[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-openais-first-custom-chip-targets-inference-costs":10,"sections":34},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},2150,"openais-first-custom-chip-targets-inference-costs","OpenAI's First Custom Chip Targets Inference Costs","OpenAI and Broadcom have built a custom processor called Jalapeño, designed to handle the specific demands of OpenAI's inference workloads.","OpenAI has a custom chip now, and it runs on Broadcom silicon.\n\nThe processor, named Jalapeño, was designed specifically for inference — the part of AI where the model actually responds to a user query, as opposed to the training runs that get most of the attention. OpenAI partnered with Broadcom to build it, marking the company's first move into custom silicon. The chip is purpose-built for OpenAI's own systems rather than sold to outside customers.\n\nThis matters because inference is where the money goes. Every ChatGPT query, every API call, every agent loop burns compute — and that compute currently runs on hardware built for general workloads. A chip tuned to OpenAI's exact inference patterns could cut costs significantly, giving the company more control over its unit economics at a moment when it needs them.\n\nGoogle has been running custom TPUs for years, and Amazon and Microsoft have both developed in-house AI silicon. OpenAI — despite being the company that made AI inference a mass-market problem — is arriving to that party late. Jalapeño is a start, but the gap between a first chip and a mature silicon program is wide.","[\"openai\",\"hardware\",\"ai\",\"inference\"]","2026-06-24T14:54:46.000Z","2026-06-24T15:52:33.543Z","2026-06-24T15:52:36.381Z","published",null,[],"ai",[26,27,24,28],"openai","hardware","inference",[30],{"name":31,"url":32},"TechCrunch","https:\u002F\u002Ftechcrunch.com\u002F2026\u002F06\u002F24\u002Fopenai-unveils-its-first-custom-chip-built-by-broadcom\u002F",0,{"sections":35},[36,40,45,50,55,60,64,69,74,79,84,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",539,"2026-06-24T18:46:18.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Deals","deals",161,"2026-06-24T13:00:00.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Security","security",148,"2026-06-24T17:25:00.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",106,"2026-06-24T17:36:44.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Consumer Tech","consumer-tech",87,"2026-06-24T14:22:56.000Z",{"name":61,"slug":27,"count":62,"latest_published_at":63},"Hardware",81,"2026-06-24T18:50:56.000Z",{"name":65,"slug":66,"count":67,"latest_published_at":68},"Software","software",64,"2026-06-24T16:15:00.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Dev Tools","dev-tools",53,"2026-06-23T18:13:40.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":78},"Science","science",39,"2026-06-23T05:25:16.000Z",{"name":80,"slug":81,"count":82,"latest_published_at":83},"Gaming","gaming",36,"2026-06-24T15:05:01.000Z",{"name":85,"slug":86,"count":87,"latest_published_at":88},"Startups","startups",33,"2026-06-24T18:27:41.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"General","general",27,"2026-06-24T08:50:14.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]