每秒110萬個token！微軟(MSFT.US)和英偉達(NVDA.US)聯手刷新AI推理紀錄

智通財經

11/04

智通財經APP獲悉，微軟(MSFT.US)宣佈，其Azure ND GB300v6虛擬機在Meta的Llama270B模型上實現了每秒推理速度達110萬token的行業新紀錄。據悉，Azure ND GB300虛擬機採用英偉達(NVDA.US)的Blackwell Ultra GPU，具體為NVIDIA GB300NVL72系統，配置72個NVIDIA Blackwell Ultra GPU和36個NVIDIA Grace CPU，採用單機架構設計。這款虛擬機專為推理工作負載優化，具有50%的GPU內存提升和16%的熱設計功率(TDP)提高。

微軟首席執行官薩提亞・納德拉在社交媒體上表示:「這一成就是我們與英偉達長期合作和在生產規模運行人工智能方面專業知識的結晶。」

資料顯示，為了驗證性能提升，微軟在一個NVIDIA GB300NVL72域下的18臺ND GB300v6虛擬機上運行了Llama270B(FP4精度)模型，使用NVIDIA TensorRT-LLM作為推理引擎。微軟表示:「一個NVL72機架的Azure ND GB300v6實現了總計110萬個token/秒的推理速度。」這一新記錄超越了微軟之前在NVIDIA GB200NVL72機架上實現的86.5萬個token/秒的成績。

對此，Signal65的實驗室副總裁拉斯・費洛斯指出:「這一里程碑不僅突破了每秒百萬token的障礙，還在一個能夠滿足現代企業動態使用和數據治理需求的平臺上實現。」他補充稱，Azure ND GB300相較於上一代NVIDIA GB200在推理性能上提升了27%，而僅增加了17%的功率規格。

免責聲明：投資有風險，本文並非投資建議，以上內容不應被視為任何金融產品的購買或出售要約、建議或邀請，作者或其他用戶的任何相關討論、評論或帖子也不應被視為此類內容。本文僅供一般參考，不考慮您的個人投資目標、財務狀況或需求。TTM對信息的準確性和完整性不承擔任何責任或保證，投資者應自行研究並在投資前尋求專業建議。

