默默过去，未来估计是噩梦。”周骏说。\n在申诉流程上，亚马逊全球副总裁、亚马逊全球开店亚太区执行总裁 Cindy Tai表示，如果一个卖家账号因为违规行为被停用，亚马逊会为卖家提供申诉机会；只要这个卖家证明亚马逊判断有误，或者证明违规行为只是暂时或无意失误，并提供如何避免再次违规方案，这些账号会恢复正常运营。\n但据36氪从多位卖家处了解到，从亚马逊申诉成功的概率几乎为0。同时，对于亚马逊的规则变化多样把握也不是很准，其中主要问题在于邮件沟通这种方式效率很低。\n出路在哪儿？\n对于整个跨境电商行业来说，目前还处于很早期阶段，甚至都还谈不上一个行业。36氪从走访的多家投资机构中发现，投资人对行业的乐观判断未受到任何影响。一直在寻找的是，哪些企业能穿越周期，有意识有能力走出一个新品牌。\n塑造一个品牌很难，需要的是企业对用户需求、营销玩法、供应链影响等综合能力，但正因为门槛高，所以才更加稀缺。特别是安克创新和Shein的成功，让所有投资人都很兴奋，期待找到下一个这样的投资标的。\n在未来出路上，行业首先要重构的是意识。过往的铺货路径已经是死胡同，越早转型机会越多。行业某大卖家就是因为晚了2年转型品牌，才丧失了很多先发红利。等回头猛追才发现，成功概率微乎其微。这个也是行业早期阶段留下的红利路径依赖，因为过去方式赚快钱更容易，所以都不太愿意走一条有积累、难而正确的路。\n然后，找到一个细分的垂直品类深耕。如果说之前是粗旷式的经营，现在肯定要进行精细化运营。而且对品类要求门槛变得更高。为什么之前那种简单粗暴模式行不通？就是因为消费者在觉醒，对产品质量、服务要求变得更高，亚马逊这样的第三方也在逐渐改变策略，鼓励品牌决心更强。如此环境之下，没有精细化的能力，或许未来压根没有利润可言。\n其次，引入一些高素质的技术人才。疫情是跨境电商出圈的开始，封号是加速，带来的直接影响就是更多创业者跨界进入。看似没有门槛的行业，实则门槛不低，特别到了实操阶段。也许直到现在还能走通一部分纯生意逻辑，但是打造竞争壁垒这件事优势就需要更多高素质人才加入。特别是偏技术端，就目前行业从业者来看，能提升的空间还是很大。\n最后，善于借助资本力量。一直以来，因为跨境电商赚钱容易，甚至不缺现金流。所以对资本的欲望也相对一般，但如果梳理安克创新和Shein就会明显发现，这两家公司之所以在资本市场有这么高的估值，特别是后者的高增长。离不开资本的助力，资本无好坏，核心在于如何去利用。资本提供资金只是底层作用，更大的赋能对做强做大帮助大。\n从任何一个行业发展周期来看，资本大规模进入是在渗透率有一定提高之后。这时候投资的是一些大厂、高学历、丰富行业经验等人才，这些创新力量在技术推动下开始走出新的路子，以及重构商业格局。换句话说，你不接受投资，资本就会投资你的竞争对手。","news_type":1,"symbols_score_info":{"AMZN":0.9}},"isVote":1,"tweetType":1,"viewCount":4909,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":860401476,"gmtCreate":1632193536795,"gmtModify":1676530722519,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Good","listText":"Good","text":"Good","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":1,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/860401476","repostId":"2169684551","repostType":4,"isVote":1,"tweetType":1,"viewCount":4592,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":881911500,"gmtCreate":1631284760100,"gmtModify":1676530520376,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Good","listText":"Good","text":"Good","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":4,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/881911500","repostId":"1144889232","repostType":4,"isVote":1,"tweetType":1,"viewCount":3702,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":883091778,"gmtCreate":1631186053494,"gmtModify":1676530490511,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Good","listText":"Good","text":"Good","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":3,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/883091778","repostId":"2166195813","repostType":4,"repost":{"id":"2166195813","kind":"highlight","weMediaInfo":{"introduction":"新型财经科技信息服务提供商，专注TMT。技术改变商业，商业改变世界，我们纪录这个过程，并聚集这些改变世界的人。","home_visible":1,"media_name":"TMTPost","id":"1065587721","head_image":"https://static.tigerbbs.com/72948639b39fd795a430fcaa2772851c"},"pubTimestamp":1631170956,"share":"https://ttm.financial/m/news/2166195813?lang=en_US&edition=fundamental","pubTime":"2021-09-09 15:02","market":"sh","language":"zh","title":"When was \"Nvidia\" born in China?","url":"https://stock-news.laohu8.com/highlight/detail?id=2166195813","media":"TMTPost","summary":"这一事件引发了人们对虚拟现实、元宇宙、AI换脸等技术和概念的激烈讨论，同时也让“英伟达”这家美国芯片霸主从半导体行业“出圈”，走入了大众视野。另外，作为全球芯片销量大国，中国却没有出现一家“英伟达”这样的芯片巨头，大市场并没有产生与之匹配的大公司。","content":"<p><img src=\"https://static.tigerbbs.com/01e0bbbfdbdc6f8f2f7ce512582f7c3e\" tg-width=\"1200\" tg-height=\"750\" referrerpolicy=\"no-referrer\" width=\"100%\" height=\"auto\"><a href=\"https://laohu8.com/S/NVDA\">NVIDIA</a>CEO Huang Renxun (Image source: Nvidia official website)</p><p>\"The product I am about to show, which integrates new GPU accelerated computing capabilities and has Mellanox high-performance network, is the world's first CPU processor designed for TB-level data center accelerated computing. Its secret code name is Grace.\"</p><p>This is a quote from NVIDIA CEO Jensen Huang during his speech at the GTC Summit in April 2021. However, unexpectedly, it was not until August 12th that NVIDIA exposed itself that people knew that the content of this speech of less than 100 words and 14 seconds was not Huang Renxun himself, but a synthetic \"digital stand-in\", that is, the \"virtual Huang Renxun\" image formed by NVIDIA GPU processor and Omniverse software platform.</p><p>This incident triggered heated discussions on technologies and concepts such as virtual reality, metaverse and AI face change. At the same time, it also made \"Nvidia\", the American chip overlord, \"out of the circle\" from the semiconductor industry and into the public eye.</p><p>Since its establishment in 1993, under the leadership of Huang Renxun, Nvidia has successfully created and led the category of GPU (graphics processor) chips, covering the whole PC device GPU to server GPU market. Over the past five years,<b>Nvidia grows from $31 billion to $505 billion in market capitalization, ranking as the world's seventh-largest semiconductor supplier, is an artificial<a href=\"https://laohu8.com/S/5RE.SI\">smart</a>A hot star enterprise in the field of (AI) chips.</b></p><p>Meanwhile, after Nvidia's market cap exceeds<a href=\"https://laohu8.com/S/INTC\">Intel</a>After that, the domestic semiconductor market saw greater market opportunities in GPU and AI chip tracks.<a href=\"https://laohu8.com/S/300474\">Jing Jiawei</a>, Tianshu Zhixin, Denglin Technology, Biren Technology, Suiyuan Technology,<a href=\"https://laohu8.com/S/688256\">Cambrian</a>Enterprises such as Muxi Integrated Circuit and Muxi Integrated Circuit gather in the general-purpose processor track.</p><p>In the past two years, the domestic semiconductor industry has undergone transformation. Thanks to Huawei HiSilicon,<a href=\"https://laohu8.com/S/SMI\">SMIC</a>Affected by the US \"entity list\" and the chain reaction caused by the global chip shortage, it is more and more difficult for Chinese enterprises to purchase American semiconductor products, and the demand for \"domestic substitution\" is stronger and stronger.</p><p>On June 29, 2020, due to the upgrade of the U.S. Export Administration Regulations and the constant adjustment of enterprise members in the entity list, Intel \"temporarily suspended\" the supply of chips to Inspur Group.</p><p>Although on July 3rd, Inspur announced that Intel had resumed its supply, one was the world's largest PC and cloud service x86 architecture chip supplier, and the other was China's largest server manufacturer. The temporary supply cutoff crisis between the two companies raised a wake-up for the whole Chinese cloud computing and semiconductor industry:<b>With the increasingly strong demand for computing power, China needs to mass-produce fully independent and controllable domestic GPU and server chip products.</b>This highlights that \"domestic substitution\" is becoming the biggest driving force for the development of China's semiconductor industry.</p><p><b>In addition, as a big chip selling country in the world, there is no chip giant like \"Nvidia\" in China, and the big market has not produced a matching big company.</b>According to the statistics of IC Insight, the global semiconductor market size in 2020 was US$395.7 billion, of which the market size in mainland China was US$43.4 billion, making it the largest market in the world, accounting for 36.24% of the global proportion. However, semiconductor companies headquartered in mainland China had a total output value of just $8.3 billion in 2020, representing only 5.9% of the market size.</p><p><b>Who could cut off a corner of a huge cake? When will the semiconductor industry create \"China's Nvidia\"?</b></p><p><h2>Missing out on the Golden Age</h2>GPU graphics processor, also known as display chip and vision processor, was first proposed by Nvidia in 1999. It is a microprocessor specialized in running drawing operations on personal computers, workstations, game consoles and mobile devices (smart phones, tablets, VR devices).</p><p>As the parallel computing advantages of GPU are gradually tapped, the application field of GPU expands from graphics processing to high-performance computing, and gradually becomes the most mature and widely used general-purpose chip in Al computing. In June 2020, Nvidia launched the A100 Tensor Core GPU based on Ampere architecture, becoming the world's most powerful AI chip.</p><p>From the perspective of application terminals, GPUs can be divided into PC GPUs, server GPUs and mobile GPUs, corresponding to three architectures, namely, independent graphics cards composed of special circuit boards and components, shared integrated graphics cards, and mobile GPUs packaged together with other chips or modules into highly integrated SoCs-which are applied to multiple application scenarios including mobile phones, automotive electronics and AI.</p><p><b>since<a href=\"https://laohu8.com/S/AMD\">AMD</a>After the acquisition of ATI, a Canadian GPU manufacturer, in 2006, at present, in the field of PC and server GPUs, the global GPU market is monopolized by Intel, AMD and Nvidia, the \"Big Three American Chips\". Intel has obvious advantages in the integrated GPU market, while Nvidia and AMD are separated in the independent GPU market.</b></p><p>According to the data of Jon Peddie Research, a Research organization, in the first quarter of 2021, in the global PC-side GPU market, Intel (Intel) topped the list with a market share of 68%, followed by AMD and Nvidia with 17% and 14% respectively, and the total share of the three companies was close to 100%; In the global field of independent GPUs, Nvidia is the market leader of data center GPUs, accounting for 81% of the market share, and has a leading advantage, while AMD ranks second with 19%.</p><p>In 2019 alone, with products such as the V100 series, Nvidia occupied 90% of China's AI training chip market, firmly grasping China's huge AI chip sales market.</p><p>One of the core reasons why Nvidia can continue to be a \"chip overlord\" lies in its \"light design mode\". Nvidia does not do chip manufacturing and packaging, leaving it to<a href=\"https://laohu8.com/S/TSM\">TSMC</a>OEM is completed, and it enjoys the dividend of advanced process technology such as 7nm. Between 2016 and 2021, Nvidia's revenue grew 233% and operating profit doubled to $4.5 billion, according to the earnings report. In the three months to May, sales jumped 84% year-on-year, while gross profit margin reached 64%.</p><p><b>In fact, China entered the field of GPU chip design very early, but the results were not satisfactory.</b></p><p><b>Since the 1970s, China began to introduce semiconductor and integrated circuit technologies and production lines. However, as a result, we have fallen into an a vicious circle of \"generation-to-generation introduction and generation-to-generation backwardness\". Coupled with the adverse impact of the \"Hanxin No.1\" fake chip incident on the society, China's \"autonomous processors\" have suffered serious setbacks, and China's active promotion of WTO globalization and other factors, thus missing the golden period of global semiconductor industry development, and downstream enterprises can only \"buy rather than make\".</b></p><p>By 2000, marked by the promulgation of the national \"Document No. 18\", China Semiconductor gradually formed an industrial organization form of \"separation of three industries\" in design, manufacturing, packaging and testing, and introduced \"<a href=\"https://laohu8.com/S/688981\">SMIC</a>\"A group of chip manufacturing (Foundry) enterprises represented by China have been built and put into production, and their technical level has been rapidly improved.</p><p>At present, Jingjiawei, Tianshu Zhixin, Denglin Technology, Biren Technology, Suiyuan Technology, Cambrian, Muxi IC and other enterprises are all gathered in the general-purpose processor track.</p><p>In 2014, Jingjiawei (300474.SH), which started as a military aircraft graphical display control module, successfully developed the military GPU chip JM5400, and then successfully developed the second generation GPU chip JM7200 with 28nm process technology in 2018. Jingjiawei has moved from military customization to general-purpose GPU, becoming one of the few companies in the world and the only company in China to realize commercial mass production of independent GPU.</p><p>In addition to Jingjiawei, in March 2021, Tianshu Zhixin released the first GPGPU (general-purpose graphics processor) manufactured by 7nm process in China, that is, 30% of the graphics rendering part of the traditional GPU was removed, and it was only born for handling artificial intelligence (AI) applications; In June this year, Suiyuan Technology released China's largest AI computing chip \"Sisi 2.0\" AI chip, \"Yunsui T20\" training accelerator card based on Sisi 2.0 and \"Yunsui T21\" training OAM module.