Miller
04-15
服务器市场:Tesla在孵化FSD,Google在孵化自己的LLM,Meta在孵化虚拟现实和LLM,微软和亚马逊都在构建基于AI的云服务……………不要质疑黄老板强劲的需求。我还没说政府的军备采购,以色利和乌克兰已经实地演示了AI时代下现代战争的军备,那些无人机、自寻址导弹,背后都是对高效算力的需求! 美国的国运都在人工智能上了,这股力量不以你我的个人意志为转移。
@安和昂:
$英伟达(NVDA)$ $台积电(TSM)$ 周末和几个AI、IC行业的朋友聊了一下,达成共识。 NVDA如此高的估值建立在两个条件上: 1.未来AI计算需求达到数千亿美元级 2.NVDA能垄断AI算力市场。 先假设条件1为真,条件2我们认为很难实现。 先上结论:长期看跌NVDA 主要是因为NV的CUDA护城河在B端并没有那么坚固。现在愿意花重金买皮衣黄的金子GPU,是急于在训练阶段跑马圈地。 等到了推理算力需求占主要的阶段,大型企业有能力且有意愿使用性价比更高的替代品,一块H100成本小于3000美元,为什么要花10倍价格给黄老板当冤大头呢? 如AMD的MI300系列+ROCm、谷歌的TPU+Tensorflow、国内华为升腾+Mindspore、Groq的LPU,微软和meta也在着手开发自己的AI芯片。而除华为,这些芯片订单绝大部分都会交给台积电代工,这也是为什么我们一致看好$台积电(TSM)$ 的原因。 黄老板为了吸引客户购买更多他们家的GPU,唯一的方法就是继续推出“性价比更高”的GPU。 但从另一种角度讲,若AI算力需求主体长期处于算法不断推陈出新的训练阶段,NVDA会依然会挣走绝大部分money~ 至于AI的C端需求和$阿斯麦(ASML)$ ,有空写篇文章细说。
免责声明:上述内容仅代表发帖人个人观点,不构成本平台的任何投资建议。
分享至
微信
复制链接
精彩评论
杰瑞冲冲冲
04-16
杰瑞冲冲冲
🔥
什么也没有了~
APP内打开
发表看法
1
3
{"i18n":{"language":"zh_CN"},"detailType":1,"isChannel":false,"data":{"magic":2,"id":295681274466608,"tweetId":"295681274466608","gmtCreate":1713196771244,"gmtModify":1713235927550,"author":{"id":11307867536043,"idStr":"11307867536043","authorId":11307867536043,"authorIdStr":"11307867536043","name":"Miller","avatar":"https://static.laohu8.com/cee66e75fe7e4ea41b12cf6399c33dcf","vip":1,"userType":1,"introduction":"","boolIsFan":false,"boolIsHead":false,"crmLevel":6,"crmLevelSwitch":0,"currentWearingBadge":{"badgeId":"e50ce593bb40487ebfb542ca54f6a561-4","templateUuid":"e50ce593bb40487ebfb542ca54f6a561","name":"明星虎友","description":"加入老虎社区2000天","bigImgUrl":"https://static.tigerbbs.com/dddf24b906c7011de2617d4fb3f76987","smallImgUrl":"https://static.tigerbbs.com/53d58ad32c97254c6f74db8b97e6ec49","grayImgUrl":"https://static.tigerbbs.com/6304700d92ad91c7a33e2e92ec32ecc1","redirectLinkEnabled":0,"hasAllocated":1,"isWearing":1,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2021.12.18","individualDisplayEnabled":0},"individualDisplayBadges":[],"fanSize":1254,"starInvestorFlag":false},"themes":[],"images":[],"coverImages":[],"html":"<html><head></head><body>服务器市场:Tesla在孵化FSD,Google在孵化自己的LLM,Meta在孵化虚拟现实和LLM,微软和亚马逊都在构建基于AI的云服务……………不要质疑黄老板强劲的需求。我还没说政府的军备采购,以色利和乌克兰已经实地演示了AI时代下现代战争的军备,那些无人机、自寻址导弹,背后都是对高效算力的需求! 美国的国运都在人工智能上了,这股力量不以你我的个人意志为转移。</body></html>","htmlText":"<html><head></head><body>服务器市场:Tesla在孵化FSD,Google在孵化自己的LLM,Meta在孵化虚拟现实和LLM,微软和亚马逊都在构建基于AI的云服务……………不要质疑黄老板强劲的需求。我还没说政府的军备采购,以色利和乌克兰已经实地演示了AI时代下现代战争的军备,那些无人机、自寻址导弹,背后都是对高效算力的需求! 美国的国运都在人工智能上了,这股力量不以你我的个人意志为转移。</body></html>","text":"服务器市场:Tesla在孵化FSD,Google在孵化自己的LLM,Meta在孵化虚拟现实和LLM,微软和亚马逊都在构建基于AI的云服务……………不要质疑黄老板强劲的需求。我还没说政府的军备采购,以色利和乌克兰已经实地演示了AI时代下现代战争的军备,那些无人机、自寻址导弹,背后都是对高效算力的需求! 