熱議股票

1
2
3
4
5
6
7
8
9
10

{"basename":"/hant","ssrTDKData":{"titleTemplate":"%s - 老虎證券","title":"老虎证券全球投资理财平台| 一站式投资美股新股港股A股","description":"老虎证券助您一站式投资美股，新股，港股，A股等全球金融理财产品。新加坡华人最信赖的在线投资平台，现在加入即享低费用，24/5 无时差炒美股投资理财！","keywords":"老虎證券,老虎證券開戶,老虎券商,老虎證券官網,老虎證券app,tigertrade老虎證券,股票,炒股,新加坡股票交易平臺,投資,投資理財","social":{"ogDescription":"老虎证券助您一站式投资美股，新股，港股，A股等全球金融理财产品。新加坡华人最信赖的在线投资平台，现在加入即享低费用，24/5 无时差炒美股投资理财！","ogImage":"https://c1.itigergrowtha.com/portal5/static/media/og-logo.be62fbe1.png","ogUrl":"https://www.itiger.com/hant/news/2580787639"},"companyName":"老虎證券"},"pageData":{"isMobile":false,"isTiger":false,"isTTM":true,"region":"SGP","license":"TBSG","edition":"fundamental"},"isCrawlerRequest":true,"__swrFallback__":{"@#url:\"https://stock-news.skytigris.cn/v3/news\",params:#id:\"2580787639\",edition:\"fundamental\",auth_exemption:1,,,undefined,":{"share":"https://ttm.financial/m/news/2580787639?lang=zh_TW&edition=fundamental","thumbnail":"","is_english":false,"pubTime":"2025-11-04 19:10","share_image_url":"https://static.laohu8.com/e9f99090a1c2ed51c021029395664489","id":"2580787639","market":"hk","top_or_hot":-1,"title":"每秒110萬個token！微軟(MSFT.US)和英偉達(NVDA.US)聯手刷新AI推理紀錄","media":"智通財經","content":"<html><body><p>智通財經APP獲悉，<strong><a href=\"https://laohu8.com/S/MSFT\">微軟</a>(MSFT.US)宣佈，其Azure ND GB300v6虛擬機在<a href=\"https://laohu8.com/S/META\">Meta</a>的Llama270B模型上實現了每秒推理速度達110萬token的行業新紀錄。</strong>據悉，Azure ND GB300虛擬機採用<a href=\"https://laohu8.com/S/NVDA\">英偉達</a>(NVDA.US)的Blackwell Ultra GPU，具體為NVIDIA GB300NVL72系統，配置72個NVIDIA Blackwell Ultra GPU和36個NVIDIA Grace CPU，採用單機架構設計。這款虛擬機專為推理工作負載優化，具有50%的GPU內存提升和16%的熱設計功率(TDP)提高。</p><p>微軟首席執行官薩提亞・納德拉在社交媒體上表示:「這一成就是我們與英偉達長期合作和在生產規模運行人工智能方面專業知識的結晶。」</p><p>資料顯示，為了驗證性能提升，微軟在一個NVIDIA GB300NVL72域下的18臺ND GB300v6虛擬機上運行了Llama270B(FP4精度)模型，使用NVIDIA TensorRT-LLM作為推理引擎。微軟表示:「一個NVL72機架的Azure ND GB300v6實現了總計110萬個token/秒的推理速度。」 這一新記錄超越了微軟之前在NVIDIA GB200NVL72機架上實現的86.5萬個token/秒的成績。</p><p>對此，Signal65的實驗室副總裁拉斯・費洛斯指出:「這一里程碑不僅突破了每秒百萬token的障礙，還在一個能夠滿足現代企業動態使用和數據治理需求的平臺上實現。」 他補充稱，Azure ND GB300相較於上一代NVIDIA GB200在推理性能上提升了27%，而僅增加了17%的功率規格。</p></body></html>","source":"stock_zhitongcaijing","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>每秒110萬個token！微軟(MSFT.US)和英偉達(NVDA.US)聯手刷新AI推理紀錄</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\">\n每秒110萬個token！微軟(MSFT.US)和英偉達(NVDA.US)聯手刷新AI推理紀錄\n</h2>\n<h4 class=\"meta\">\n<p class=\"head\">\n<strong class=\"h-name small\">智通財經</strong><span class=\"h-time small\">2025-11-04 19:10</span>\n</p>\n</h4>\n</header>\n<article>\n<html><body><p>智通財經APP獲悉，<strong><a href=\"https://laohu8.com/S/MSFT\">微軟</a>(MSFT.US)宣佈，其Azure ND GB300v6虛擬機在<a href=\"https://laohu8.com/S/META\">Meta</a>的Llama270B模型上實現了每秒推理速度達110萬token的行業新紀錄。</strong>據悉，Azure ND GB300虛擬機採用<a href=\"https://laohu8.com/S/NVDA\">英偉達</a>(NVDA.US)的Blackwell Ultra GPU，具體為NVIDIA GB300NVL72系統，配置72個NVIDIA Blackwell Ultra GPU和36個NVIDIA Grace CPU，採用單機架構設計。