</p><p>However, it is worth noting that the performance of JM7200 chip developed by Jingjiawei is only equivalent to that of Nvidia GTX 640 in 2012, which is difficult to meet the application needs of enterprise customers. Even Suiyuan Technology's \"Suisi 2.0\" AI chip only tied with Nvidia's A100, and two of the six projects tested by Benchmark greatly surpassed the performance of Nvidia A100. (For details, see the previous article of Titanium Media App: \"Suiyuan Technology releases China's largest AI computing chip to accelerate the implementation of three major business directions\")</p><p><img src=\"https://static.tigerbbs.com/15cb73d2787595547a9ed01926f4dfbe\" tg-width=\"570\" tg-height=\"1926\" referrerpolicy=\"no-referrer\" width=\"100%\" height=\"auto\"></p><p><b>The reason behind this is mainly that China's semiconductor industry started late, the technical threshold of chips is high, the cost elasticity is high, and the industry is highly concentrated, which makes the overall R&D investment, technology and talents of Chinese GPU chip enterprises lag behind foreign countries. Therefore, there is still a gap with chip giants in product performance and technology, and it is still difficult for downstream enterprises to break away from the position of \"American Chip Big Three\".</b></p><p>Take R&D investment as an example. In the ten years from 2011 to 2020, Jingjiawei's R&D investment totaled RMB 627 million, while Nvidia's R&D investment in 2020 reached USD 3.924 billion, or about RMB 25.323 billion. In the ten years, Nvidia's total investment exceeded RMB 120 billion, with a difference of more than 190 times.</p><p>In terms of talent, as of the first half of 2021, Nvidia had a whopping 18,975 employees and Jingjiawei had a total of 1,174 employees, far lower than AMD's 2,000 employees in the Shanghai R&D center.</p><p>\"AI chip and GPU chip markets are quite special. Unlike traditional special-purpose processors, the technology is very complicated. It requires a lot of data and needs to be combined with specific algorithms before it can be put into the market.\"<a href=\"https://laohu8.com/S/SNPS\">Synopsys</a>Xie Zhonghui, deputy general manager of China, said in an exclusive interview with Titanium Media App in April this year that if an enterprise wants to make the first AI chip solid, it usually takes more than two or three years.</p><p>In his view, chip semiconductor itself is an industry with large investment, long cycle and slow results. Complete localization of technology requires long-term and sustained accumulation of capital, talents and technology, and it is difficult to deal with it with the Internet thinking of \"spending money and seeing returns\".</p><p><b>In addition, the software and hardware ecology combined with CUDA technology is another important reason for the big gap between domestic chip companies and Nvidia.</b></p><p>In 2006, Nvidia released the parallel computing platform CUDA, which contains a series of development tools. Only by installing and using this platform can complex parallel computing be performed. Anyone who owns a laptop equipped with Nvidia GPU can use CUDA to perform scientific and convenient programming calculations, such as deep learning, AI algorithms, etc., and develop related software. Over the past decade or so, NVIDIA has persistently promoted CUDA, so that more government and enterprise-level software has been developed based on this platform, and NVIDIA's self-developed GPU hardware has been combined with CUDA software to efficiently realize the application landing.</p><p>In contrast, at present, there is no series of platforms similar to CUDA and NVIDIA hardware deep binding in China, and the gap in technical barriers is very obvious. Most domestic GPU manufacturers have adopted the strategy of compatibility with CUDA open source framework, such as Tianshu Zhixin and Denglin, and are ready to cultivate their own software ecosystem on this basis.</p><p>\"In the short term, it is easier to develop domestic GPU compatible with CUDA. After all, writing operators is a manpower-intensive industry. If users migrate, they need to migrate 100%, and the whole set of codes have to run on your chip. If the amount of code is small, there are not so many operators needed, and the difficulty will be relatively low. But in the long run, it is still necessary to get rid of compatibility ideas and develop its own core technologies.\" According to chip industry insiders, choosing compatibility is mainly to ensure that the existing software is still available. In the future, it will continue to improve its own platform to make it more matched with its own chips, thus attracting developers to migrate.</p><p>However, some enterprises choose to be incompatible with CUDA ecology. For example, Suiyuan Technology, which does AI training and inference chips at the same time, has comprehensively upgraded its \"TopsRider\" software platform and a brand-new \"Yunsui Cluster\" this year, hoping to have ecological dominance.</p><p><b>To sum up, these domestic chip companies that benchmark Nvidia are still in the initial stage of development. The industrialization and marketization ability of AI chip technology is weak, and it has not actually been used on a large scale. There is still a long way to go before surpassing or replacing \"China Nvidia\".</b></p><p>Yan Guihai, CEO of Zhongke Yushu, said in an interview with Titanium Media App that although China's demand side is still the largest single market in the world, and its growth rate is among the best, the \"demand side\" is still very strong, but there is still a big gap in design and manufacturing of high-end chips, and the \"supply side\" is not strong enough. He pointed out that the advantages and disadvantages of the supply side depend not only on one enterprise, but on the capacity of the whole industrial chain. In the short term, it is still unrealistic to build such a large-scale and comprehensive leading enterprise.</p><p>Founded in 2018, Zhongke Yushu is a developer of dedicated computing architecture. It was incubated from the State Key Laboratory of Computer Architecture of Institute of Computing Technology, Chinese Academy of Sciences. Now the company's valuation has exceeded 1 billion yuan. On July 27th this year, Zhongke Yushu completed hundreds of millions of yuan in Series A financing, led by Huatai Innovation, followed by Lingjun Investment and the old shareholder Guoxin Sichuang.</p><p>Huang Liming, a partner at Hillhouse and head of software and hard technology at Hillhouse Ventures, said in an exclusive interview with Titanium Media App that although the GPU market is promising, China's<a href=\"https://laohu8.com/S/V03.SI\">Startups</a>It is difficult to make \"Nvidia\" directly. In addition to technical difficulty, it has to be done in combination with strong applications-involving software system software ecology, which is extremely demanding for startups.</p><p>Hillhouse launched Hillhouse Venture Capital, an independent VC brand, in February 2020. Since then, it has invested in the field of chip semiconductors, including semiconductor IP company Xinyaohui, EDA manufacturer Xinhuazhang, GPU platform Biren Technology, DPU company Nebula Zhilian, plus Tianke Heda in silicon carbide, Minxin Semiconductor in optical chip field, and Xingsi Semiconductor in mobile phone baseband.</p><p>Huang Liming emphasized that there will not be many companies that can run out in this direction, whether overseas or domestic, in the end.</p><p><h2>The wind has arrived</h2>\"Let's not struggle with how to replace Nvidia now. The road is step by step. I think first of all, China has to have domestic AI chips, general-purpose GPUs, FPGAs and other underlying computing power. As long as there is domestic market demand, we must have many opportunities.\" Zhang Gaonan, managing partner of Huaying Capital, told Titanium Media App that the domestic semiconductor industry has arrived, and China's entry into the GPU market is \"the right time, the right place and the right people\". In particular, semiconductors and the next generation AI technology are areas that China must break through.</p><p>Zhang Gaonan gave an example,<a href=\"https://laohu8.com/S/688111\">Kingsoft Office</a>The product is inferior to<a href=\"https://laohu8.com/S/MSFT\">Microsoft</a>Office suite, but the market gives a market value of more than 110 billion yuan. One of the important reasons behind it is that China must have domestic Office, and the same reason applies to the domestic GPU market.</p><p>Huaying Capital is one of the earliest private equity funds in China To deploy mobile Internet and cultural industry. In recent years, To B field has also become the key investment area of Huaying Capital. At present, Huaying Capital has invested more than 30 projects in the To B field, with a total investment of over 700 million yuan, which are built by Zhang Gaonan, a technology investor. In the fields of data center and underlying computing power, Huaying Capital has invested in projects such as Biren Technology, Tianyun Big Data and Zhongke Haiwei.</p><p>In fact, as a general platform spanning visual computing and AI computing, GPU has a huge market space.<b>According to<a href=\"https://laohu8.com/S/601555\">Soochow Securities</a>It is estimated that by 2027, the market space of domestic substitution in GPU field will exceed 34.1 billion USD. In addition to the existing game market, there is room for further development in industrial, medical, military aerospace and other directions.</b></p><p>In March this year, \"Biren Technology\", a general intelligent chip designer co-founded by Zhang Wen, the former president of Shangtang Technology, completed the B round of financing. Since its establishment in September 2019, the company's total financing amount has exceeded RMB 4.7 billion, and its investors include Hillhouse Venture Capital, Huaying Capital,<a href=\"https://laohu8.com/S/601318\">Ping An of China</a>China Merchants Capital, BAI Capital, Guosheng Group National Reform Fund, etc., with a valuation of over 10 billion yuan, making it one of the \"unicorn\" enterprises with the fastest momentum in the growing semiconductor industry.</p><p><b>In addition to Wall Ren, Mu Xi IC, Moore Thread and other enterprises in the GPU field have also completed financing.</b></p><p>On August 25th, Muxi Integrated Circuit, a GPU manufacturer, announced the completion of RMB 1 billion Series A financing. The founders Chen Weiliang and Yang Jian are all from American chip giant AMD, and the investors include China State-owned Enterprise Structural Adjustment Fund Co., Ltd.,<a href=\"https://laohu8.com/S/00810\">China Internet Investment</a>Fund, Matrix Partners China, Holy Capital, Sequoia China, Lightspeed China, Guochuang Zhongding, Smart Internet Industry Fund, Shanghai Science and Technology Fund, Lenovo Venture Capital, etc.; Moore Thread, established in 2020, claimed to have completed two rounds of multi-billion yuan financing within 100 days. The team members are mainly from Nvidia, and the investors include Shenzhen Venture Capital and Sequoia Capital<a href=\"https://laohu8.com/S/CHN\">China Fund</a>China Merchants Venture Capital, ByteDance, Pony Zhixing, Wuyuan Capital, etc.</p><p><b>However, an interesting phenomenon is that the above three start-ups, namely Bi Ren, Mu Xi and Moore Thread, are veritable \"PPT financing core-making\", and none of them completed the taping of the first chip (trial production of assembly line) during financing.</b></p><p>Why are market investors willing to open their wallets to this? In an interview with Titanium Media App, several investors said that these projects can get a lot of capital support, all because of the early investment, mainly depending on the team and the track: the AI chip track has arrived, and the senior executive team is also from the \"American Chip Big Three\".</p><p>\"I think we need to give these enterprises opportunities and patience. It is impossible for 500 people to write PPT. Manufacturing chips is a matter of 5 to 10 years. The reason why we are willing to invest is not speculation or fooling. I think the risk of investing in the semiconductor track itself is high, and it is necessary to make long-term plans and have strong enough risk-taking capacity, which is completely different from the model innovation of investing in the Internet.\" The above-mentioned investors told Titanium Media App.</p><p>However, some investors in the semiconductor industry pointed out that the above-mentioned investment projects essentially hope to raise the market value and have a higher rate of return. Especially in the \"chip fever\" environment, venture capital institutions need to constantly look for these GPU and AI chip enterprise targets in the middle and early stages, hoping to gamble on a higher return.