美国的国运都在人工智能上了,这股力量不以你我的个人意志为转移。","highlighted":1,"essential":1,"paper":1,"likeSize":3,"commentSize":1,"repostSize":0,"favoriteSize":0,"link":"https://laohu8.com/post/295681274466608","repostId":282492482809976,"repostType":1,"repost":{"magic":2,"id":282492482809976,"tweetId":"282492482809976","gmtCreate":1709996336511,"gmtModify":1710042538623,"gmtUserModify":1710034011389,"author":{"id":4172278615319592,"idStr":"4172278615319592","authorId":4172278615319592,"authorIdStr":"4172278615319592","name":"安和昂","avatar":"https://static.tigerbbs.com/14fd583c36b4d1429436c40219c04456","vip":1,"userType":1,"introduction":"","boolIsFan":false,"boolIsHead":false,"crmLevel":1,"crmLevelSwitch":1,"currentWearingBadge":{"badgeId":"1f74612b6f95455f970edd02b5316db8-6","templateUuid":"1f74612b6f95455f970edd02b5316db8","name":"收益排行","description":"Top 3%","bigImgUrl":"https://static.tigerbbs.com/6fbaa8e552e02eae8e6a2ce5d6566431","smallImgUrl":"https://static.tigerbbs.com/98862e418516dab972bed7aaf561dec6","grayImgUrl":"https://static.tigerbbs.com/3a5ca933a5695ed2e862b228f6c09ee5","redirectLinkEnabled":0,"redirectLink":"https://www.itigerup.com/activity/market/2023/stock-contest?adcode=StockContest2023","hasAllocated":1,"isWearing":1,"stampPosition":0,"hasStamp":0,"allocationCount":1,"allocatedDate":"2024.03.09","individualDisplayEnabled":0},"individualDisplayBadges":[],"fanSize":7,"starInvestorFlag":false},"themes":[{"themeId":"e6c3b9d055e246edb3c48a9d63d9e0ac","categoryId":"d316d8bd4a714fab9ea15fd7157c45e7","name":"英伟达股价下跌,绩前能回调到多少价位?","type":3,"rnLink":"https://laohu8.com/RN?name=RNTheme&page=/theme/special/discussion&rndata={\"themeId\":e6c3b9d055e246edb3c48a9d63d9e0ac}&rnconfig={\"headerBarHidden\": true}","description":"美股三大指数集体收跌,英伟达跌近4%,【下周英伟达发布财报,你希望绩前英伟达回调到什么价位?本次财报后英伟达能否大涨?】","image":"https://static.tigerbbs.com/f40d4c9ce2b914622d5bc108dd736592"}],"images":[],"coverImages":[],"html":"<html><head></head><body><p><a href=\"https://laohu8.com/S/NVDA\">$英伟达(NVDA)$ </a><v-v data-views=\"0\"></v-v> <a href=\"https://laohu8.com/S/TSM\">$台积电(TSM)$ </a><v-v data-views=\"1\"></v-v> 周末和几个AI、IC行业的朋友聊了一下,达成共识。</p>\n<p>NVDA如此高的估值建立在两个条件上:</p>\n<p>1.未来AI计算需求达到数千亿美元级</p>\n<p>2.NVDA能垄断AI算力市场。</p>\n<p>先假设条件1为真,条件2我们认为很难实现。</p>\n<p>先上结论:长期看跌NVDA</p>\n<p>主要是因为NV的CUDA护城河在B端并没有那么坚固。现在愿意花重金买皮衣黄的金子GPU,是急于在训练阶段跑马圈地。</p>\n<p>等到了推理算力需求占主要的阶段,大型企业有能力且有意愿使用性价比更高的替代品,一块H100成本小于3000美元,为什么要花10倍价格给黄老板当冤大头呢? 如AMD的MI300系列+ROCm、谷歌的TPU+Tensorflow、国内华为升腾+Mindspore、Groq的LPU,微软和meta也在着手开发自己的AI芯片。而除华为,这些芯片订单绝大部分都会交给台积电代工,这也是为什么我们一致看好<a href=\"https://laohu8.com/S/TSM\">$台积电(TSM)$ </a> 的原因。</p>\n<p>黄老板为了吸引客户购买更多他们家的GPU,唯一的方法就是继续推出“性价比更高”的GPU。</p>\n<p>但从另一种角度讲,若AI算力需求主体长期处于算法不断推陈出新的训练阶段,NVDA会依然会挣走绝大部分money~</p>\n<p>至于AI的C端需求和<a href=\"https://laohu8.com/S/ASML\">$阿斯麦(ASML)$ </a> ,有空写篇文章细说。</p></body></html>","htmlText":"<html><head></head><body><p><a href=\"https://laohu8.com/S/NVDA\">$英伟达(NVDA)$ </a><v-v data-views=\"0\"></v-v> <a href=\"https://laohu8.com/S/TSM\">$台积电(TSM)$ </a><v-v data-views=\"1\"></v-v> 周末和几个AI、IC行业的朋友聊了一下,达成共识。</p>\n<p>NVDA如此高的估值建立在两个条件上:</p>\n<p>1.未来AI计算需求达到数千亿美元级</p>\n<p>2.NVDA能垄断AI算力市场。</p>\n<p>先假设条件1为真,条件2我们认为很难实现。