{"rewrite":{"id":"r_0398031318c12481c639cd7b","clusterId":"c_921d0d5e37cf6d53489bec93","slug":"openai-says-its-models-cut-amd-gpu-design-work-from-a-year-to-five-months","model":"deepseek-v4-1-flash","headline":"OpenAI Says Its Models Cut AMD GPU Design Work From a Year to Five Months","summary":"OpenAI used its o3-mini model to help design Jalapeño, the inference chip it is co-developing with Broadcom, shrinking vector unit cell area by about 13% and matrix unit cell area by 10%. Work with AMD on GPU circuit code achieved up to 14% PPA improvement against an optimized baseline, and a gain OpenAI says would have taken about a year arrived in five months.","whyItMatters":"OpenAI is selling its models as chip design tools, and the named customers are the companies whose silicon it also buys and builds alongside.","webCardHtml":"\u003cp\u003eAMD is folding the changes into product designs, with PPA up by a single-digit percentage across the main blocks of its next-generation flagship CPU. AMD CEO Lisa Su said engineers at both companies are pushing the frontier of chip design.\u003c/p\u003e\u003cp\u003eOpenAI also tested its models against PicoRV32, a RISC-V core built for FPGAs. GPT-6.2 Astra gave the largest performance gain at the lowest cost among the five models tried. With Synopsys, OpenAI developed GPT-Synopsys, a model for operating the company\u0026#39;s EDA tools.\u003c/p\u003e","blueskyPost":"OpenAI's o3-mini shrank Jalapeño's vector unit cell area by about 13%, a design task the company frames as AI doing engineering work rather than assisting it. The chip is co-developed with Broadcom.","twitterPost":"Jalapeño's 13% vector unit cell area cut came from OpenAI's o3-mini, not a human layout team. The chip is co-developed with Broadcom.","threadsPost":"Jalapeño is the detail worth sitting with: OpenAI says o3-mini shrank its vector unit cell area by about 13% and its matrix unit cell area by 10%. That is a model making layout calls on a chip OpenAI is co-developing with Broadcom, not drafting text about one.","newsletterBlurb":"OpenAI reports that o3-mini helped design Jalapeño, its Broadcom co-developed inference chip, cutting vector unit cell area about 13% and matrix unit cell area 10%. Work with AMD on GPU circuit code reached up to 14% PPA improvement, and AMD is putting the changes into product designs. OpenAI and Synopsys also built GPT-Synopsys for operating EDA tools.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20261119-openai-jalapeno-amd-advancing-chip/\",\"title\":\"OpenAI Successfully Uses AI to Improve AMD GPUs and Its Own AI Chips, Also Develops Dedicated AI Model for Semiconductor Design\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":5606,"outputTokens":703,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1795080755,"createdAt":"2026-11-19T09:29:09.000Z","publishedAt":"2026-11-19T09:29:14.000Z","updatedAt":"2026-11-19T09:29:14.000Z"},"cluster":{"id":"c_921d0d5e37cf6d53489bec93","canonicalTitle":"OpenAIがAIを用いてAMD製GPUや独自開発AIチップの改善に成功、半導体設計の専用AIモデルも開発","representativeArticleId":"a_0a84d1fb1f16564c9c6856f0","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Jalapeño\"],\"studios\":[\"OpenAI\",\"AMD\"],\"people\":[],\"type\":\"news\",\"domain\":\"industry\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-11-19T07:58:00.000Z","lastSeenAt":"2026-11-19T07:58:00.000Z","updatedAt":"2026-11-19T09:29:12.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20261119-openai-jalapeno-amd-advancing-chip/","title":"OpenAIがAIを用いてAMD製GPUや独自開発AIチップの改善に成功、半導体設計の専用AIモデルも開発"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Jalapeño"],"studios":["OpenAI","AMD"],"people":[],"type":"news","domain":"industry","is_roundup":false},"keyFacts":["OpenAI co-develops the inference chip Jalapeño with Broadcom.","o3-mini reduced Jalapeño's vector unit cell area by about 13% and matrix unit cell area by 10%.","OpenAI's work with AMD on GPU circuit code achieved up to 14% PPA improvement over a highly optimized baseline.","AMD CEO Lisa Su commented on the chip design work.","OpenAI and Synopsys developed GPT-Synopsys, a model for operating Synopsys EDA tools."],"pageLinks":{"healthSlugs":["jalapeno","openai","amd"],"storySlugs":[],"entities":[{"slug":"jalapeno","supported":2,"disputed":0,"contradicted":0,"unverified":3,"overall":"green","story_count":2,"has_dossier":false},{"slug":"openai","supported":8,"disputed":0,"contradicted":1,"unverified":28,"overall":"red","story_count":37,"has_dossier":true},{"slug":"amd","supported":2,"disputed":0,"contradicted":0,"unverified":8,"overall":"green","story_count":12,"has_dossier":true}],"stories":[],"entitiesOk":true,"storiesOk":true}}
