{"id":5900,"date":"2024-03-21T09:28:32","date_gmt":"2024-03-21T01:28:32","guid":{"rendered":"https:\/\/www.1ai.net\/?p=5900"},"modified":"2024-03-21T09:28:32","modified_gmt":"2024-03-21T01:28:32","slug":"meta-%e5%8f%91%e8%a8%80%e4%ba%ba%e7%a1%ae%e8%ae%a4%e5%b0%86%e7%8e%87%e5%85%88%e4%bd%bf%e7%94%a8%e8%8b%b1%e4%bc%9f%e8%be%be%e6%97%97%e8%88%b0-ai-%e8%8a%af%e7%89%87-b200%ef%bc%8c%e9%a2%84%e4%bc%b0","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/5900.html","title":{"rendered":"Meta spokesperson confirmed that it will be the first to use Nvidia's flagship AI chip B200, and it is estimated that it will receive"},"content":{"rendered":"

Meta<\/a> The company expects to receive a $2.5 million payment later this year, a spokesperson told Reuters on Thursday.Nvidia<\/a>Latest flagship AI Chips<\/a> \u2014\u2014B200<\/a>, and it is NVIDIA\u2019s first batch of shipments.<\/p>\n

Nvidia\u2019s chief financial officer, Colette Kress, told financial analysts that \u201cwe think (the product) will be available later this year,\u201d but she also said thatShipments of new GPUs will not increase until 2025<\/span>.<\/p>\n

It is reported that social media giant Meta is one of Nvidia's largest customers and has purchased hundreds of thousands of previous generation accelerator cards (H100). Meta's CEO Zuckerberg announced in January that the company plans to reserve about 350,000 H100s by the end of this year. He added that combined with other GPUs, Meta will have the equivalent of about 600,000 H100s by then.<\/p>\n

At the GTC Developer Conference in the early morning of the 19th Beijing time, NVIDIA released the most powerful AI accelerator card GB200, which uses the new generation of AI graphics processor architecture Blackwell. Huang Renxun said at the GTC conference: "Hopper is already very good, but we need more powerful GPUs<\/span>\u201d.<\/p>\n

The Blackwell GPU is huge and is etched using TSMC's 4-nanometer (4NP) process. It integrates two independently manufactured dies with a total of 208 billion transistors, and then NVLink 5.0 is used to bundle the chips like a zipper.<\/p>\n

GB200 includes two B200 Blackwell GPUs and an Arm-based Grace CPU<\/span>The performance of reasoning large language models is 30 times higher than that of H100, and the cost and energy consumption are reduced to one 25th.<\/p>\n

\"36600c07-98d8-4ade-b1cc-853af4f7a433\"<\/p>\n

Huang Renxun said that Blackwell's AI performance can reach 20 petaflops, while H100 is only 4 petaflops. According to Nvidia officials, the additional processing power will enable artificial intelligence companies to train larger and more complex models.<\/p>","protected":false},"excerpt":{"rendered":"

A Meta spokesperson told Reuters on 19 EST that the company expects to receive the B200, NVIDIA's latest flagship AI chip, later this year and for NVIDIA's first shipments. NVIDIA's chief financial officer, Colette Kress, told financial analysts that \"we think [the product] will be available later this year,\" but also said that shipments of the new GPUs won't increase until 2025. Social media giant Meta is reportedly one of NVIDIA's biggest customers, having purchased hundreds of thousands of last-generation accelerator cards (H100.) Meta's CEO Zuckerberg had announced in January that the company planned to stockpile around 350,000 H100s by the end of this year. he added that the kno<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[1099,1783,297,239],"collection":[],"class_list":["post-5900","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-b200","tag-meta","tag-239"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/5900","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=5900"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/5900\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=5900"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=5900"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=5900"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=5900"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}