這款虛擬機專為推理工作負載優化，具有50%的GPU內存提升和16%的熱設計功率(TDP)提高。</p><p>微軟首席執行官薩提亞・納德拉在社交媒體上表示:「這一成就是我們與英偉達長期合作和在生產規模運行人工智能方面專業知識的結晶。」</p><p>資料顯示，為了驗證性能提升，微軟在一個NVIDIA GB300NVL72域下的18臺ND GB300v6虛擬機上運行了Llama270B(FP4精度)模型，使用NVIDIA TensorRT-LLM作為推理引擎。微軟表示:「一個NVL72機架的Azure ND GB300v6實現了總計110萬個token/秒的推理速度。」 這一新記錄超越了微軟之前在NVIDIA GB200NVL72機架上實現的86.5萬個token/秒的成績。</p><p>對此，Signal65的實驗室副總裁拉斯・費洛斯指出:「這一里程碑不僅突破了每秒百萬token的障礙，還在一個能夠滿足現代企業動態使用和數據治理需求的平臺上實現。」 他補充稱，Azure ND GB300相較於上一代NVIDIA GB200在推理性能上提升了27%，而僅增加了17%的功率規格。</p></body></html>\n<div class=\"bt-text\">\n\n\n<p> 來源：<a href=\"http://www.zhitongcaijing.com/content/detail/1365177.html\">智通財經</a></p>\n<p>為提升您的閱讀體驗，我們對本頁面進行了排版優化</p>\n\n\n</div>\n</article>\n</div>\n</body>\n</html>\n","isBrief":false,"type":0,"news_type":1,"symbol":"LU0985320562.USD","symbol_name":"NORDEA 1 GLOBAL STARS EQUITY \"BP\" (USD) ACC","start_time":0,"source_url":"http://www.zhitongcaijing.com/content/detail/1365177.html","article_id":"2580787639","we_media_id":null,"thumbnails":[],"rights":null,"url":"https://stock-news.laohu8.com/highlight/detail?id=2580787639","pubTimestamp":1762254609,"columns":[],"sourceInfo":{"source_id":"stock_zhitongcaijing","name":"智通财经网"},"weMediaInfo":null,"summary":"智通財經APP獲悉，微軟宣佈，其Azure ND GB300v6虛擬機在Meta的Llama270B模型上實現了每秒推理速度達110萬token的行業新紀錄。據悉，Azure ND GB300虛擬機採用英偉達的Blackwell Ultra GPU，具體為NVIDIA GB300NVL72系統，配置72個NVIDIA Blackwell Ultra GPU和36個NVIDIA Grace CPU，採用單機架構設計。資料顯示，為了驗證性能提升，微軟在一個NVIDIA GB300NVL72域下的18台ND GB300v6虛擬機上運行了Llama270B模型，使用NVIDIA TensorRT-LLM作為推理引擎。微軟表示:「一個NVL72機架的Azure ND GB300v6實現了總計110萬個token/秒的推理速度。」","collect":0,"end_time":0,"defaultTopTitle":"zhitongcaijing.com","property":[],"viewcount":null,"language":"zh","relate_stocks":{"LU0985320562.USD":"NORDEA 1 GLOBAL STARS EQUITY \"BP\" (USD) ACC","LU2491050071.SGD":"WELLINGTON SUSTAINABLE OUTCOMES \"A\" (SGDHDG) ACC","LU1670711123.USD":"M&G (LUX) GLOBAL DIVIDEND \"A\" (USD) INC","LU0171293334.USD":"贝莱德英国基金A2","NVDD":"Direxion Daily NVDA Bear 1X Shares","LU0109391861.USD":"富兰克林美国机遇基金A Acc","NVIW.SI":"NVDA 3xLongSG261006","LU2357305700.SGD":"Allianz Global Artificial Intelligence ET H2-SGD","SNVD.UK":"LS -1X NVIDIA","LU0289961442.SGD":"SUSTAINABLE GLOBAL THEMATIC PORTFOLIO \"AX\" (SGD) ACC","LU1366192091.USD":"ALLIANZ US EQUITY PLUS \"AM\" (USD) INC","LU1582987324.SGD":"M&G (LUX) INCOME ALLOCATION \"A-H\" (SGDHDG) ACC","LU0342679015.USD":"ALLIANZ GLOBAL EQUITY UNCONSTRAINED \"AT\" (USD) ACC","LU2491050154.USD":"WELLINGTON SUSTAINABLE OUTCOMES \"A\" (USD) ACC","LU0823417653.USD":"BNP PARIBAS SUSTAINABLE GLOBAL LOW VOL EQUITY \"C\" (USD) ACC","GB00BDT5M118.USD":"天利环球扩展Alpha基金A Acc","LU2461242641.AUD":"WELLINGTON US QUALITY GROWTH \"A\" (AUDHDG) ACC","LU1582987597.SGD":"M&G (LUX) INCOME ALLOCATION \"A-H\" (SGDHDG) INC","LU0011850046.USD":"贝莱德全球长线股票 A2 USD","3NVD.UK":"LS 3X NVIDIA","IE00B1XK9C88.USD":"PINEBRIDGE US LARGE CAP RESEARCH ENHANCED \"A\" (USD) ACC","LU0466842654.USD":"HSBC ISLAMIC GLOBAL EQUITY INDEX \"A\" (USD) ACC","LU0048584097.USD":"FIDELITY FUNDS  GLOBAL THEMATIC OPPORTUNITIES \"A\" (USD) INC","LU0097036916.USD":"贝莱德美国增长A2 USD","LU1280957306.USD":"THREADNEEDLE (LUX) US CONTRARIAN CORE EQUITIES \"AUP\" (USD) INC","LU2860962120.EUR":"CPR INVEST - ARTIFICIAL INTELLIGENCE \"A2\" (EUR) ACC A","LU2089284900.SGD":"Allianz Global Sustainability Cl AM Dis H2-SGD","NVDA":"英伟达","LU1032955483.USD":"NORDEA 1 STABLE RETURN \"HB\" (USDHDG) ACC","LU2592432038.USD":"WELLINGTON MULTI-ASSET HIGH INCOME \"A\" (USD) ACC","NVDS":"Tradr 1.5X Short NVDA Daily ETF","2NVD.UK":"2X NVIDIA ETP","IE0034235188.USD":"PINEBRIDGE GLOBAL FOCUS EQUITY \"A\" (USD) ACC","LU0456855351.SGD":"JPMorgan Funds - Global Equity A (acc) SGD","NVD2.UK":"2X NVIDIA ETP","NVDU":"Direxion Daily NVDA Bull 2X Shares","NVD3.UK":"LS 3X NVIDIA","LU0310799852.SGD":"FTIF - Templeton Global Equity Income A MDIS SGD","NVDS.UK":"LS -1X NVIDIA","NVDX":"T-REX 2X LONG NVIDIA DAILY TARGET ETF","NVD":"GraniteShares 2x Short NVDA Daily ETF","NVDY":"YIELDMAX NVDA OPTION INCOME STRATEGY ETF","LU0528227936.USD":"富达环球人口趋势基金A-ACC","LU2237443978.SGD":"Aberdeen Standard SICAV I - Global Dynamic Dividend A Acc SGD-H","IE00B3M56506.USD":"NEUBERGER BERMAN EMERGING MARKETS EQUITY \"A\" (USD) ACC","LU2065170008.USD":"M&G (LUX) GLOBAL MAXIMA \"A\" (USD) INC","MSFT":"微软","LU1059921491.USD":"NORDEA 1 GLOBAL STABLE EQUITY \"HB\" (USDHDG) ACC","LU2491049909.HKD":"WELLINGTON SUSTAINABLE OUTCOMES \"A\" (HKD) ACC"},"translate_title":"1.1 million tokens per second! Microsoft (MSFT.US) and Nvidia (NVDA.US) join forces to set a new AI inference record","themeId":null,"isJumpTheme":false,"ttsUrl":null,"symbols_score_info":{"2NVD.UK":0.6,"3NVD.UK":0.6,"MSFT":1.5,"NVD":0.6,"NVD2.UK":0.6,"NVD3.UK":0.6,"NVDA":1.5,"NVDD":0.6,"NVDS":0.6,"NVDS.UK":0.6,"NVDU":0.6,"NVDX":0.6,"NVDY":0.6,"NVIW.SI":0.6,"SNVD.UK":0.6},"content_text":"智通財經APP獲悉，微軟(MSFT.US)宣佈，其Azure ND GB300v6虛擬機在Meta的Llama270B模型上實現了每秒推理速度達110萬token的行業新紀錄。據悉，Azure ND GB300虛擬機採用英偉達(NVDA.US)的Blackwell Ultra GPU，具體為NVIDIA GB300NVL72系統，配置72個NVIDIA Blackwell Ultra GPU和36個NVIDIA Grace CPU，採用單機架構設計。這款虛擬機專為推理工作負載優化，具有50%的GPU內存提升和16%的熱設計功率(TDP)提高。微軟首席執行官薩提亞・納德拉在社交媒體上表示:「這一成就是我們與英偉達長期合作和在生產規模運行人工智能方面專業知識的結晶。」資料顯示，為了驗證性能提升，微軟在一個NVIDIA GB300NVL72域下的18台ND GB300v6虛擬機上運行了Llama270B(FP4精度)模型，使用NVIDIA TensorRT-LLM作為推理引擎。微軟表示:「一個NVL72機架的Azure ND GB300v6實現了總計110萬個token/秒的推理速度。」 這一新記錄超越了微軟之前在NVIDIA GB200NVL72機架上實現的86.5萬個token/秒的成績。對此，Signal65的實驗室副總裁拉斯・費洛斯指出:「這一里程碑不僅突破了每秒百萬token的障礙，還在一個能夠滿足現代企業動態使用和數據治理需求的平台上實現。」 他補充稱，Azure ND GB300相較於上一代NVIDIA GB200在推理性能上提升了27%，而僅增加了17%的功率規格。","kind":"news","is_publish_news":true,"is_publish_highlight":false,"is_publish_live":false,"is_publish_wemedia":null,"editions":null,"column":"","sentiment":"2","news_tag":"corporation","news_rank":0,"symbols":[],"gpt_button":0,"need_auth":false,"code":"91000000","status":"200"}}}