</p><p>In addition, in this wave of GPU entrepreneurship, the founding team teached the \"American Chip Big Three\". For example, Zheng Jinshan, chief scientist of Tianshu Zhixin, was the chief technical expert of AMD; Muxi's founding team is mainly from AMD. CEO Chen Weiliang once served as senior director of graphics research and development in AMD, and CTO Yang Jian once served as AMD Fellow (academician); Li Xinrong, the newly appointed co-CEO of Biren Technology, was once the global vice president of AMD, and Chen Wenzhong, senior vice president of Biren Technology, also served in AMD.</p><p>In this regard, Zhang Gaonan said that AMD is the top two chip giants in the field of GPU. The core research and development of GPU is all in Shanghai, while the research and development of graphics rendering is in the United States. Enterprises can go to the executives of AMD and Nvidia to communicate, and the final choice is definitely the leader in the semiconductor industry.</p><p>Yan Guihai believes that in segmenting emerging tracks, with the application \"potential energy\" on the demand side, Chinese chip enterprises have concentrated their superior forces, based on serving local enterprises, and highlighted the agility of development, which is an opportunity to surpass chip giants such as \"Nvidia\" in product definition and scheme iteration cycle.<b>He predicts that a number of domestic GPU and DPU enterprises with leading technology will appear within 10 years.</b></p><p>\"Among the five links of the chip industry: design, manufacturing, packaging and testing, materials and EDA, the most relevant to application is design, and our biggest advantage lies in application, so we have a great opportunity to make a breakthrough in the\" design \"link, and then gradually expand the territory of victory from point to point. The so-called \"corner overtaking\" is still a catch-up strategy. It is more likely to occupy a new strategic commanding height by cutting into a new future-oriented track and accelerating at full speed. I hope that within 10 years, a number of enterprises with leading technology, solid products and strategic awareness will emerge. \" Yan Guihai told Titanium Media App.</p><p>Zhang Wen, founder of Biren Technology, said that the capability requirements for chip companies have been upgraded from product level to system level and ecological level. In less than five years, China has caught up with or even led the international standard in the field of AI chip design. He emphasized that going beyond Nvidia requires redefining a product, as well as redefining a market.</p><p><h2>Ten billion DPU chip market \"explodes\"</h2>In Huang Renxun's view, the data processing unit (DPU) responsible for transmitting and processing data in the data center is forming \"the three pillars of future computing\" together with CPU and GPU. When Chinese chip companies developed GPU, Nvidia set its sights on the two new markets of CPU and DPU.</p><p>In September 2020, Nvidia announced that it planned to acquire British chip designer Arm from a Japanese software group for US$40 billion, and it is expected to write the largest merger and acquisition case in the semiconductor industry. But the deal is controversial and awaits approval from governments including the European Union, Britain, the United States and China. But in April 2021, Nvidia announced its entry into the data center CPU market and released the Grace CPU processor, which is what Huang Renxun said at the beginning of this article. (For details, see the previous article of Titanium Media App: \"The British government intervened, and Nvidia's $40 billion acquisition of Arm transaction changed\")</p><p>In addition to CPU and GPU, Nvidia is also laying out DPU. In 2019, Nvidia announced the acquisition of Mellanox, an Israeli network chipmaker, for $6.9 billion in all-cash, and eventually won it. In this largest acquisition in Nvidia's history, what Huang Renxun values most is Meros's unique ability in data center technology and other aspects.<b>In October 2020, NVIDIA debuted its DPU — the NVIDIA BlueField series of data processors.</b></p><p>Fundamentally, on the one hand, DPU is more flexible and safe, and more importantly, DPU can free the computing power of CPU, release the load of server, significantly reduce the comprehensive cost with low power consumption, and even improve the performance of AI and machine learning applications.</p><p>According to IDC statistics, the global demand for computing power will double every 3.5 months, far exceeding the current growth rate of computing power. Driven by this, the global computing, storage and network infrastructure is also undergoing fundamental changes: some workloads with excessive data occupy too much CPU resources, and various \"X\" PU chips that cooperate with them have emerged. Besides GPU, FPGA and other chips, DPU is the next \"X\" PU.</p><p>People in the industry have made a vivid metaphor for this. The network is like building roads. In the past, the roads were not wide enough in the 1G and 10G era, and there were more and more cars. In order to balance the pressure, it was necessary to increase traffic lights and invest more traffic police to coordinate resources, which has improved the original efficiency a lot, but it is still not enough. First, we must expand the road, which is the increase of bandwidth. However, the road has changed from 2 to 4. Just relying on traffic lights and limited traffic police will still be blocked. However, we can't increase the traffic police indefinitely, which requires the road to be more intelligent and help solve the congestion.</p><p>Zhang Gaonan pointed out that a large amount of network management is in the CPU, which occupies the container capacity, while the DPU intelligently provides the space capacity of the server, and a large amount of virtualization space can increase the computing power demand.</p><p><b>With 2020, DPU's reputation exceeded the infrastructure processor (IPU) and SmartNIC launched by rival Intel, and it also made every enterprise eyeing data center business get a piece of this field. DPU has become a new track for major chip giants and start-up companies to compete for research and development. Domestic DPU is now almost in a state of blooming, and Sequoia, Hillhouse Venture Capital, CDH and SoftBank China have all begun to enter the market.</b></p><p>In April this year, Tianyan Check data showed that the domestic DPU chip supplier \"Cloud Leopard Intelligence\" completed<a href=\"https://laohu8.com/S/00700\">Tencent</a>Investment, Sequoia Capital, Yaotu Capital and other joint angel round financing; At the end of May, Xinqiyuan completed hundreds of millions of yuan of Pre-A round of financing, with investors including SoftBank China and Pudong Science and Technology Group; On July 27th, DPU chip developer \"Zhongke Yushu\" completed hundreds of millions of yuan of Series A financing led by Huatai Innovation; On August 30th, Nebula Zhilian, a developer of DPU chips, announced the completion of hundreds of millions of angel round financing, led by Hillhouse Venture Capital, and CDH VGC and Walden International China Fund participated in the follow-up investment; At the beginning of September, IDG Capital threw the \"Yunmai Xinlian\" angel round financing project.</p><p>\"DPU may become the third computing chip after CPU and GPU, but from the structural point of view, DPU will be more heterogeneous and more dedicated.\" Yan Guihai said in an interview with Titanium Media App and others that the background of DPU is the demand for computing delay, data security and resource virtualization brought by the end-edge-cloud integration trend caused by the data explosion in the intelligent era. The CPU is overwhelmed with these non-business loads, and there is an urgent need for an ideal object to share these computational loads.</p><p><b>Leopard Research Institute predicts that China's DPU market is expected to reach 3.74 billion USD in 2025. The global DPU market size is expected to reach USD 13.57 billion in 2025.</b>At the same time, the report also pointed out that data circulation is the largest application market of DPU, among which bare metal services are just in demand for DPU. The application of DPU in the telecom market is mainly in edge computing scenarios, with a penetration rate of less than 5%. DPU for the field of intelligent driving is still in the exploratory stage, and it is expected that DPU will be deployed in the field of intelligent driving only in 2023.<img src=\"https://static.tigerbbs.com/7170c585dab44018726df81a32c63d62\" tg-width=\"808\" tg-height=\"500\" referrerpolicy=\"no-referrer\" width=\"100%\" height=\"auto\"></p><p>China DPU Market Size, Forecast 2020-2025, Source: Toubao Research Institute</p><p>Yan Guihai pointed out that the performance of CPU increased from 30% per year 5-10 years ago to less than 3% per year three years ago. However, the network bandwidth is still increasing by about 35% every year.</p><p>With<a href=\"https://laohu8.com/S/600918\">Zhongtai Securities</a>For example, the challenge that the company encountered at that time was that transaction statement compliance checks were too slow, and it needed to improve transaction efficiency. Therefore, Zhongke Yushu, Zhongtai Securities and Shanghai Stock Exchange Technology Co., Ltd. jointly developed a set of extremely fast risk control system solutions to speed up this process. Zhongke Yushu has successively developed a number of products and solutions such as ultra-low latency smart network card and data computing accelerator card, mainly for high bandwidth, low latency and data-intensive scenarios. The company has achieved quarterly revenue of tens of millions this year. Zhongke Yushu's next-generation DPU chip is expected to be designed by the end of 2021 and is expected to handle up to 200G of network bandwidth data.</p><p><b>However, although the DPU market is hot, the concept is newer, there are more unknowns, and the investment risk will be greater.</b></p><p>Lu Sheng, CEO of Xinqiyuan, pointed out that at present, the barriers of DPU segmentation track are still relatively high. In addition to technical barriers, there are also market barriers, which require customers to iterate continuously, especially with the continuous upgrading of open source software to adapt to customers' rapidly changing software and hardware environment. Therefore, VC (venture capital) must recognize the track and have enough patience before investing. He emphasized that investors need to constantly observe and adjust their judgments on the market. Now no one can predict the future development prospects of DPU.</p><p>Some media also believe that when Nvidia enters the newly opened battlefield of CPU and DPU, it may be good for Chinese GPU manufacturers. In particular, Nvidia still spends a lot of energy on the M&A transaction of acquiring British chip designer Arm for $40 billion, which may be an opportunity for new GPU companies to catch up.</p><p>As Zhang Gaonan said to Titanium Media App, \"Logically speaking, things with low thresholds are usually not scarce. (Chip semiconductor track) Some things are difficult and require a long time of investment. Although it is high risk, someone has to do it. This is a long-term investment that is really beneficial to the country. In fact, investment should be encouraged. Otherwise, no one will do these difficult things that require a long time of investment, and you will never get up the steps.\"</p><p>Zhang Gaonan emphasized that although venture capital must pursue returns, he believes that it is not only of strong social significance to invest in some start-ups in the semiconductor track under the condition that the whole capital is allocated reasonably, but also an important embodiment of some long-term value investment.</p>","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>When was \"Nvidia\" born in China?</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: 12.5px; color: #7E829C; margin: 0;}\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\">\nWhen was \"Nvidia\" born in China?\n</h2>\n<h4 class=\"meta\">\n<a class=\"head\" href=\"https://laohu8.com/wemedia/1065587721\">\n\n<div class=\"h-thumb\" style=\"background-image:url(https://static.tigerbbs.com/72948639b39fd795a430fcaa2772851c);background-size:cover;\"></div>\n\n<div class=\"h-content\">\n<p class=\"h-name\">TMTPost </p>\n<p class=\"h-time smaller\">2021-09-09 15:02</p>\n</div>\n</a>\n</h4>\n</header>\n<article>\n<p><img src=\"https://static.tigerbbs.com/01e0bbbfdbdc6f8f2f7ce512582f7c3e\" tg-width=\"1200\" tg-height=\"750\" referrerpolicy=\"no-referrer\" width=\"100%\" height=\"auto\"><a href=\"https://laohu8.com/S/NVDA\">NVIDIA</a>CEO Huang Renxun (Image source: Nvidia official website)</p><p>\"The product I am about to show, which integrates new GPU accelerated computing capabilities and has Mellanox high-performance network, is the world's first CPU processor designed for TB-level data center accelerated computing. Its secret code name is Grace.