{"id":9562,"date":"2025-03-12T04:01:13","date_gmt":"2025-03-12T04:01:13","guid":{"rendered":"https:\/\/fusionmindlabs.com\/blogs\/meta-tests-in-house-ai-chip\/"},"modified":"2025-03-12T04:01:13","modified_gmt":"2025-03-12T04:01:13","slug":"meta-tests-in-house-ai-chip","status":"publish","type":"post","link":"https:\/\/fusionmindlabs.com\/blogs\/meta-tests-in-house-ai-chip\/","title":{"rendered":"Meta tests in-house AI chip"},"content":{"rendered":"\n<\/p>\n<div>\n<p>Breaking News: Meta, a major Big Tech company, is accelerating its AI development efforts. (Quite a surprise!) In this scenario, the key player is Meta, which, according to <em>Reuters<\/em>, is testing its first in-house AI training chip. The goal is to reduce its massive infrastructure costs and dependence on NVIDIA, a company that seems to <a data-i13n=\"cpos:1;pos:1\" href=\"https:\/\/www.engadget.com\/mark-zuckerberg-says-fck-that-to-closed-platforms-235700788.html\" data-ylk=\"slk:brings out Mark Zuckerberg&#039;s &quot;adult language&quot;;cpos:1;pos:1;elm:context_link;itc:0;sec:content-canvas\" class=\"link \" rel=\"nofollow noopener\" target=\"_blank\">elicit strong reactions from Mark Zuckerberg<\/a>. If successful, Meta plans to utilize this chip for training purposes by 2026.<\/p>\n<p>Meta has reportedly initiated a small-scale deployment of the dedicated accelerator chip, designed specifically for AI tasks, making it more power-efficient than <a data-i13n=\"cpos:2;pos:1\" href=\"https:\/\/www.engadget.com\/gaming\/pc\/nvidia-geforce-rtx-5070-review-basically-a-4070-ti-thats-better-at-4k-140016718.html\" data-ylk=\"slk:general-purpose NVIDIA GPUs;cpos:2;pos:1;elm:context_link;itc:0;sec:content-canvas\" class=\"link \" rel=\"nofollow noopener\" target=\"_blank\">general-purpose NVIDIA GPUs<\/a>. This deployment follows the completion of its first &#8220;tape-out,&#8221; a critical phase in silicon development where a complete design is sent for a manufacturing test run.<\/p>\n<p>The chip is part of the Meta Training and Inference Accelerator (MTIA) <a data-i13n=\"elm:context_link;elmt:doNotAffiliate;cpos:3;pos:1\" class=\"link \" href=\"https:\/\/ai.meta.com\/blog\/next-generation-meta-training-inference-accelerator-AI-MTIA\/\" rel=\"nofollow noopener\" target=\"_blank\" data-ylk=\"slk:series;elm:context_link;elmt:doNotAffiliate;cpos:3;pos:1;itc:0;sec:content-canvas\">series<\/a>, the company&#8217;s family of custom in-house silicon focused on generative AI, recommendation systems, and advanced research.<\/p>\n<p>Last year, the company began using an MTIA chip for <a data-i13n=\"cpos:4;pos:1\" href=\"https:\/\/www.engadget.com\/2019-01-07-intel-nervana-processor-for-inference.html\" data-ylk=\"slk:inference;cpos:4;pos:1;elm:context_link;itc:0;sec:content-canvas\" class=\"link \" rel=\"nofollow noopener\" target=\"_blank\">inference<\/a>, a predictive process that occurs behind the scenes in AI models. Meta utilized the inference chip for its Facebook and Instagram news feed recommendation systems. <em>Reuters<\/em> reports that it plans to start using the training silicon for these purposes as well. The long-term plan for both chips is to begin with recommendations and eventually use them for generative products like the <a data-i13n=\"cpos:5;pos:1\" href=\"https:\/\/www.engadget.com\/social-media\/meta-ai-will-now-use-your-facebook-and-instagram-activity-to-inform-its-recommendations-201218403.html\" data-ylk=\"slk:Meta AI chatbot;cpos:5;pos:1;elm:context_link;itc:0;sec:content-canvas\" class=\"link \" rel=\"nofollow noopener\" target=\"_blank\">Meta AI chatbot<\/a>.<\/p>\n<p>As one of NVIDIA&#8217;s largest customers, after <a data-i13n=\"elm:context_link;elmt:doNotAffiliate;cpos:6;pos:1\" class=\"link \" href=\"https:\/\/www.reuters.com\/technology\/inside-metas-scramble-catch-up-ai-2023-04-25\/\" rel=\"nofollow noopener\" target=\"_blank\" data-ylk=\"slk:placing;elm:context_link;elmt:doNotAffiliate;cpos:6;pos:1;itc:0;sec:content-canvas\">placing orders<\/a> for billions of dollars&#8217; worth of GPUs in 2022, Meta is now shifting its focus to in-house development. This move comes after the company previously abandoned an in-house inference silicon that failed a small-scale test deployment, much like the current test for the training chip.<\/p>\n<\/div>\n\n<br \/><a href=\"https:\/\/www.engadget.com\/ai\/meta-is-reportedly-testing-its-first-in-house-ai-training-chip-174550664.html?src=rss\" rel=\"nofollow noopener\" target=\"_blank\">Source Link<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Breaking News: Meta, a major Big Tech company, is accelerating its AI development efforts. (Quite a surprise!) In this scenario, the key player is Meta, which, according to Reuters, is&#8230;<\/p>\n","protected":false},"author":0,"featured_media":9563,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[47],"tags":[65,64,151,2132,63],"class_list":["post-9562","post","type-post","status-publish","format-standard","has-post-thumbnail","category-news","tag-ai","tag-gear","tag-meta","tag-meta-ai","tag-news"],"jetpack_featured_media_url":"https:\/\/fusionmindlabs.com\/blogs\/wp-content\/uploads\/sites\/9\/2025\/03\/ebfb0510-fe9a-11ef-afd2-678bdb728f8b.jpeg","_links":{"self":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/posts\/9562","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/comments?post=9562"}],"version-history":[{"count":0,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/posts\/9562\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/media\/9563"}],"wp:attachment":[{"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/media?parent=9562"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/categories?post=9562"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fusionmindlabs.com\/blogs\/wp-json\/wp\/v2\/tags?post=9562"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}