{"id":35561,"date":"2025-12-04T07:44:57","date_gmt":"2025-12-04T07:44:57","guid":{"rendered":"https:\/\/metaverseplanet.net\/blog\/?p=35561"},"modified":"2026-01-05T08:00:40","modified_gmt":"2026-01-05T08:00:40","slug":"amazon-introduces-highly-ambitious-next-generation-ai-chip-trainium3","status":"publish","type":"post","link":"https:\/\/metaverseplanet.net\/blog\/amazon-introduces-highly-ambitious-next-generation-ai-chip-trainium3\/","title":{"rendered":"Amazon Introduces Highly Ambitious Next-Generation AI Chip Trainium3"},"content":{"rendered":"\n<p><strong>AWS<\/strong> has introduced the new <strong>Trainium3 chip<\/strong>. The new chip offers a <strong>4x increase<\/strong> in performance and memory. <strong>Trainium4<\/strong> will be compatible with <strong><em><a href=\"https:\/\/metaverseplanet.net\/blog\/tag\/nvidia-news-and-content\/\" data-type=\"post_tag\" data-id=\"102\">Nvidia<\/a><\/em> GPUs<\/strong> and will be used in large <strong>AI clusters<\/strong>.<\/p>\n\n\n\n<p><strong>Amazon Web Services (AWS)<\/strong> continues its long-developed series of <strong>AI training <em><a href=\"https:\/\/metaverseplanet.net\/blog\/tag\/chip-technology\/\" data-type=\"post_tag\" data-id=\"156\">chips<\/a><\/em><\/strong> with <strong>Trainium3<\/strong>. At the <strong>re:Invent 2025<\/strong> event, <strong>AWS<\/strong> introduced the new training processor <strong>Trainium3<\/strong> and the large-scale <strong>Trainium3 UltraServer<\/strong> system. The company also confirmed that its next-generation processor, <strong>Trainium4<\/strong>, is in the development phase.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity is-style-wide\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>4x Improvement in Training and Inference<\/strong><\/h2>\n\n\n\n<figure class=\"wp-block-image size-full\"><img decoding=\"async\" width=\"720\" height=\"405\" src=\"https:\/\/metaverseplanet.net\/blog\/wp-content\/uploads\/2025\/12\/Amazon-Introduces-Highly-Ambitious-Next-Generation-AI-Chip-Trainium3-2.webp\" alt=\"\" class=\"wp-image-35563\" srcset=\"https:\/\/metaverseplanet.net\/blog\/wp-content\/uploads\/2025\/12\/Amazon-Introduces-Highly-Ambitious-Next-Generation-AI-Chip-Trainium3-2.webp 720w, https:\/\/metaverseplanet.net\/blog\/wp-content\/uploads\/2025\/12\/Amazon-Introduces-Highly-Ambitious-Next-Generation-AI-Chip-Trainium3-2-300x169.webp 300w, https:\/\/metaverseplanet.net\/blog\/wp-content\/uploads\/2025\/12\/Amazon-Introduces-Highly-Ambitious-Next-Generation-AI-Chip-Trainium3-2-390x220.webp 390w, https:\/\/metaverseplanet.net\/blog\/wp-content\/uploads\/2025\/12\/Amazon-Introduces-Highly-Ambitious-Next-Generation-AI-Chip-Trainium3-2-150x84.webp 150w\" sizes=\"(max-width: 720px) 100vw, 720px\" \/><\/figure>\n\n\n\n<p><strong>Trainium3<\/strong> offers a fourfold increase in both <strong>training<\/strong> and <strong>inference performance<\/strong>, as well as a fourfold increase in <strong>memory capacity<\/strong>, compared to its predecessor, <strong>Trainium2<\/strong>. Each <strong>UltraServer<\/strong> hosts 144 <strong>Trainium3 chips<\/strong>, and <strong>AWS<\/strong> allows customers to connect thousands of these systems together. In a single deployment at maximum scale, 1 million <strong>Trainium3 chips<\/strong> can be used. This represents a tenfold increase compared to the previous generation.<\/p>\n\n\n\n<p><strong>Energy efficiency<\/strong> is another important feature of <strong>Trainium3<\/strong>. According to <strong>AWS<\/strong>, these systems consume approximately <strong>40 percent less energy<\/strong> while providing higher transaction volume. Early adopters of <strong>Trainium3<\/strong> include <strong>Anthropic<\/strong>, the Japanese <strong>LLM<\/strong> startup <strong>Karakuri<\/strong>, <strong>SplashMusic<\/strong>, and <strong>Decart<\/strong>. These customers have reported faster inference, faster iteration, and lower billable compute hours for the models they develop.