\"</p><p>This is a quote from NVIDIA CEO Jensen Huang during his speech at the GTC Summit in April 2021. However, unexpectedly, it was not until August 12th that NVIDIA exposed itself that people knew that the content of this speech of less than 100 words and 14 seconds was not Huang Renxun himself, but a synthetic \"digital stand-in\", that is, the \"virtual Huang Renxun\" image formed by NVIDIA GPU processor and Omniverse software platform.</p><p>This incident triggered heated discussions on technologies and concepts such as virtual reality, metaverse and AI face change. At the same time, it also made \"Nvidia\", the American chip overlord, \"out of the circle\" from the semiconductor industry and into the public eye.</p><p>Since its establishment in 1993, under the leadership of Huang Renxun, Nvidia has successfully created and led the category of GPU (graphics processor) chips, covering the whole PC device GPU to server GPU market. Over the past five years,<b>Nvidia grows from $31 billion to $505 billion in market capitalization, ranking as the world's seventh-largest semiconductor supplier, is an artificial<a href=\"https://laohu8.com/S/5RE.SI\">smart</a>A hot star enterprise in the field of (AI) chips.</b></p><p>Meanwhile, after Nvidia's market cap exceeds<a href=\"https://laohu8.com/S/INTC\">Intel</a>After that, the domestic semiconductor market saw greater market opportunities in GPU and AI chip tracks.<a href=\"https://laohu8.com/S/300474\">Jing Jiawei</a>, Tianshu Zhixin, Denglin Technology, Biren Technology, Suiyuan Technology,<a href=\"https://laohu8.com/S/688256\">Cambrian</a>Enterprises such as Muxi Integrated Circuit and Muxi Integrated Circuit gather in the general-purpose processor track.</p><p>In the past two years, the domestic semiconductor industry has undergone transformation. Thanks to Huawei HiSilicon,<a href=\"https://laohu8.com/S/SMI\">SMIC</a>Affected by the US \"entity list\" and the chain reaction caused by the global chip shortage, it is more and more difficult for Chinese enterprises to purchase American semiconductor products, and the demand for \"domestic substitution\" is stronger and stronger.</p><p>On June 29, 2020, due to the upgrade of the U.S. Export Administration Regulations and the constant adjustment of enterprise members in the entity list, Intel \"temporarily suspended\" the supply of chips to Inspur Group.</p><p>Although on July 3rd, Inspur announced that Intel had resumed its supply, one was the world's largest PC and cloud service x86 architecture chip supplier, and the other was China's largest server manufacturer. The temporary supply cutoff crisis between the two companies raised a wake-up for the whole Chinese cloud computing and semiconductor industry:<b>With the increasingly strong demand for computing power, China needs to mass-produce fully independent and controllable domestic GPU and server chip products.</b>This highlights that \"domestic substitution\" is becoming the biggest driving force for the development of China's semiconductor industry.</p><p><b>In addition, as a big chip selling country in the world, there is no chip giant like \"Nvidia\" in China, and the big market has not produced a matching big company.</b>According to the statistics of IC Insight, the global semiconductor market size in 2020 was US$395.7 billion, of which the market size in mainland China was US$43.4 billion, making it the largest market in the world, accounting for 36.24% of the global proportion. However, semiconductor companies headquartered in mainland China had a total output value of just $8.3 billion in 2020, representing only 5.9% of the market size.</p><p><b>Who could cut off a corner of a huge cake? When will the semiconductor industry create \"China's Nvidia\"?</b></p><p><h2>Missing out on the Golden Age</h2>GPU graphics processor, also known as display chip and vision processor, was first proposed by Nvidia in 1999. It is a microprocessor specialized in running drawing operations on personal computers, workstations, game consoles and mobile devices (smart phones, tablets, VR devices).</p><p>As the parallel computing advantages of GPU are gradually tapped, the application field of GPU expands from graphics processing to high-performance computing, and gradually becomes the most mature and widely used general-purpose chip in Al computing. In June 2020, Nvidia launched the A100 Tensor Core GPU based on Ampere architecture, becoming the world's most powerful AI chip.</p><p>From the perspective of application terminals, GPUs can be divided into PC GPUs, server GPUs and mobile GPUs, corresponding to three architectures, namely, independent graphics cards composed of special circuit boards and components, shared integrated graphics cards, and mobile GPUs packaged together with other chips or modules into highly integrated SoCs-which are applied to multiple application scenarios including mobile phones, automotive electronics and AI.</p><p><b>since<a href=\"https://laohu8.com/S/AMD\">AMD</a>After the acquisition of ATI, a Canadian GPU manufacturer, in 2006, at present, in the field of PC and server GPUs, the global GPU market is monopolized by Intel, AMD and Nvidia, the \"Big Three American Chips\". Intel has obvious advantages in the integrated GPU market, while Nvidia and AMD are separated in the independent GPU market.</b></p><p>According to the data of Jon Peddie Research, a Research organization, in the first quarter of 2021, in the global PC-side GPU market, Intel (Intel) topped the list with a market share of 68%, followed by AMD and Nvidia with 17% and 14% respectively, and the total share of the three companies was close to 100%; In the global field of independent GPUs, Nvidia is the market leader of data center GPUs, accounting for 81% of the market share, and has a leading advantage, while AMD ranks second with 19%.</p><p>In 2019 alone, with products such as the V100 series, Nvidia occupied 90% of China's AI training chip market, firmly grasping China's huge AI chip sales market.</p><p>One of the core reasons why Nvidia can continue to be a \"chip overlord\" lies in its \"light design mode\". Nvidia does not do chip manufacturing and packaging, leaving it to<a href=\"https://laohu8.com/S/TSM\">TSMC</a>OEM is completed, and it enjoys the dividend of advanced process technology such as 7nm. Between 2016 and 2021, Nvidia's revenue grew 233% and operating profit doubled to $4.5 billion, according to the earnings report. In the three months to May, sales jumped 84% year-on-year, while gross profit margin reached 64%.</p><p><b>In fact, China entered the field of GPU chip design very early, but the results were not satisfactory.</b></p><p><b>Since the 1970s, China began to introduce semiconductor and integrated circuit technologies and production lines. However, as a result, we have fallen into an a vicious circle of \"generation-to-generation introduction and generation-to-generation backwardness\". Coupled with the adverse impact of the \"Hanxin No.1\" fake chip incident on the society, China's \"autonomous processors\" have suffered serious setbacks, and China's active promotion of WTO globalization and other factors, thus missing the golden period of global semiconductor industry development, and downstream enterprises can only \"buy rather than make\".</b></p><p>By 2000, marked by the promulgation of the national \"Document No. 18\", China Semiconductor gradually formed an industrial organization form of \"separation of three industries\" in design, manufacturing, packaging and testing, and introduced \"<a href=\"https://laohu8.com/S/688981\">SMIC</a>\"A group of chip manufacturing (Foundry) enterprises represented by China have been built and put into production, and their technical level has been rapidly improved.</p><p>At present, Jingjiawei, Tianshu Zhixin, Denglin Technology, Biren Technology, Suiyuan Technology, Cambrian, Muxi IC and other enterprises are all gathered in the general-purpose processor track.</p><p>In 2014, Jingjiawei (300474.SH), which started as a military aircraft graphical display control module, successfully developed the military GPU chip JM5400, and then successfully developed the second generation GPU chip JM7200 with 28nm process technology in 2018. Jingjiawei has moved from military customization to general-purpose GPU, becoming one of the few companies in the world and the only company in China to realize commercial mass production of independent GPU.</p><p>In addition to Jingjiawei, in March 2021, Tianshu Zhixin released the first GPGPU (general-purpose graphics processor) manufactured by 7nm process in China, that is, 30% of the graphics rendering part of the traditional GPU was removed, and it was only born for handling artificial intelligence (AI) applications; In June this year, Suiyuan Technology released China's largest AI computing chip \"Sisi 2.0\" AI chip, \"Yunsui T20\" training accelerator card based on Sisi 2.0 and \"Yunsui T21\" training OAM module.</p><p>However, it is worth noting that the performance of JM7200 chip developed by Jingjiawei is only equivalent to that of Nvidia GTX 640 in 2012, which is difficult to meet the application needs of enterprise customers. Even Suiyuan Technology's \"Suisi 2.0\" AI chip only tied with Nvidia's A100, and two of the six projects tested by Benchmark greatly surpassed the performance of Nvidia A100. (For details, see the previous article of Titanium Media App: \"Suiyuan Technology releases China's largest AI computing chip to accelerate the implementation of three major business directions\")</p><p><img src=\"https://static.tigerbbs.com/15cb73d2787595547a9ed01926f4dfbe\" tg-width=\"570\" tg-height=\"1926\" referrerpolicy=\"no-referrer\" width=\"100%\" height=\"auto\"></p><p><b>The reason behind this is mainly that China's semiconductor industry started late, the technical threshold of chips is high, the cost elasticity is high, and the industry is highly concentrated, which makes the overall R&D investment, technology and talents of Chinese GPU chip enterprises lag behind foreign countries. Therefore, there is still a gap with chip giants in product performance and technology, and it is still difficult for downstream enterprises to break away from the position of \"American Chip Big Three\".</b></p><p>Take R&D investment as an example. In the ten years from 2011 to 2020, Jingjiawei's R&D investment totaled RMB 627 million, while Nvidia's R&D investment in 2020 reached USD 3.924 billion, or about RMB 25.323 billion. In the ten years, Nvidia's total investment exceeded RMB 120 billion, with a difference of more than 190 times.</p><p>In terms of talent, as of the first half of 2021, Nvidia had a whopping 18,975 employees and Jingjiawei had a total of 1,174 employees, far lower than AMD's 2,000 employees in the Shanghai R&D center.</p><p>\"AI chip and GPU chip markets are quite special. Unlike traditional special-purpose processors, the technology is very complicated. It requires a lot of data and needs to be combined with specific algorithms before it can be put into the market.\"<a href=\"https://laohu8.com/S/SNPS\">Synopsys</a>Xie Zhonghui, deputy general manager of China, said in an exclusive interview with Titanium Media App in April this year that if an enterprise wants to make the first AI chip solid, it usually takes more than two or three years.</p><p>In his view, chip semiconductor itself is an industry with large investment, long cycle and slow results. Complete localization of technology requires long-term and sustained accumulation of capital, talents and technology, and it is difficult to deal with it with the Internet thinking of \"spending money and seeing returns\".</p><p><b>In addition, the software and hardware ecology combined with CUDA technology is another important reason for the big gap between domestic chip companies and Nvidia.</b></p><p>In 2006, Nvidia released the parallel computing platform CUDA, which contains a series of development tools. Only by installing and using this platform can complex parallel computing be performed. Anyone who owns a laptop equipped with Nvidia GPU can use CUDA to perform scientific and convenient programming calculations, such as deep learning, AI algorithms, etc., and develop related software. Over the past decade or so, NVIDIA has persistently promoted CUDA, so that more government and enterprise-level software has been developed based on this platform, and NVIDIA's self-developed GPU hardware has been combined with CUDA software to efficiently realize the application landing.</p><p>In contrast, at present, there is no series of platforms similar to CUDA and NVIDIA hardware deep binding in China, and the gap in technical barriers is very obvious. Most domestic GPU manufacturers have adopted the strategy of compatibility with CUDA open source framework, such as Tianshu Zhixin and Denglin, and are ready to cultivate their own software ecosystem on this basis.