</p>\n<p>先上结论:长期看跌NVDA</p>\n<p>主要是因为NV的CUDA护城河在B端并没有那么坚固。现在愿意花重金买皮衣黄的金子GPU,是急于在训练阶段跑马圈地。</p>\n<p>等到了推理算力需求占主要的阶段,大型企业有能力且有意愿使用性价比更高的替代品,一块H100成本小于3000美元,为什么要花10倍价格给黄老板当冤大头呢? 如AMD的MI300系列+ROCm、谷歌的TPU+Tensorflow、国内华为升腾+Mindspore、Groq的LPU,微软和meta也在着手开发自己的AI芯片。而除华为,这些芯片订单绝大部分都会交给台积电代工,这也是为什么我们一致看好<a href=\"https://laohu8.com/S/TSM\">$台积电(TSM)$ </a> 的原因。</p>\n<p>黄老板为了吸引客户购买更多他们家的GPU,唯一的方法就是继续推出“性价比更高”的GPU。</p>\n<p>但从另一种角度讲,若AI算力需求主体长期处于算法不断推陈出新的训练阶段,NVDA会依然会挣走绝大部分money~</p>\n<p>至于AI的C端需求和<a href=\"https://laohu8.com/S/ASML\">$阿斯麦(ASML)$ </a> ,有空写篇文章细说。</p></body></html>","text":"$英伟达(NVDA)$ $台积电(TSM)$ 周末和几个AI、IC行业的朋友聊了一下,达成共识。 NVDA如此高的估值建立在两个条件上: 1.未来AI计算需求达到数千亿美元级 2.NVDA能垄断AI算力市场。 先假设条件1为真,条件2我们认为很难实现。 先上结论:长期看跌NVDA 主要是因为NV的CUDA护城河在B端并没有那么坚固。现在愿意花重金买皮衣黄的金子GPU,是急于在训练阶段跑马圈地。 等到了推理算力需求占主要的阶段,大型企业有能力且有意愿使用性价比更高的替代品,一块H100成本小于3000美元,为什么要花10倍价格给黄老板当冤大头呢? 如AMD的MI300系列+ROCm、谷歌的TPU+Tensorflow、国内华为升腾+Mindspore、Groq的LPU,微软和meta也在着手开发自己的AI芯片。而除华为,这些芯片订单绝大部分都会交给台积电代工,这也是为什么我们一致看好$台积电(TSM)$ 的原因。 黄老板为了吸引客户购买更多他们家的GPU,唯一的方法就是继续推出“性价比更高”的GPU。 但从另一种角度讲,若AI算力需求主体长期处于算法不断推陈出新的训练阶段,NVDA会依然会挣走绝大部分money~ 至于AI的C端需求和$阿斯麦(ASML)$ ,有空写篇文章细说。","highlighted":1,"essential":2,"paper":1,"link":"https://laohu8.com/post/282492482809976","repostId":0,"isVote":1,"tweetType":1,"commentLimit":10,"symbols":["NVDA","TSM","ASML"],"verified":2,"subType":0,"readableState":1,"langContent":"CN","currentLanguage":"CN","warmUpFlag":false,"orderFlag":false,"shareable":true,"causeOfNotShareable":"","featuresForAnalytics":[],"commentAndTweetFlag":false,"andRepostAutoSelectedFlag":false,"upFlag":false,"length":884,"xxTargetLangEnum":"ZH_CN"},"isVote":1,"tweetType":1,"viewCount":8596,"commentLimit":10,"likeStatus":false,"favoriteStatus":false,"reportStatus":false,"symbols":["TSLA","GOOG"],"verified":2,"subType":0,"readableState":1,"langContent":"CN","currentLanguage":"CN","warmUpFlag":false,"orderFlag":false,"shareable":true,"causeOfNotShareable":"","featuresForAnalytics":[],"commentAndTweetFlag":false,"andRepostAutoSelectedFlag":false,"upFlag":true,"length":317,"xxTargetLangEnum":"ZH_CN"},"commentList":[{"id":295843232047280,"commentId":"295843232047280","gmtCreate":1713236232839,"gmtModify":1713341112896,"authorId":9000000000000290,"author":{"id":9000000000000290,"idStr":"9000000000000290","authorId":9000000000000290,"name":"杰瑞冲冲冲","avatar":"https://static.tigerbbs.com/6658c473a6cc65ee66ff79bb186e9aac","vip":8,"crmLevel":1,"crmLevelSwitch":0,"individualDisplayBadges":[]},"repliedAuthorId":0,"objectId":295681274466608,"objectIdStr":"295681274466608","type":1,"supId":0,"supIdStr":"0","prevId":0,"prevIdStr":"0","content":"🔥","text":"🔥","html":"🔥","likeSize":0,"commentSize":0,"subComments":[],"verified":10,"allocateAmount":0,"commentType":"invalid","coins":0,"score":0}],"isCommentEnd":false,"isTiger":false,"isWeiXinMini":false,"url":"/m/post/295681274466608"}
精彩评论