[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-tesla-renames-its-ai-chips-si-and-promises-80x-for-si6":10,"sections":35},{"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":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},12564,"tesla-renames-its-ai-chips-si-and-promises-80x-for-si6","Tesla Renames Its AI Chips SI and Promises 80x for SI6","Tesla's chip roadmap swaps AI for SI, claims an 80x jump for an inference chip due in 2029-2030, and moves Dojo 3 to a chiplet design with stacked HBM.","Tesla's updated chip roadmap promises 80x the inference performance of today's in-car computer by 2029 or 2030, and it relabels the whole lineup. The AI4, AI5 and AI6 chips are now SI4, SI5 and SI6, short for super intelligence.\n\nThe baseline is SI4, the chip (formerly HW4) in Tesla cars today. According to the roadmap, SI5 delivers 40x that performance. Tesla attributes it to an eightfold gain in compute and a fivefold gain from silicon optimization, a split Elon Musk has described. SI5 uses 3nm-class silicon at TSMC and 2nm-class at Samsung, a multi-chip module, 96 GB of memory, and a power draw of 250W to 800W depending on the application. SI6 then reaches 80x, with 144 GB of memory.\n\nThe training side changed more. Musk cancelled the wafer-scale Dojo project in August 2025, and in August 2026 said SpaceX and xAI would use only Nvidia hardware. Dojo 3 is back anyway, now described as a chiplet design using 2.5D and 3D packaging with stacked HBM (high-bandwidth memory). That replaces the original approach of 25 D1 processors sharing 11 GB of on-chip SRAM per wafer. Tesla said D3 is optimized for space, training and scaling, but gave no performance, process or schedule details.\n\nThe 80x figure is softer than it sounds. Tesla does not publish absolute numbers for SI4. Working from Musk's statement that HW4 is 3x to 8x faster than the 72 TOPS HW3, estimates for SI6 span 17,280 to 46,080 TOPS, a range wide enough to hold two different chips. Tesla has not said which data formats SI6 supports, so comparing it to Nvidia's Rubin, rated at 50 NVFP4 PFLOPS for inference and shipping now, tells little.\n\nThe rename is the cheapest change on the slide. Whether D3 keeps any wafer-scale integration is the expensive one: stacking HBM on wafer-scale logic would add cooling, yield and cost problems, which is why a conventional design looks likelier.","[\"tesla\",\"ai-chips\",\"dojo\",\"inference\"]","2026-11-06T12:00:00.000Z","2026-11-08T10:01:58.166Z","2026-11-08T10:01:58.342Z","published",null,[],"hardware",[26,27,28,29],"tesla","ai-chips","dojo","inference",[31],{"name":32,"url":33},"Tom's Hardware","https:\u002F\u002Fwww.tomshardware.com\u002Ftech-industry\u002Fartificial-intelligence\u002Ftesla-unveils-ai-accelerator-roadmap-adopts-new-si-branding-in-super-intelligence-push-si6-to-bring-80x-more-compute-performance-dojo-3-headed-to-space",0,{"sections":36},[37,42,47,52,57,62,66,71,76,81,86,91,96,101],{"name":38,"slug":39,"count":40,"latest_published_at":41},"AI","ai",7110,"2026-11-08T14:34:09.000Z",{"name":43,"slug":44,"count":45,"latest_published_at":46},"Security","security",1099,"2026-11-08T15:24:01.000Z",{"name":48,"slug":49,"count":50,"latest_published_at":51},"Deals","deals",620,"2026-11-08T08:00:00.000Z",{"name":53,"slug":54,"count":55,"latest_published_at":56},"Policy","policy",604,"2026-11-07T15:48:26.000Z",{"name":58,"slug":59,"count":60,"latest_published_at":61},"Consumer Tech","consumer-tech",336,"2026-11-08T21:00:00.000Z",{"name":63,"slug":24,"count":64,"latest_published_at":65},"Hardware",316,"2026-11-08T14:36:39.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Science","science",222,"2026-11-07T16:19:17.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Software","software",154,"2026-11-08T15:42:52.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",150,"2026-11-06T16:50:01.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Dev Tools","dev-tools",125,"2026-11-06T22:15:59.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",111,"2026-11-08T21:27:29.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Gaming","gaming",94,"2026-11-08T20:52:05.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"Reviews","reviews",38,"2026-10-20T12:00:00.000Z",{"name":102,"slug":103,"count":104,"latest_published_at":105},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]