</p><p>\"In the short term, it is easier to develop domestic GPU compatible with CUDA. After all, writing operators is a manpower-intensive industry. If users migrate, they need to migrate 100%, and the whole set of codes have to run on your chip. If the amount of code is small, there are not so many operators needed, and the difficulty will be relatively low. But in the long run, it is still necessary to get rid of compatibility ideas and develop its own core technologies.\" According to chip industry insiders, choosing compatibility is mainly to ensure that the existing software is still available. In the future, it will continue to improve its own platform to make it more matched with its own chips, thus attracting developers to migrate.</p><p>However, some enterprises choose to be incompatible with CUDA ecology. For example, Suiyuan Technology, which does AI training and inference chips at the same time, has comprehensively upgraded its \"TopsRider\" software platform and a brand-new \"Yunsui Cluster\" this year, hoping to have ecological dominance.</p><p><b>To sum up, these domestic chip companies that benchmark Nvidia are still in the initial stage of development. The industrialization and marketization ability of AI chip technology is weak, and it has not actually been used on a large scale. There is still a long way to go before surpassing or replacing \"China Nvidia\".</b></p><p>Yan Guihai, CEO of Zhongke Yushu, said in an interview with Titanium Media App that although China's demand side is still the largest single market in the world, and its growth rate is among the best, the \"demand side\" is still very strong, but there is still a big gap in design and manufacturing of high-end chips, and the \"supply side\" is not strong enough. He pointed out that the advantages and disadvantages of the supply side depend not only on one enterprise, but on the capacity of the whole industrial chain. In the short term, it is still unrealistic to build such a large-scale and comprehensive leading enterprise.</p><p>Founded in 2018, Zhongke Yushu is a developer of dedicated computing architecture. It was incubated from the State Key Laboratory of Computer Architecture of Institute of Computing Technology, Chinese Academy of Sciences. Now the company's valuation has exceeded 1 billion yuan. On July 27th this year, Zhongke Yushu completed hundreds of millions of yuan in Series A financing, led by Huatai Innovation, followed by Lingjun Investment and the old shareholder Guoxin Sichuang.</p><p>Huang Liming, a partner at Hillhouse and head of software and hard technology at Hillhouse Ventures, said in an exclusive interview with Titanium Media App that although the GPU market is promising, China's<a href=\"https://laohu8.com/S/V03.SI\">Startups</a>It is difficult to make \"Nvidia\" directly. In addition to technical difficulty, it has to be done in combination with strong applications-involving software system software ecology, which is extremely demanding for startups.</p><p>Hillhouse launched Hillhouse Venture Capital, an independent VC brand, in February 2020. Since then, it has invested in the field of chip semiconductors, including semiconductor IP company Xinyaohui, EDA manufacturer Xinhuazhang, GPU platform Biren Technology, DPU company Nebula Zhilian, plus Tianke Heda in silicon carbide, Minxin Semiconductor in optical chip field, and Xingsi Semiconductor in mobile phone baseband.</p><p>Huang Liming emphasized that there will not be many companies that can run out in this direction, whether overseas or domestic, in the end.</p><p><h2>The wind has arrived</h2>\"Let's not struggle with how to replace Nvidia now. The road is step by step. I think first of all, China has to have domestic AI chips, general-purpose GPUs, FPGAs and other underlying computing power. As long as there is domestic market demand, we must have many opportunities.\" Zhang Gaonan, managing partner of Huaying Capital, told Titanium Media App that the domestic semiconductor industry has arrived, and China's entry into the GPU market is \"the right time, the right place and the right people\". In particular, semiconductors and the next generation AI technology are areas that China must break through.</p><p>Zhang Gaonan gave an example,<a href=\"https://laohu8.com/S/688111\">Kingsoft Office</a>The product is inferior to<a href=\"https://laohu8.com/S/MSFT\">Microsoft</a>Office suite, but the market gives a market value of more than 110 billion yuan. One of the important reasons behind it is that China must have domestic Office, and the same reason applies to the domestic GPU market.</p><p>Huaying Capital is one of the earliest private equity funds in China To deploy mobile Internet and cultural industry. In recent years, To B field has also become the key investment area of Huaying Capital. At present, Huaying Capital has invested more than 30 projects in the To B field, with a total investment of over 700 million yuan, which are built by Zhang Gaonan, a technology investor. In the fields of data center and underlying computing power, Huaying Capital has invested in projects such as Biren Technology, Tianyun Big Data and Zhongke Haiwei.</p><p>In fact, as a general platform spanning visual computing and AI computing, GPU has a huge market space.<b>According to<a href=\"https://laohu8.com/S/601555\">Soochow Securities</a>It is estimated that by 2027, the market space of domestic substitution in GPU field will exceed 34.1 billion USD. In addition to the existing game market, there is room for further development in industrial, medical, military aerospace and other directions.</b></p><p>In March this year, \"Biren Technology\", a general intelligent chip designer co-founded by Zhang Wen, the former president of Shangtang Technology, completed the B round of financing. Since its establishment in September 2019, the company's total financing amount has exceeded RMB 4.7 billion, and its investors include Hillhouse Venture Capital, Huaying Capital,<a href=\"https://laohu8.com/S/601318\">Ping An of China</a>China Merchants Capital, BAI Capital, Guosheng Group National Reform Fund, etc., with a valuation of over 10 billion yuan, making it one of the \"unicorn\" enterprises with the fastest momentum in the growing semiconductor industry.</p><p><b>In addition to Wall Ren, Mu Xi IC, Moore Thread and other enterprises in the GPU field have also completed financing.</b></p><p>On August 25th, Muxi Integrated Circuit, a GPU manufacturer, announced the completion of RMB 1 billion Series A financing. The founders Chen Weiliang and Yang Jian are all from American chip giant AMD, and the investors include China State-owned Enterprise Structural Adjustment Fund Co., Ltd.,<a href=\"https://laohu8.com/S/00810\">China Internet Investment</a>Fund, Matrix Partners China, Holy Capital, Sequoia China, Lightspeed China, Guochuang Zhongding, Smart Internet Industry Fund, Shanghai Science and Technology Fund, Lenovo Venture Capital, etc.; Moore Thread, established in 2020, claimed to have completed two rounds of multi-billion yuan financing within 100 days. The team members are mainly from Nvidia, and the investors include Shenzhen Venture Capital and Sequoia Capital<a href=\"https://laohu8.com/S/CHN\">China Fund</a>China Merchants Venture Capital, ByteDance, Pony Zhixing, Wuyuan Capital, etc.</p><p><b>However, an interesting phenomenon is that the above three start-ups, namely Bi Ren, Mu Xi and Moore Thread, are veritable \"PPT financing core-making\", and none of them completed the taping of the first chip (trial production of assembly line) during financing.</b></p><p>Why are market investors willing to open their wallets to this? In an interview with Titanium Media App, several investors said that these projects can get a lot of capital support, all because of the early investment, mainly depending on the team and the track: the AI chip track has arrived, and the senior executive team is also from the \"American Chip Big Three\".</p><p>\"I think we need to give these enterprises opportunities and patience. It is impossible for 500 people to write PPT. Manufacturing chips is a matter of 5 to 10 years. The reason why we are willing to invest is not speculation or fooling. I think the risk of investing in the semiconductor track itself is high, and it is necessary to make long-term plans and have strong enough risk-taking capacity, which is completely different from the model innovation of investing in the Internet.\" The above-mentioned investors told Titanium Media App.</p><p>However, some investors in the semiconductor industry pointed out that the above-mentioned investment projects essentially hope to raise the market value and have a higher rate of return. Especially in the \"chip fever\" environment, venture capital institutions need to constantly look for these GPU and AI chip enterprise targets in the middle and early stages, hoping to gamble on a higher return.</p><p>In addition, in this wave of GPU entrepreneurship, the founding team teached the \"American Chip Big Three\". For example, Zheng Jinshan, chief scientist of Tianshu Zhixin, was the chief technical expert of AMD; Muxi's founding team is mainly from AMD. CEO Chen Weiliang once served as senior director of graphics research and development in AMD, and CTO Yang Jian once served as AMD Fellow (academician); Li Xinrong, the newly appointed co-CEO of Biren Technology, was once the global vice president of AMD, and Chen Wenzhong, senior vice president of Biren Technology, also served in AMD.</p><p>In this regard, Zhang Gaonan said that AMD is the top two chip giants in the field of GPU. The core research and development of GPU is all in Shanghai, while the research and development of graphics rendering is in the United States. Enterprises can go to the executives of AMD and Nvidia to communicate, and the final choice is definitely the leader in the semiconductor industry.</p><p>Yan Guihai believes that in segmenting emerging tracks, with the application \"potential energy\" on the demand side, Chinese chip enterprises have concentrated their superior forces, based on serving local enterprises, and highlighted the agility of development, which is an opportunity to surpass chip giants such as \"Nvidia\" in product definition and scheme iteration cycle.<b>He predicts that a number of domestic GPU and DPU enterprises with leading technology will appear within 10 years.</b></p><p>\"Among the five links of the chip industry: design, manufacturing, packaging and testing, materials and EDA, the most relevant to application is design, and our biggest advantage lies in application, so we have a great opportunity to make a breakthrough in the\" design \"link, and then gradually expand the territory of victory from point to point. The so-called \"corner overtaking\" is still a catch-up strategy. It is more likely to occupy a new strategic commanding height by cutting into a new future-oriented track and accelerating at full speed. I hope that within 10 years, a number of enterprises with leading technology, solid products and strategic awareness will emerge. \" Yan Guihai told Titanium Media App.</p><p>Zhang Wen, founder of Biren Technology, said that the capability requirements for chip companies have been upgraded from product level to system level and ecological level. In less than five years, China has caught up with or even led the international standard in the field of AI chip design. He emphasized that going beyond Nvidia requires redefining a product, as well as redefining a market.</p><p><h2>Ten billion DPU chip market \"explodes\"</h2>In Huang Renxun's view, the data processing unit (DPU) responsible for transmitting and processing data in the data center is forming \"the three pillars of future computing\" together with CPU and GPU. When Chinese chip companies developed GPU, Nvidia set its sights on the two new markets of CPU and DPU.</p><p>In September 2020, Nvidia announced that it planned to acquire British chip designer Arm from a Japanese software group for US$40 billion, and it is expected to write the largest merger and acquisition case in the semiconductor industry. But the deal is controversial and awaits approval from governments including the European Union, Britain, the United States and China. But in April 2021, Nvidia announced its entry into the data center CPU market and released the Grace CPU processor, which is what Huang Renxun said at the beginning of this article. (For details, see the previous article of Titanium Media App: \"The British government intervened, and Nvidia's $40 billion acquisition of Arm transaction changed\")</p><p>In addition to CPU and GPU, Nvidia is also laying out DPU. In 2019, Nvidia announced the acquisition of Mellanox, an Israeli network chipmaker, for $6.9 billion in all-cash, and eventually won it. In this largest acquisition in Nvidia's history, what Huang Renxun values most is Meros's unique ability in data center technology and other aspects.