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity is-style-wide\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Trainium4 and Hybrid Ecosystem Strategy<\/strong><\/h2>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<iframe title=\"AWS Trainium3-Powered Amazon EC2 Trn3 UltraServers | Amazon Web Services\" width=\"500\" height=\"281\" src=\"https:\/\/www.youtube.com\/embed\/4y3pMGIS6DU?feature=oembed\" frameborder=\"0\" allow=\"accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share\" referrerpolicy=\"strict-origin-when-cross-origin\" allowfullscreen><\/iframe>\n<\/div><\/figure>\n\n\n\n<p><strong>AWS<\/strong> also gave hints about the next-generation <strong>Trainium4<\/strong>. Although an official release date was not announced, <strong>Trainium4<\/strong> promises a significant leap in performance, and its most notable feature is <strong>Nvidia NVLink Fusion<\/strong> support. This approach shows that <strong>AWS<\/strong> positions its hardware not as an alternative to <strong>Nvidia<\/strong> solutions, but as a complementary element in a hybrid ecosystem. Thus, companies using <strong>CUDA<\/strong>-based applications will be able to integrate <strong>AWS<\/strong> proprietary hardware without rebuilding their software stacks.<\/p>\n\n\n\n<p>The <strong>AWS<\/strong> and <strong>Nvidia partnership<\/strong> covers connectivity technology, <strong>data center infrastructure<\/strong>, open model support, and rack-scale <strong>AI system<\/strong> deployment. <strong>NVLink Fusion<\/strong> support will be integrated into future platforms, including <strong>Trainium4<\/strong>, <strong>Graviton CPUs<\/strong>, and <strong>Nitro virtualization hardware<\/strong>. On the other hand, the company did not share pricing and availability information for the new chips.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">You Might Also Like;<\/h3>\n\n\n<ul class=\"wp-block-latest-posts__list wp-block-latest-posts\"><li><a class=\"wp-block-latest-posts__post-title\" href=\"https:\/\/metaverseplanet.net\/blog\/the-dark-side-of-nanotechnology\/\">The Dark Side of Nanotechnology: Could Microscopic Swarms Erase Billions?<\/a><\/li>\n<li><a class=\"wp-block-latest-posts__post-title\" href=\"https:\/\/metaverseplanet.net\/blog\/the-illusion-of-digital-immortality\/\">The Illusion of Digital Immortality: Are You Really Uploading Your Mind?<\/a><\/li>\n<li><a class=\"wp-block-latest-posts__post-title\" href=\"https:\/\/metaverseplanet.net\/blog\/artemis-2s-deep-space-eclipse\/\">The View That Changes Everything: Artemis 2\u2019s Deep Space Eclipse<\/a><\/li>\n<\/ul>","protected":false},"excerpt":{"rendered":"<p>AWS has introduced the new Trainium3 chip. The new chip offers a 4x increase in performance and memory. Trainium4 will be compatible with Nvidia GPUs and will be used in large AI clusters. Amazon Web Services (AWS) continues its long-developed series of AI training chips with Trainium3. At the re:Invent 2025 event, AWS introduced the &hellip;<\/p>\n","protected":false},"author":1,"featured_media":35564,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"googlesitekit_rrm_CAown96uCw:productID":"","footnotes":""},"categories":[332],"tags":[335,156],"class_list":["post-35561","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-information","tag-ai-news","tag-chip-technology"],"amp_enabled":true,"_links":{"self":[{"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/posts\/35561","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/comments?post=35561"}],"version-history":[{"count":0,"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/posts\/35561\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/media\/35564"}],"wp:attachment":[{"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/media?parent=35561"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/categories?post=35561"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/metaverseplanet.net\/blog\/wp-json\/wp\/v2\/tags?post=35561"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}