[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-intel-d270-ai-chip-swaps-hbm-for-480gb-of-lpddr5x":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},11431,"intel-d270-ai-chip-swaps-hbm-for-480gb-of-lpddr5x","Intel D270 AI Chip Swaps HBM for 480GB of LPDDR5X","Intel's new D270 data center GPU trades the industry's usual HBM memory for up to 480GB of cheaper LPDDR5X, betting on capacity over bandwidth.","Intel's newest AI accelerator skips the industry's favorite memory and loads up on capacity instead.\n\nIntel quietly listed the D270, a discrete data center GPU built for AI workloads, on its product site. The headline spec is memory, not compute: up to 480 GB of LPDDR5X per card, the kind of memory found in laptops and phones rather than the HBM stacks Nvidia and AMD use in their top AI chips. The listing surfaced on Hacker News, where it drew a modest round of discussion. Intel has not detailed pricing, availability, or performance benchmarks.\n\nThat tradeoff is the real story. HBM is fast but expensive and supply-constrained, which is part of why Nvidia's top chips stay scarce and pricey. LPDDR5X is slower but far cheaper and easier to source in bulk, letting Intel pack a large memory pool onto a card for a fraction of the cost. For inference workloads that need to hold big models in memory more than they need raw bandwidth, that could be a genuinely useful niche.\n\nIntel has talked a big game on AI chips before, from Gaudi to earlier discrete GPUs, without landing a hit against Nvidia. A memory-capacity bet is different enough to watch, but different isn't the same as winning.","[\"ai-hardware\",\"intel\",\"data-center-gpu\",\"lpddr5x\"]","2026-10-12T21:35:35.000Z","2026-10-13T00:48:00.074Z","2026-10-13T00:48:05.687Z","published",null,[],"hardware",[26,27,28,29],"ai-hardware","intel","data-center-gpu","lpddr5x",[31],{"name":32,"url":33},"Hacker News","https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fproducts\u002Fdetails\u002Fdiscrete-gpus\u002Fdata-center-gpu\u002Fd-series.html",0,{"sections":36},[37,42,47,52,57,60,65,70,75,80,85,90,95,100],{"name":38,"slug":39,"count":40,"latest_published_at":41},"AI","ai",6974,"2026-10-12T23:20:00.000Z",{"name":43,"slug":44,"count":45,"latest_published_at":46},"Security","security",962,"2026-10-13T00:20:00.000Z",{"name":48,"slug":49,"count":50,"latest_published_at":51},"Policy","policy",500,"2026-10-12T18:34:57.000Z",{"name":53,"slug":54,"count":55,"latest_published_at":56},"Deals","deals",492,"2026-10-12T14:55:53.000Z",{"name":58,"slug":24,"count":59,"latest_published_at":18},"Hardware",243,{"name":61,"slug":62,"count":63,"latest_published_at":64},"Science","science",197,"2026-10-12T23:34:22.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Consumer Tech","consumer-tech",192,"2026-10-12T14:07:13.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Startups","startups",122,"2026-10-12T17:19:14.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Software","software",116,"2026-10-12T19:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Dev Tools","dev-tools",113,"2026-10-12T22:34:47.000Z",{"name":86,"slug":87,"count":88,"latest_published_at":89},"General","general",72,"2026-10-12T21:54:35.000Z",{"name":91,"slug":92,"count":93,"latest_published_at":94},"Gaming","gaming",63,"2026-10-12T20:59:50.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"Reviews","reviews",35,"2026-10-12T14:00:00.000Z",{"name":101,"slug":102,"count":103,"latest_published_at":104},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]