<b>In October 2020, NVIDIA debuted its DPU — the NVIDIA BlueField series of data processors.</b></p><p>Fundamentally, on the one hand, DPU is more flexible and safe, and more importantly, DPU can free the computing power of CPU, release the load of server, significantly reduce the comprehensive cost with low power consumption, and even improve the performance of AI and machine learning applications.</p><p>According to IDC statistics, the global demand for computing power will double every 3.5 months, far exceeding the current growth rate of computing power. Driven by this, the global computing, storage and network infrastructure is also undergoing fundamental changes: some workloads with excessive data occupy too much CPU resources, and various \"X\" PU chips that cooperate with them have emerged. Besides GPU, FPGA and other chips, DPU is the next \"X\" PU.</p><p>People in the industry have made a vivid metaphor for this. The network is like building roads. In the past, the roads were not wide enough in the 1G and 10G era, and there were more and more cars. In order to balance the pressure, it was necessary to increase traffic lights and invest more traffic police to coordinate resources, which has improved the original efficiency a lot, but it is still not enough. First, we must expand the road, which is the increase of bandwidth. However, the road has changed from 2 to 4. Just relying on traffic lights and limited traffic police will still be blocked. However, we can't increase the traffic police indefinitely, which requires the road to be more intelligent and help solve the congestion.</p><p>Zhang Gaonan pointed out that a large amount of network management is in the CPU, which occupies the container capacity, while the DPU intelligently provides the space capacity of the server, and a large amount of virtualization space can increase the computing power demand.</p><p><b>With 2020, DPU's reputation exceeded the infrastructure processor (IPU) and SmartNIC launched by rival Intel, and it also made every enterprise eyeing data center business get a piece of this field. DPU has become a new track for major chip giants and start-up companies to compete for research and development. Domestic DPU is now almost in a state of blooming, and Sequoia, Hillhouse Venture Capital, CDH and SoftBank China have all begun to enter the market.</b></p><p>In April this year, Tianyan Check data showed that the domestic DPU chip supplier \"Cloud Leopard Intelligence\" completed<a href=\"https://laohu8.com/S/00700\">Tencent</a>Investment, Sequoia Capital, Yaotu Capital and other joint angel round financing; At the end of May, Xinqiyuan completed hundreds of millions of yuan of Pre-A round of financing, with investors including SoftBank China and Pudong Science and Technology Group; On July 27th, DPU chip developer \"Zhongke Yushu\" completed hundreds of millions of yuan of Series A financing led by Huatai Innovation; On August 30th, Nebula Zhilian, a developer of DPU chips, announced the completion of hundreds of millions of angel round financing, led by Hillhouse Venture Capital, and CDH VGC and Walden International China Fund participated in the follow-up investment; At the beginning of September, IDG Capital threw the \"Yunmai Xinlian\" angel round financing project.</p><p>\"DPU may become the third computing chip after CPU and GPU, but from the structural point of view, DPU will be more heterogeneous and more dedicated.\" Yan Guihai said in an interview with Titanium Media App and others that the background of DPU is the demand for computing delay, data security and resource virtualization brought by the end-edge-cloud integration trend caused by the data explosion in the intelligent era. The CPU is overwhelmed with these non-business loads, and there is an urgent need for an ideal object to share these computational loads.</p><p><b>Leopard Research Institute predicts that China's DPU market is expected to reach 3.74 billion USD in 2025. The global DPU market size is expected to reach USD 13.57 billion in 2025.</b>At the same time, the report also pointed out that data circulation is the largest application market of DPU, among which bare metal services are just in demand for DPU. The application of DPU in the telecom market is mainly in edge computing scenarios, with a penetration rate of less than 5%. DPU for the field of intelligent driving is still in the exploratory stage, and it is expected that DPU will be deployed in the field of intelligent driving only in 2023.<img src=\"https://static.tigerbbs.com/7170c585dab44018726df81a32c63d62\" tg-width=\"808\" tg-height=\"500\" referrerpolicy=\"no-referrer\" width=\"100%\" height=\"auto\"></p><p>China DPU Market Size, Forecast 2020-2025, Source: Toubao Research Institute</p><p>Yan Guihai pointed out that the performance of CPU increased from 30% per year 5-10 years ago to less than 3% per year three years ago. However, the network bandwidth is still increasing by about 35% every year.</p><p>With<a href=\"https://laohu8.com/S/600918\">Zhongtai Securities</a>For example, the challenge that the company encountered at that time was that transaction statement compliance checks were too slow, and it needed to improve transaction efficiency. Therefore, Zhongke Yushu, Zhongtai Securities and Shanghai Stock Exchange Technology Co., Ltd. jointly developed a set of extremely fast risk control system solutions to speed up this process. Zhongke Yushu has successively developed a number of products and solutions such as ultra-low latency smart network card and data computing accelerator card, mainly for high bandwidth, low latency and data-intensive scenarios. The company has achieved quarterly revenue of tens of millions this year. Zhongke Yushu's next-generation DPU chip is expected to be designed by the end of 2021 and is expected to handle up to 200G of network bandwidth data.</p><p><b>However, although the DPU market is hot, the concept is newer, there are more unknowns, and the investment risk will be greater.</b></p><p>Lu Sheng, CEO of Xinqiyuan, pointed out that at present, the barriers of DPU segmentation track are still relatively high. In addition to technical barriers, there are also market barriers, which require customers to iterate continuously, especially with the continuous upgrading of open source software to adapt to customers' rapidly changing software and hardware environment. Therefore, VC (venture capital) must recognize the track and have enough patience before investing. He emphasized that investors need to constantly observe and adjust their judgments on the market. Now no one can predict the future development prospects of DPU.</p><p>Some media also believe that when Nvidia enters the newly opened battlefield of CPU and DPU, it may be good for Chinese GPU manufacturers. In particular, Nvidia still spends a lot of energy on the M&A transaction of acquiring British chip designer Arm for $40 billion, which may be an opportunity for new GPU companies to catch up.</p><p>As Zhang Gaonan said to Titanium Media App, \"Logically speaking, things with low thresholds are usually not scarce. (Chip semiconductor track) Some things are difficult and require a long time of investment. Although it is high risk, someone has to do it. This is a long-term investment that is really beneficial to the country. In fact, investment should be encouraged. Otherwise, no one will do these difficult things that require a long time of investment, and you will never get up the steps.\"</p><p>Zhang Gaonan emphasized that although venture capital must pursue returns, he believes that it is not only of strong social significance to invest in some start-ups in the semiconductor track under the condition that the whole capital is allocated reasonably, but also an important embodiment of some long-term value investment.</p>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/01e0bbbfdbdc6f8f2f7ce512582f7c3e","relate_stocks":{"NVDA":"英伟达"},"is_english":false,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"2166195813","content_text":"英伟达CEO黄仁勋（图片来源：Nvidia官网）\n“我即将展示的产品，融合了新的GPU加速计算能力，拥有Mellanox高性能网络，补足我们最后一块拼图的产品是——全球首款专为TB级数据中心加速计算而设计的CPU处理器，它的秘密代号是Grace。”\n这是2021年4月英伟达（NVIDIA）CEO黄仁勋在GTC峰会演讲中的一段话。然而，让人意想不到的是，直到8月12日英伟达自曝后人们才知道，这段不足100字、14秒的演讲内容竟然不是黄仁勋本人出镜，而是使用了合成的“数字替身”，即利用英伟达GPU处理器与Omniverse软件平台共同形成的“虚拟黄仁勋”形象。\n这一事件引发了人们对虚拟现实、元宇宙、AI换脸等技术和概念的激烈讨论，同时也让“英伟达”这家美国芯片霸主从半导体行业“出圈”，走入了大众视野。\n自1993年成立至今，在黄仁勋的带领下，英伟达成功创造且引领了GPU（图形处理器）芯片这一类别，产品覆盖整个PC设备GPU至服务器GPU市场。过去五年内，英伟达市值从310亿美元增长到5050亿美元，跻身成为全球第七大半导体供应商，是人工智能（AI）芯片领域炙手可热的明星企业。\n与此同时，在英伟达市值超过英特尔之后，国内半导体市场看到了GPU、AI芯片赛道更大的市场机会，景嘉微、天数智芯、登临科技、壁仞科技、燧原科技、寒武纪、沐曦集成电路等企业均在通用处理器这一赛道中集聚。\n近两年，国内半导体产业发生着转变。由于华为海思、中芯国际等企业受美方“实体清单”影响，以及全球芯片短缺引发的连锁反应，使得中国企业愈加难采购到美国半导体产品，对于“国产替代”需求愈加强烈。\n2020年6月29日，由于美国《出口管理条例》再升级，实体清单企业成员在不断调整，英特尔“临时性暂停”对浪潮集团的芯片供货。\n尽管随后在7月3日，浪潮方面宣布英特尔已恢复对其供货，但一家是全球最大的PC、云服务x86架构芯片供应商，另一方是中国最大的服务器厂商，两家公司之间的临时断供危机，为整个中国云计算、半导体行业提了一个醒：随着算力需求越来越强烈，中国需要大规模生产全面自主可控的国产GPU、服务器芯片产品。从而凸显了“国产替代”正成为中国半导体行业发展的最大驱动力。\n另外，作为全球芯片销量大国，中国却没有出现一家“英伟达”这样的芯片巨头，大市场并没有产生与之匹配的大公司。根据IC Insight的统计显示，2020年全球半导体市场规模为3957亿美元，其中，中国大陆市场规模是434亿美元，为全球最大市场，占全球比例达到36.24%。然而，总部位于中国大陆的半导体公司2020年总产值仅为83亿美元，仅占市场规模的5.9%。\n偌大的蛋糕，究竟谁能切下一角？半导体产业何时才能造出“中国英伟达”？\n错失黄金时代\nGPU图形处理器又被称为显示芯片、视觉处理器，最初于1999年由英伟达提出，是个人电脑、工作站、游戏主机以及移动设备（智能手机、平板电脑、VR设备）上专门运行绘图运算的微处理器。\n随着GPU的并行计算优势被逐步挖掘，GPU的应用领域从图形处理扩展到高性能计算，逐步成为Al计算最成熟、应用最广泛的通用型芯片。2020年6月，英伟达推出基于安培（Ampere）架构的A100 Tensor Core GPU，成为全球性能最强的AI芯片。\n以应用终端角度分类，GPU可分为PC端GPU、服务器GPU和移动端GPU，对应三种架构，即与专用电路板及组件组成的独立显卡，共享集成显卡，以及移动端GPU与其他芯片或模块一起封装成高集成度的SoC——应用于手机、汽车电子、AI在内的多个应用场景。\n自从AMD在2006年收购加拿大GPU厂商ATI之后，目前，在PC及服务器GPU领域，全球GPU市场呈现“美国芯片三巨头”——英特尔、AMD和英伟达垄断的局面。集成GPU市场英特尔优势明显，独立GPU市场英伟达和AMD两强割据。\n根据研究机构Jon Peddie Research的数据显示，2021年第一季度，全球PC端GPU市场中，英特尔（Intel）以68%市场份额位居榜首，AMD和英伟达分别为17%和14%，三家共计份额接近100%；全球独立GPU领域中，英伟达是数据中心GPU市场领导者，占据81%的市场份额，拥有领先优势，AMD则以占比19%位居第二。\n仅2019年，英伟达凭借V100系列等产品，占据了中国AI训练芯片市场90%份额，牢牢掌握着中国这一庞大的AI芯片销售市场。\n英伟达能持续作为“芯片霸主”地位的核心原因之一在于其“轻设计模式”。英伟达不做芯片制造和封装，交由台积电代工完成，自身享受7nm等先进制程工艺技术红利。根据财报显示，2016年至2021年期间，英伟达收入增长了233%，营业利润翻了一番，达到45亿美元。在截至今年5月的三个月内，销售额同比猛增84%，毛利率则达到了64%。\n事实上，中国很早就进入了GPU芯片设计领域，但结果并不如意。\n从20世纪70年代开始，中国开始引进半导体与集成电路技术和生产线。但结果却是陷入了“代代引进、代代落后”的恶性循环，加上“汉芯一号”假芯片事件给社会带来的不良影响，让中国的“自主处理器”遭受严重挫败，以及中国积极推动WTO全球化等因素，从而错失了全球半导体产业发展的黄金时期，下游企业只能“造不如买”。\n到2000年，以国家“18号文件”出台为标志，中国半导体才逐渐形成设计、制造、封装测试“三业分离”的产业组织形态，引进以“中芯国际”为代表的一批芯片制造（Foundry）企业在华建设、投产，技术水平也因此得到快速提升。\n目前，景嘉微、天数智芯、登临科技、壁仞科技、燧原科技、寒武纪、沐曦集成电路等企业均在通用处理器这一赛道中集聚。\n2014年，以军机图形显示控制模块起家的“景嘉微”（300474.SH）成功研制出军用GPU芯片JM5400，随后在2018年成功研发出28nm制程工艺的第二代GPU芯片JM7200。景嘉微从军用定制走向通用GPU，成为全球少数、国内唯一实现独立GPU商用量产的公司。\n除景嘉微外，2021年3月天数智芯发布了国内首颗7纳米工艺制造的GPGPU（通用图形处理器），即去掉了传统GPU 30%的图形渲染部分，只为处理人工智能（AI）应用而生；燧原科技则在今年6月发布了迄今中国最大的AI计算芯片“邃思2.0”AI芯片、基于邃思2.0的“云燧T20”训练加速卡和“云燧T21”训练OAM模组。\n但值得注意的是，景嘉微研发的JM7200芯片，性能只相当于2012年英伟达GTX 640水平，难以满足企业客户的应用需求。即便燧原科技的“邃思2.0”AI芯片，也仅和英伟达的A100达成平手，Benchmark测试的6个项目中有2项大幅超越了英伟达A100的性能表现。（详见钛媒体App前文：《燧原科技发布中国最大的AI计算芯片，加速推进三大业务方向落地》）\n\n背后的原因，主要由于中国半导体产业起步晚，芯片的技术门槛高、成本弹性大、产业高度集中，使得中国GPU芯片企业的整体研发投入、技术、人才都滞后于国外，从而在产品性能和技术上依然和芯片巨头有差距，下游企业依然难以脱离“美国芯片三巨头”的境地。\n以研发投入为例，2011年至2020年的十年间，景嘉微的研发投入费用总额为人民币6.27亿元，而英伟达2020年这一年的研发投入就达到39.24亿美元，约合人民币253.23亿元，十年间英伟达总计投入超过1200亿元人民币，两者相差超190倍。\n在人才方面，截至2021年上半年，英伟达员工人数高达18975人，景嘉微总员工人数为1174人，远低于AMD在上海研发中心的2000名员工。\n“AI芯片、GPU芯片市场比较特殊，跟传统的专用处理器不一样，技术十分复杂。它需要大量的数据，需要和特定的算法结合，才能够付诸市场运用。”新思科技中国副总经理谢仲辉在今年4月接受钛媒体App独家专访时表示，如果企业想把首颗AI芯片做扎实，通常需要两三年以上。\n在他看来，芯片半导体本身是一个投入大、周期长、见效慢的行业，技术完全国产化需要长期持续的资金、人才和技术积累，很难用“砸金钱见回报”这种互联网思维来处理。\n此外，结合CUDA技术的软硬件生态，也是国内芯片企业与英伟达形成较大差距的另一重要原因。\n2006年，英伟达就发布了并行计算平台CUDA，其中包含一系列开发工具，只有安装使用这个平台才能够进行复杂的并行计算，任何人只要拥有一台配有英伟达GPU的笔记本电脑，就可以利用CUDA可以进行科学、便捷编程计算，比如深度学习、AI算法等，开发相关软件。过去十多年，英伟达坚持不懈地推广CUDA，使更多政企级类型软件都基于该平台开发，将英伟达自研GPU硬件与CUDA软件相结合，高效实现应用落地。\n相比之下，目前国内却没有一个类似CUDA和英伟达硬件深度绑定的系列平台，技术壁垒差距十分明显。大部分国产GPU厂商均采取兼容CUDA开源框架的策略，如天数智芯、登临等，准备在此基础上培育自己的软件生态。\n“短期来看，国产GPU兼容CUDA更容易发展，毕竟写算子是人力密集型行业，用户迁移的话是需要100%迁移、整套代码都要在你的片上跑，如果代码量很小，需要的算子不那么多，难度就比较低。但是长期来看，还是要摆脱兼容思路，发展自有的核心技术。”芯片行业内人士表示，选择兼容主要是确保已有软件依然可用，未来会不断改进自家平台，使其更加匹配自己的芯片，从而吸引开发者迁移。\n但也有企业选择不兼容CUDA生态，比如同时做AI训练和推理芯片的燧原科技，今年全面升级了其“驭算TopsRider”软件平台以及全新的“云燧集群”，希望能拥有生态主导权。\n总结来看，对标英伟达的这些国内芯片企业依然处在发展的初级阶段，AI芯片技术的产业化、市场化能力较弱，没有产生实际的大规模使用，距离超越或取代“中国英伟达”仍然有很长的路要走。\n中科驭数CEO鄢贵海在接受钛媒体App采访时表示，虽然目前中国需求侧虽然还是全球最大的单一市场，增速也名列前茅，“需求侧”还是很强劲的，但在高端芯片方面无论是设计还是制造还有不小差距，“供给侧”不够强大。他指出，供给侧的优劣不仅取决于一家企业，而是全产业链能力。短期内要想打造出这样大体量和全面引领性的企业还是不太现实的。\n中科驭数成立于2018年，是一家专用计算架构研发商，孵化自中科院计算所的计算机体系结构国家重点实验室，如今公司估值已超10亿元。今年7月27日，中科驭数完成数亿元A轮融资，由华泰创新领投，灵均投资以及老股东国新思创跟投。\n高瓴合伙人、高瓴创投软件与硬科技负责人黄立明在接受钛媒体App的独家专访时表示，虽然GPU市场前景广阔，但中国创业公司很难直接做成“英伟达”。除了技术难度外，还要结合很强的应用来做——涉及到软件系统软件生态，这对创业公司来说要求是极高的。\n高瓴于2020年2月推出独立VC品牌高瓴创投，此后其对芯片半导体领域进行投资入局，其中包括半导体IP企业芯耀辉、EDA厂商芯华章，GPU平台壁仞科技、DPU公司星云智联，加上碳化硅方面的天科合达、光芯片领域的敏芯半导体、以及手机基带星思半导体等。