{"id":191858,"date":"2024-06-26T06:22:55","date_gmt":"2024-06-26T11:22:55","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2024\/06\/ai-startup-etched-unveils-transformer-asic-claiming-20x-speed-up-over-nvidia-h100"},"modified":"2024-06-26T06:22:55","modified_gmt":"2024-06-26T11:22:55","slug":"ai-startup-etched-unveils-transformer-asic-claiming-20x-speed-up-over-nvidia-h100","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2024\/06\/ai-startup-etched-unveils-transformer-asic-claiming-20x-speed-up-over-nvidia-h100","title":{"rendered":"AI Startup Etched Unveils Transformer ASIC Claiming 20x Speed-up Over NVIDIA H100"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/ai-startup-etched-unveils-transformer-asic-claiming-20x-speed-up-over-nvidia-h1003.jpg\"><\/a><\/p>\n<p>A new startup emerged out of stealth mode today to power the next generation of generative AI. Etched is a company that makes an application-specific integrated circuit (ASIC) to process \u201cTransformers.\u201d The transformer is an architecture for designing deep learning models developed by Google and is now the powerhouse behind models like OpenAI\u2019s GPT-4o in ChatGPT, Antrophic Claude, Google Gemini, and Meta\u2019s Llama family. Etched wanted to create an ASIC for processing only the transformer models, making a chip called Sohu. The claim is Sohu outperforms NVIDIA\u2019s latest and greatest by an entire order of magnitude. Where a server configuration with eight NVIDIA H100 GPU clusters pushes Llama-3 70B models at 25,000 tokens per second, and the latest eight B200 \u201cBlackwell\u201d GPU cluster pushes 43,000 tokens\/s, the eight Sohu clusters manage to output 500,000 tokens per second.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>A new startup emerged out of stealth mode today to power the next generation of generative AI. Etched is a company that makes an application-specific integrated circuit (ASIC) to process \u201cTransformers.\u201d The transformer is an architecture for designing deep learning models developed by Google and is now the powerhouse behind models like OpenAI\u2019s GPT-4o in [\u2026]<\/p>\n","protected":false},"author":718,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[],"class_list":["post-191858","post","type-post","status-publish","format-standard","hentry","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/191858","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/718"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=191858"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/191858\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=191858"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=191858"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=191858"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}