\n黄立明强调，能在这个方向跑出来的公司，无论海外还是国内，高瓴判断最终都不会有很多。\n风口已至\n“我们现在先不纠结于怎么去取代英伟达，路都是一步一步走的。我觉得首先中国得有国产AI芯片、通用GPU、FPGA等底层算力。只要国内有市场需求，我们一定有很多机会。”华映资本主管合伙人章高男对钛媒体App表示，国内半导体产业风口已至，中国现在切入GPU市场是“天时、地利、人和”皆备，尤其半导体和下一代AI技术都是中国必须突围的领域。\n章高男举了一个例子，金山办公产品虽然逊于微软Office套件，但市场给出1100多亿元市值，背后重要原因之一是，中国必须得有国产office，同样道理也适用于国产的GPU市场。\n华映资本是国内最早布局移动互联网和文化产业的私募股权基金之一，近几年To B领域也成为华映资本重点关注的投资领域。目前华映资本在To B领域投资的30余个项目，投资总额超7亿元生态，由技术型投资人章高男负责搭建。在数据中台及底层算力相关领域，华映资本投资布局了壁仞科技、天云大数据，中科海微等项目。\n实际上，作为横跨视觉计算和AI计算的通用平台，GPU拥有巨大的市场空间。据东吴证券测算，预计到2027年，GPU领域国产替代的市场空间规模超过341亿美元。除了既有的游戏市场，在工业、医疗、军事航天等方向都有进一步的发挥空间。\n今年3月，原商汤科技总裁张文联合创立的通用智能芯片设计商“壁仞科技”完成了B轮融资。2019年9月成立以来，公司总融资额超47亿元人民币，投资方包括高瓴创投、华映资本、中国平安、招商局资本、BAI资本、国盛集团国改基金等，估值已超过100亿元，成长半导体行业势头最为迅猛的“独角兽”企业之一。\n除壁仞外，沐曦集成电路、摩尔线程等入局GPU领域的企业也都完成了融资。\n8月25日，GPU厂商沐曦集成电路宣布完成10亿元人民币的A轮融资，创始人陈维良、杨建等均来自美国芯片巨头AMD，投资方包括中国国有企业结构调整基金股份有限公司、中国互联网投资基金、经纬中国、和利资本、红杉中国、光速中国、国创中鼎、智慧互联产业基金、上海科创基金、联想创投等；而2020年成立的摩尔线程，宣称100天内就完成了两轮数十亿元融资，团队成员主要来自英伟达，投资方包括深创投、红杉资本中国基金、招商局创投、字节跳动、小马智行、五源资本等。\n不过，一个有趣的现象是，壁仞、沐曦、摩尔线程上述三家初创企业是名副其实的“PPT融资造芯”，融资时无一家完成首颗芯片的流片（流水线试生产）。\n为何市场投资人愿意对此敞开钱包？数位投资人在接受钛媒体App采访时表示，这些项目能够获得大量资本支持，原因都为投资早期，主要看的还是团队、赛道两部分：AI芯片赛道风口已至，高管团队也均出自“美国芯片三巨头”。\n“我觉得需要给这些企业机会和耐心，不可能500个人都在写PPT。制造芯片是一个5年到10年的事情，我们愿意去投的原因，并非是投机或者是忽悠。我认为，投半导体赛道本身风险就高，需要做好长周期的打算，需要有足够强的风险承担能力，这和投资互联网的模式创新完全不一样。”上述投资人对钛媒体App表示。\n但也有半导体行业投资人指出，上述投资项目本质上还是希望市值撑高，有更高的回报率，尤其“芯片热”环境下，风投机构需要不断在中早期寻找这些GPU、AI芯片企业标的，希望从中赌得一份更高的回报。\n此外，在这一波GPU创业浪潮中，创始团队师出“美国芯片三巨头”。例如，天数智芯首席科学家郑金山曾任AMD首席技术专家；沐曦的创始团队主要来自AMD，CEO陈维良曾在AMD担任图形研发高级总监，CTO杨建曾任AMD Fellow（院士）；壁仞科技最新上任的联席CEO李新荣，曾任AMD全球副总裁，壁仞科技高级副总裁陈文中也曾在AMD任职。\n对此，章高男表示，AMD是GPU领域排名前二的芯片巨头，关于GPU核心研发都在上海，而图形渲染的研发是在美国，企业可以去找AMD和英伟达两家公司高管去沟通，而最终选择的人肯定是半导体行业内的佼佼者。\n鄢贵海认为，在细分新兴赛道，凭借需求侧的应用“势能”，中国芯片企业集中优势兵力，立足服务本土企业，突出开发的敏捷性，是有机会在产品定义、方案迭代周期上超越“英伟达”这些芯片巨头。他预计，10年内会出现一批技术领先的国产GPU、DPU企业。\n“芯片产业五个环节：设计、制造、封测、材料、EDA五个环节中，与应用最相关的是设计，我们最大的优势又在于应用，所以非常有机会在“设计”这一环节取得突破，然后以点带面，逐步扩大胜利版图。所谓“弯道超车”还是追赶策略，切入面向未来的新赛道并且全力加速才更有可能占据新的战略制高点。希望能在10年内能出现一批技术领先、产品扎实而且富有战略意识的企业。”鄢贵海对钛媒体App表示。\n壁仞科技创始人张文表示，对芯片公司的能力要求从产品级提升到系统级和生态级。时间上不超过5年，中国在AI芯片设计领域赶上甚至领先国际水准。他强调，超越英伟达，需要重新定义一个产品，以及重新定义一个市场。\n百亿DPU芯片市场“爆火”\n在黄仁勋看来，负责在数据中心传输和处理数据的数据处理单元（DPU），正与CPU、GPU共同组成“未来计算的三大支柱”。当中国芯片企业发力GPU时，英伟达则把目光放在了CPU、DPU这两个新市场中。\n2020年9月，英伟达宣布拟以400亿美元，从日本软件集团处收购英国芯片设计商Arm，预计写下半导体行业最大的并购案。但这笔交易存有争议，目前还等待欧盟、英国、美国和中国等政府的批准。但2021年4月，英伟达则宣布进军数据中心CPU市场，发布Grace CPU处理器，也就是本文开头黄仁勋所讲的那一段话。（详见钛媒体App前文：《英国政府出手干预，英伟达400亿美元并购Arm交易生变》）\nCPU和GPU之外，英伟达还在布局DPU。2019年，英伟达宣布以69亿美元全现金的形式收购以色列网络芯片商迈络思（Mellanox），并最终将其拿下。而这笔英伟达有史以来规模最大的收购，黄仁勋最看重的就是迈络思在数据中心技术等方面独步天下的能力。2020年10月，英伟达首次推出了DPU — NVIDIA BlueField系列数据处理器。\n究其根本，一方面DPU更灵活安全，更重要的是，DPU可以解放CPU的算力，释放服务器的负载，并凭借低功耗显著降低综合成本，甚至还可以改善AI和机器学习应用的性能。\n据IDC统计，全球算力的需求每3.5个月就会翻一倍，远远超过了当前算力的增长速度。在此驱动下，全球计算、存储和网络基础设施也在发生根本转变：一些数据量过大的工作负载，过多占用CPU资源，与之协同作战的各种“X”PU芯片便应运而生，GPU、FPGA等芯片之外，DPU就是下一个“X”PU。\n业内人士就此做了一个形象的比喻，网络就像造马路，以前1G 10G时代马路已经不够宽了，车子越来越多，为了平衡压力，通过增加红绿灯和投入更多的交警来更高的协调资源，这样已经让原来的效率提高很多，但是仍然不够。必须第一扩大马路，这就是带宽增加，但是马路从2道变为4道，仅仅依靠红绿灯和有限的交警还是会堵塞，但是我们不能无限增加交警，这就需要马路能更加智能，帮助解决拥堵。\n章高男指出，大量的网络管理在CPU里面，占据了容器能力，而DPU则是将服务器智能提供空间能力，大量虚拟化空间可以提高算力需求。\n随着2020年，DPU的名声超出了竞争对手英特尔所推出的基础设施处理器(IPU)和SmartNIC，也让每个对数据中心业务虎视眈眈的企业都要在这个领域分一杯羹。DPU成为了各大芯片巨头、初创公司争相研发的新赛道，国产DPU现在几乎处在百花齐放的状态，红杉、高瓴创投、鼎晖、软银中国都开始入场。\n今年4月，天眼查数据显示，国产DPU芯片供应商“云豹智能”完成腾讯投资、红杉资本、耀途资本等联合的天使轮融资；5月末，芯启源完成数亿元Pre-A轮融资，投资方包括软银中国、浦东科创集团等；7月27日，DPU芯片研发商“中科驭数”完成华泰创新领投的数亿元A轮融资；8月30日，DPU芯片研发商星云智联宣布完成了数亿元天使轮融资，由高瓴创投领投，鼎晖VGC、华登国际中国基金参与跟投；9月初，IDG资本豪掷“云脉芯联”天使轮融资项目。\n“DPU有可能成为继CPU和GPU之后的第三颗算力芯片，但从结构上来看，DPU会更异构、也更专用。”鄢贵海在接受钛媒体App等采访时表示，DPU产生的背景是智能时代数据爆发导致的端-边-云一体化趋势带来的对计算延迟、数据安全、资源虚拟化需求。CPU对这些非业务性负载已不堪重负，迫切需要一个理想的对象来分担这些计算负载。\n头豹研究院则预测，中国DPU市场规模预计将在2025年达到37.4亿美元。全球DPU市场规模2025年预计将达到135.7亿美元。同时报告也指出，数据流通是DPU最大的应用市场，其中裸金属服务其对DPU存在刚需。DPU在电信市场的应用主要为边缘计算场景，渗透率不足5%。针对智能驾驶领域的DPU仍在探索阶段，预计在2023年DPU才有望布局在智能驾驶领域。\n中国DPU市场规模，2020-2025年预测，来源：头豹研究院\n鄢贵海指出，CPU的性能从5-10年前每年30%的增幅，到三年前大概只有每年不到3%的性能增幅。而网络带宽每年依旧还有35%左右的增长。\n以中泰证券为例，当时该公司遇到的挑战是，交易报单合规检查太慢，需要提高交易效率。于是，中科驭数与中泰证券、上交所技术有限责任公司联合研发了一套极速风控系统解决方案，来加速这一流程。中科驭数相继研发了超低时延智能网卡、数据计算加速卡等多套产品和解决方案，主要面向高带宽、低时延、数据密集型等场景。该公司今年已经实现千万级别的季度营收。中科驭数的下一代DPU芯片预计将于2021年底完成设计，预计可处理高达200G网络带宽数据。\n不过，DPU市场虽然火爆，但概念较新，未知更多，投资风险也会更大。\n芯启源CEO卢笙指出，目前DPU细分赛道的壁垒还是相对较高的，除了技术壁垒之外，还有市场的壁垒，需要客户不断迭代，尤其是配合开源软件不断升级去适配客户快速变化的软硬件环境。因此VC（风险投资）在投资之前，一定要先认可赛道，且有足够的耐心。他强调，投资人需要对市场进行不断地观察并调整判断，现在谁也无法预料未来DPU发展前景。\n也有媒体认为，当英伟达进入新开辟的CPU和DPU战场，对中国的GPU厂商或许是个利好，尤其英伟达依然花大量精力放在400亿美元收购英国芯片设计商Arm公司的并购交易上，这对新创GPU企业而言，可能是个追赶的时机。\n正如章高男对钛媒体App所说，“从逻辑上讲，门槛不高的事情通常稀缺性都不高。（芯片半导体赛道）有些事情是很难的，需要长时间投入，虽然是高风险，但总归得有人去做。这是真正对国家有利的长远投入，其实应该鼓励投资。否则的话，这些需要长时间投入的难事，谁都不去做，你永远上不了台阶。”\n章高男强调，虽然风险投资肯定要追求回报，但他认为，在整个资金分配合理情况下，拿出一部分投资半导体赛道的初创企业，不仅有极强的社会意义，更是某种长期价值投资的重要体现。","news_type":1,"symbols_score_info":{"NVDA":0.9}},"isVote":1,"tweetType":1,"viewCount":4066,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":889800809,"gmtCreate":1631122579301,"gmtModify":1676530474613,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Omg","listText":"Omg","text":"Omg","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":7,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/889800809","repostId":"2165994363","repostType":4,"repost":{"id":"2165994363","kind":"highlight","pubTimestamp":1631085017,"share":"https://ttm.financial/m/news/2165994363?lang=en_US&edition=fundamental","pubTime":"2021-09-08 15:10","market":"hk","language":"zh","title":"Multiple Indicators in the United States Alarm!","url":"https://stock-news.laohu8.com/highlight/detail?id=2165994363","media":"华尔街见闻","summary":"美国人曾憧憬，这个夏天经济能回归常态，上班族回到办公室，孩子们重返校园，街角的咖啡店再次开业。但一切事与愿违，随着疫情复燃，复苏在8月踩下了急刹车。\n上周五，让人大跌眼镜的非农就业数据带来了最直接的报","content":"<p>Americans had envisioned a summer when the economy would return to normal, with office workers returning to the office, kids returning to school and corner coffee shops opening again. But everything backfired, and with the resurgence of the pandemic, the recovery stepped on the brakes in August.</p><p><b>Last Friday, the shocking data on non-farm payrolls brought the most direct alarm signal.</b>U.S. nonfarm payrolls increased by just 235,000 in August, a significant miss the market expectation of 733,000, down from about 1 million in June and July, the smallest increase since January 2021.</p><p>The University of Michigan's consumer confidence index also fell to its lowest level in 10 years in August, with Americans worried by the spreading Delta variant and persistently rising inflation.</p><p>According to data released by Johns Hopkins University in the United States, as of 6 p.m. ET on September 7, 2021, a total of 40,238,083 confirmed cases of COVID-19 and 650,345 deaths were reported across the United States.</p><p><b>Over the past week, the U.S. has averaged more than 161,000 new cases a day, a whopping 1,560 new deaths and an average of more than 102,000 daily hospitalizations, only slightly below last winter's peak.</b></p><p><img src=\"https://static.tigerbbs.com/f7c5b880f937edae73659acd8fa2e181\" tg-width=\"731\" tg-height=\"455\" referrerpolicy=\"no-referrer\"></p><p>(Image credit: Worldometers)</p><p><h2>'Feelings of uncertainty and anxiety are back'</h2>Amid the resurgence of the pandemic, U.S. offices and schools have been delayed, and travel and performance plans have been canceled.</p><p><b>Beginning in August, including<a href=\"https://laohu8.com/S/AAPL\">Apple</a>、<a href=\"https://laohu8.com/S/AMZN\">Amazon</a>Companies, large and small, have abandoned plans to reopen offices, and some have even pushed back their return dates to 2022.</b></p><p>Visitor arrivals to Hawaii have recovered to near pre-pandemic levels in mid-July, down only about 10% from the same time in 2019, official figures show. But starting in August, the pace has slowed somewhat, and in the last seven days of August, average daily visitor arrivals were down 34% from 2019.</p><p>Theater footfall at the end of August more than halved compared to the pandemic peak in mid-July, according to market research firm TOP. Paramount Pictures has delayed the release of Top Gun: Maverick and Mission: Impossible 7.</p><p>Meanwhile, many schools in the United States have closed or resumed online instruction.</p><p>The Wall Street Journal quoted Columbia Business School economist Stephan Meier as saying the reopening of schools was supposed to be a big moment for the economy. About a quarter of families have school-aged children, and reliable child care can get many Americans, especially women, back to work. But for now, the Delta variant and the lack of a vaccine for children under the age of 12 may keep some parents from working away from home.</p><p><b>\"Our sense of uncertainty and anxiety from last year is back, and that uncertainty is enough to dampen labor supply,\" Meier said.</b></p><p><h2>Institutions are cutting U.S. growth forecasts</h2>Economists believe the Delta variant virus won't push the U.S. back into recession, but losing growth momentum could prolong the recovery of millions of jobs lost during the pandemic, leaving many job markets on the sidelines and discouraging companies from investing amid new uncertainty.</p><p>In the report released on Monday,<a href=\"https://laohu8.com/S/GS\">Goldman Sachs</a>This is the third time in less than three months that Goldman Sachs has lowered its U.S. GDP forecast this year to 5.7%, and the unemployment rate this year is expected to rise from 4.1% to 4.2%.</p><p>In explaining why they cut their forecasts this week, Goldman Sachs economists cited several major factors,<b>It is expected that due to the raging virus variant Delta, weakening government financial support, and the shift in demand from goods to services, American consumers may reduce spending, and the consumption situation is more severe than previously expected:</b></p><p>\"The obstacles to strong consumption growth in the future seem to be much more: the Delta variant virus is already pressuring growth in the third quarter, fiscal stimulus is decreasing, and the recovery of the service industry is slowing down, which will all be negative factors in the medium term.\" Earlier this month,<a href=\"https://laohu8.com/S/MS\">Morgan Stanley</a>It also sharply lowered the U.S. GDP forecast in the third quarter, from 6.5% to only 2.9%, and the GDP forecast in the fourth quarter remained at 6.7%. After revising the GDP forecast for the third quarter, Morgan Stanley's GDP growth rate for the United States this year is expected to be 5.6%, which is lower than the expected growth rate for the whole year after Goldman Sachs lowered this week.</p><p>Morgan Stanley said at the time,<b>The reduction is mainly due to the fact that the driving force of economic growth has been released ahead of schedule, the stimulus expenditure of government has decreased to the economy, and the bottleneck of supply chain continues to drag down the economy, resulting in a decrease in consumer expenditure on large-scale durable goods such as automobiles.</b></p><p><h2>Two consecutive losses in U.S. stocks, a bigger test is behind</h2>On Tuesday, the Dow dropped more than 200 points, hitting a new closing low since Aug. 19 and its biggest closing drop since Aug. 18. The S&P closed down 0.34% at 4520.03 points, a new low since Aug. 27.</p><p>This is the second consecutive day that the Dow and S&P closed lower after the non-farm payrolls report, which was far worse than expected, was released last Friday. Even the three-day long weekend, including Monday, failed to ease the decline of most sectors of U.S. stocks. The prices of U.S. stocks and U.S. bonds both fell during the session, the yield of Treasury Bond rose, and the benchmark 10-year U.S. bond yield reached a new high since mid-July.</p><p>The analysis pointed out that under the raging Delta variant virus, the market is reevaluating the economic growth prospects of the United States and its impact on corporate earnings.</p><p><b>With the gradual withdrawal of fiscal stimulus measures in the United States, the economic recovery in the United States will face further tests.</b></p><p>More than 7.5 million Americans will lose $300 a week in additional unemployment benefits during the pandemic this week.<b>In addition to halting pandemic subsidies, risk factors such as tax increases and debt hikes have created additional tests for the U.S. economy.</b></p><p>Wall Street News mentioned earlier that although it seems likely that the US Congress will pass an infrastructure bill this autumn, compared with the multiple rounds of stimulus plans launched since March 2020, the economic impact of the recent stimulus plan is relatively limited.</p><p>The U.S. Congress could vote to raise taxes on businesses and high-income individuals, which would offset the boost from spending — another short-term risk to markets.</p><p>At the same time, U.S. stocks may also be affected by the increase in the debt ceiling this autumn. The U.S. Congress needs to pass a debt ceiling increase later this month to raise government funds, as well as a temporary spending bill to avoid a Washington shutdown in October.</p>","source":"wallstreetcn_api","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>Multiple Indicators in the United States Alarm!</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: 12.5px; color: #7E829C; margin: 0;}\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\">\nMultiple Indicators in the United States Alarm!\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">华尔街见闻</strong><span class=\"h-time small\">2021-09-08 15:10</span>\n</p>\n</h4>\n</header>\n<article>\n<p>Americans had envisioned a summer when the economy would return to normal, with office workers returning to the office, kids returning to school and corner coffee shops opening again. But everything backfired, and with the resurgence of the pandemic, the recovery stepped on the brakes in August.</p><p><b>Last Friday, the shocking data on non-farm payrolls brought the most direct alarm signal.</b>U.S. nonfarm payrolls increased by just 235,000 in August, a significant miss the market expectation of 733,000, down from about 1 million in June and July, the smallest increase since January 2021.</p><p>The University of Michigan's consumer confidence index also fell to its lowest level in 10 years in August, with Americans worried by the spreading Delta variant and persistently rising inflation.</p><p>According to data released by Johns Hopkins University in the United States, as of 6 p.m. ET on September 7, 2021, a total of 40,238,083 confirmed cases of COVID-19 and 650,345 deaths were reported across the United States.</p><p><b>Over the past week, the U.S. has averaged more than 161,000 new cases a day, a whopping 1,560 new deaths and an average of more than 102,000 daily hospitalizations, only slightly below last winter's peak.</b></p><p><img src=\"https://static.tigerbbs.com/f7c5b880f937edae73659acd8fa2e181\" tg-width=\"731\" tg-height=\"455\" referrerpolicy=\"no-referrer\"></p><p>(Image credit: Worldometers)</p><p><h2>'Feelings of uncertainty and anxiety are back'</h2>Amid the resurgence of the pandemic, U.S. offices and schools have been delayed, and travel and performance plans have been canceled.</p><p><b>Beginning in August, including<a href=\"https://laohu8.com/S/AAPL\">Apple</a>、<a href=\"https://laohu8.com/S/AMZN\">Amazon</a>Companies, large and small, have abandoned plans to reopen offices, and some have even pushed back their return dates to 2022.</b></p><p>Visitor arrivals to Hawaii have recovered to near pre-pandemic levels in mid-July, down only about 10% from the same time in 2019, official figures show. But starting in August, the pace has slowed somewhat, and in the last seven days of August, average daily visitor arrivals were down 34% from 2019.</p><p>Theater footfall at the end of August more than halved compared to the pandemic peak in mid-July, according to market research firm TOP. Paramount Pictures has delayed the release of Top Gun: Maverick and Mission: Impossible 7.</p><p>Meanwhile, many schools in the United States have closed or resumed online instruction.</p><p>The Wall Street Journal quoted Columbia Business School economist Stephan Meier as saying the reopening of schools was supposed to be a big moment for the economy. About a quarter of families have school-aged children, and reliable child care can get many Americans, especially women, back to work. But for now, the Delta variant and the lack of a vaccine for children under the age of 12 may keep some parents from working away from home.</p><p><b>\"Our sense of uncertainty and anxiety from last year is back, and that uncertainty is enough to dampen labor supply,\" Meier said.</b></p><p><h2>Institutions are cutting U.S. growth forecasts</h2>Economists believe the Delta variant virus won't push the U.S. back into recession, but losing growth momentum could prolong the recovery of millions of jobs lost during the pandemic, leaving many job markets on the sidelines and discouraging companies from investing amid new uncertainty.</p><p>In the report released on Monday,<a href=\"https://laohu8.com/S/GS\">Goldman Sachs</a>This is the third time in less than three months that Goldman Sachs has lowered its U.S. GDP forecast this year to 5.7%, and the unemployment rate this year is expected to rise from 4.1% to 4.2%.</p><p>In explaining why they cut their forecasts this week, Goldman Sachs economists cited several major factors,<b>It is expected that due to the raging virus variant Delta, weakening government financial support, and the shift in demand from goods to services, American consumers may reduce spending, and the consumption situation is more severe than previously expected:</b></p><p>\"The obstacles to strong consumption growth in the future seem to be much more: the Delta variant virus is already pressuring growth in the third quarter, fiscal stimulus is decreasing, and the recovery of the service industry is slowing down, which will all be negative factors in the medium term.\" Earlier this month,<a href=\"https://laohu8.com/S/MS\">Morgan Stanley</a>It also sharply lowered the U.S. GDP forecast in the third quarter, from 6.5% to only 2.9%, and the GDP forecast in the fourth quarter remained at 6.7%. After revising the GDP forecast for the third quarter, Morgan Stanley's GDP growth rate for the United States this year is expected to be 5.6%, which is lower than the expected growth rate for the whole year after Goldman Sachs lowered this week.</p><p>Morgan Stanley said at the time,<b>The reduction is mainly due to the fact that the driving force of economic growth has been released ahead of schedule, the stimulus expenditure of government has decreased to the economy, and the bottleneck of supply chain continues to drag down the economy, resulting in a decrease in consumer expenditure on large-scale durable goods such as automobiles.</b></p><p><h2>Two consecutive losses in U.S. stocks, a bigger test is behind</h2>On Tuesday, the Dow dropped more than 200 points, hitting a new closing low since Aug. 19 and its biggest closing drop since Aug. 18. The S&P closed down 0.34% at 4520.03 points, a new low since Aug. 27.</p><p>This is the second consecutive day that the Dow and S&P closed lower after the non-farm payrolls report, which was far worse than expected, was released last Friday. Even the three-day long weekend, including Monday, failed to ease the decline of most sectors of U.S. stocks. The prices of U.S. stocks and U.S. bonds both fell during the session, the yield of Treasury Bond rose, and the benchmark 10-year U.S. bond yield reached a new high since mid-July.</p><p>The analysis pointed out that under the raging Delta variant virus, the market is reevaluating the economic growth prospects of the United States and its impact on corporate earnings.</p><p><b>With the gradual withdrawal of fiscal stimulus measures in the United States, the economic recovery in the United States will face further tests.</b></p><p>More than 7.5 million Americans will lose $300 a week in additional unemployment benefits during the pandemic this week.<b>In addition to halting pandemic subsidies, risk factors such as tax increases and debt hikes have created additional tests for the U.S. economy.</b></p><p>Wall Street News mentioned earlier that although it seems likely that the US Congress will pass an infrastructure bill this autumn, compared with the multiple rounds of stimulus plans launched since March 2020, the economic impact of the recent stimulus plan is relatively limited.</p><p>The U.S. Congress could vote to raise taxes on businesses and high-income individuals, which would offset the boost from spending — another short-term risk to markets.</p><p>At the same time, U.S. stocks may also be affected by the increase in the debt ceiling this autumn. The U.S. Congress needs to pass a debt ceiling increase later this month to raise government funds, as well as a temporary spending bill to avoid a Washington shutdown in October.</p>\n<div class=\"bt-text\">\n\n\n<p> source：<a href=\"https://wallstreetcn.com/articles/3639891\">华尔街见闻</a></p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","type":0,"thumbnail":"https://static.tigerbbs.com/627bc890436e46f74a0fe8143398a725","relate_stocks":{"159934":"黄金ETF","518880":"黄金ETF华安","NUGT":"二倍做多黄金矿业指数ETF-Direxion","IAU":"黄金信托ETF-iShares","GDX":"黄金矿业ETF-VanEck","DUST":"二倍做空黄金矿业指数ETF-Direxion","GLD":"黄金ETF-SPDR"},"source_url":"https://wallstreetcn.com/articles/3639891","is_english":false,"share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","article_id":"2165994363","content_text":"美国人曾憧憬，这个夏天经济能回归常态，上班族回到办公室，孩子们重返校园，街角的咖啡店再次开业。但一切事与愿违，随着疫情复燃，复苏在8月踩下了急刹车。\n上周五，让人大跌眼镜的非农就业数据带来了最直接的报警信号。美国8月非农就业人口仅增加23.5万人，大幅不及市场预期的73.3万人，低于6月和7月的约100万个，创2021年1月以来最小增幅。\n8月，密歇根大学消费者信心指数也跌至10年来的最低水平，不断蔓延的Delta变种病毒和持续上升的通胀令美国人感到担忧。\n据美国约翰斯·霍普金斯大学发布的数据，截至美国东部时间2021年9月7日下午6点，全美共报告新冠肺炎确诊40238083例，死亡650345例。\n过去一周，美国平均每天新增病例超过 161000 例，新增死亡病例高达 1560 例，平均每天住院人数超过 102000 人，仅比去年冬天的峰值略低。\n\n(图片来源：Worldometers)\n“不确定性和焦虑感又回来了”\n疫情复燃之下，美国办公室和学校推迟开放，旅行和演出计划纷纷取消。\n从8月份开始，包括苹果、亚马逊在内的大小的公司都放弃了重新开放办公室的计划，一些公司甚至将返回日期推迟到2022年。\n官方数据显示，7月中旬，前往夏威夷的游客人数已经恢复至接近疫情前的水平，仅比2019年同期下降了约10%。但从8月开始，速度有所放缓，在8月的最后7天，日均游客抵达数量比2019年下降了34%。\n市场研究公司TOP的数据显示，与7月中旬的疫情高峰相比，8月底的影院客流量下降了一半以上。派拉蒙影业公司推迟了《壮志凌云:特立独行》和《碟中谍7》的上映。\n与此同时，美国许多学校已经关闭或恢复线上授课。\n华尔街日报援引哥伦比亚商学院经济学家Stephan Meier表示，学校重新开学本应是经济的一个重要时刻。大约四分之一的家庭有学龄儿童，可靠的儿童看护可以让许多美国人，尤其是女性重返工作岗位。但目前，Delta变种和缺乏针对12岁以下儿童的疫苗可能会让一些父母不愿离家工作。\nMeier称：“我们去年的不确定性和焦虑感又回来了，这种不确定性足以抑制劳动力供应。”\n机构纷纷下调美国经济增长预期\n经济学家认为，Delta变种病毒不会将美国推回衰退，但失去增长动力可能延长疫情期间失去的数百万就业岗位的复苏，令许多就业市场处于观望状态，并阻碍企业在新的不确定性下投资。\n本周一发布的报告中，高盛将今年的美国GDP预期增速降至5.7%，今年失业率预期从4.1%升至4.2%，这是不到三个月里，高盛第三次下调今年美国GDP预期。\n在本周解释为何下调预期时，高盛经济学家提到几大影响因素，预计由于变种病毒Delta肆虐、政府的财政支持在减弱，加之需求从商品转换到服务，美国消费者可能减少支出，消费形势比之前预期的更严峻：\n\n “未来消费强劲增长的阻碍看来多得多：Delta变异病毒已经在施压三季度增长，财政刺激在减少，服务业复苏放缓，这些都将是中期内的负面因素。”\n\n本月初，摩根士丹利也大幅下调了三季度美国GDP预期，从6.5%猛砍至仅2.9%，四季度GDP预期维持在6.7%。下修三季度GDP预期后，摩根士丹利的美国今年全年GDP增速预期为5.6%，比高盛本周调降后的全年预期增速还低。\n摩根士丹利当时称，调降预期主要由于经济增长的动力已提前释放，政府刺激性支出对经济的推动减少，加之供应链瓶颈持续拖累经济，汽车等大件耐用品的消费者支出由此减少。\n美股两连跌，更大的考验在后面\n周二，道指跌超200点，创8月19日以来收盘新低以及8月18日以来最大收盘跌幅。标普收跌0.34%，报4520.03点，创8月27日以来新低。\n这是上周五远逊预期的非农就业报告公布后，道指、标普连续两日收跌。即使是包括周一在内的三日长周末也没能缓和美股多数板块的跌势，美股和美债价格盘中齐跌，国债收益率上行，基准10年期美债收益率刷新7月中旬以来高位。\n分析指出，Delta变种病毒肆虐之下，市场正在重新评估美国经济增长前景，以及其对企业盈利的影响。\n随着美国财政刺激措施逐渐退场，美国经济复苏将面临进一步的考验。\n本周，超过750万的美国人将失去疫情时期每周300美元的额外失业救济。除了停止疫情补贴外，增税、债务上调等风险因素也给美国经济带来了额外的考验。\n华尔街见闻此前提及，尽管美国国会似乎有可能在今年秋季通过一项基础设施法案，但与2020年3月以来推出的多轮刺激计划相比，近期的刺激计划经济影响相对有限。\n美国国会可能会投票决定提高对企业和高收入个人的税收，这将抵消支出带来的提振——这是市场面临的另一个短期风险。\n与此同时，美股今秋还可能受到债务上限上调的影响。美国国会需要在本月晚些时候通过提高债务上限以筹集政府资金，以及一项临时开支法案，以避免华盛顿在10月份关闭。","news_type":1,"symbols_score_info":{"159934":0.9,"518880":0.9,"SGCmain":0.9,"DUST":0.9,"SGUmain":0.9,"GDX":0.9,"NUGT":0.9,"GLD":0.9,"IAU":0.9,"MGCmain":0.9}},"isVote":1,"tweetType":1,"viewCount":4473,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":834684230,"gmtCreate":1629797562520,"gmtModify":1676530134416,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Good&nbsp;","listText":"Good&nbsp;","text":"Good","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":1,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/834684230","repostId":"1140946559","repostType":4,"isVote":1,"tweetType":1,"viewCount":3811,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":835240012,"gmtCreate":1629723743636,"gmtModify":1676530111553,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Hi&nbsp;","listText":"Hi&nbsp;","text":"Hi","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":4,"commentSize":1,"repostSize":0,"link":"https://ttm.financial/post/835240012","repostId":"1197421899","repostType":4,"isVote":1,"tweetType":1,"viewCount":5769,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":835257533,"gmtCreate":1629723694922,"gmtModify":1676530111545,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"Good&nbsp;","listText":"Good&nbsp;","text":"Good","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":1,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/835257533","repostId":"1145459282","repostType":4,"isVote":1,"tweetType":1,"viewCount":3926,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0},{"id":832049713,"gmtCreate":1629548892632,"gmtModify":1676530069184,"author":{"id":"3578895363297107","authorId":"3578895363297107","name":"TIDARAT","avatar":"https://static.tigerbbs.com/70741f8f742a273f9c4a9f268fca7aa3","crmLevel":11,"crmLevelSwitch":0,"followedFlag":false,"idStr":"3578895363297107","authorIdStr":"3578895363297107"},"themes":[],"title":"","htmlText":"555","listText":"555","text":"555","images":[],"top":1,"highlighted":1,"essential":1,"paper":1,"likeSize":1,"commentSize":0,"repostSize":0,"link":"https://ttm.financial/post/832049713","repostId":"2161422746","repostType":2,"isVote":1,"tweetType":1,"viewCount":3859,"authorTweetTopStatus":1,"verified":2,"comments":[],"imageCount":0,"langContent":"EN","totalScore":0}],"defaultTab":"posts","isTTM":true}