{"id":42021,"date":"2025-12-04T14:35:43","date_gmt":"2025-12-04T19:35:43","guid":{"rendered":"https:\/\/www.eng.ufl.edu\/news\/?p=42021"},"modified":"2025-12-04T16:06:51","modified_gmt":"2025-12-04T21:06:51","slug":"americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus","status":"publish","type":"post","link":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/","title":{"rendered":"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019"},"content":{"rendered":"\n<p>AI has entered a phase where the limits are no longer set by algorithms or transistor counts but by energy and speed.&nbsp;&nbsp;<\/p>\n\n\n\n<p>For years, the industry relied on a simple strategy: Deploy more Graphics Processing Units (GPUs). That approach is now colliding with physics. Datacenters push megawatts of power, airborne and orbital systems cannot dissipate power, and edge computing devices live under strict energy consumption. Scaling electronic computing alone is no longer feasible.&nbsp;&nbsp;<\/p>\n\n\n\n<p>Researchers at the University of Florida are seeking a new direction. In work published in <a href=\"https:\/\/opg.optica.org\/prj\/fulltext.cfm?uri=prj-13-12-B120\" target=\"_blank\" rel=\"noreferrer noopener\"><em>Photonics Research<\/em><\/a>, UF engineers demonstrate a hybrid photonic\u2013electronic neural network that shifts one of the most energy-intensive stages of AI \u2014 feature extraction \u2014 into light itself. The result is dramatic: 6.5 times faster training and nearly 1,000 times higher energy efficiency than an all-electronic baseline.&nbsp;<\/p>\n\n\n\n<p>\u201cAI acceleration is becoming an energy problem as much as a computing problem,\u201d said Elham Heidari, Ph.D., assistant professor in the UF Department of Electrical and Computer Engineering (ECE) and an American Chemical Society Photonics Rising Star. \u201cOur results show a practical path where light performs the front-end neural processing, so electronics don\u2019t have to carry the entire load.\u201d&nbsp;&nbsp;<\/p>\n\n\n\n<div class=\"wp-block-create-block-blockquote\"><section class=\"blockquote-section w-100 position-relative\"><div class=\"container-fluid\"><div class=\"row d-flex justify-content-center align-items-center\"><div class=\"col col-lg-12\"><div class=\"card\"><blockquote class=\"blockquote\"><span><span class=\"quote\">\n<p><span class=\"TextRun SCXW176974774 BCX2\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"auto\"><span class=\"NormalTextRun SCXW176974774 BCX2\">\u201cStart with photons, end with electrons \u2014 that\u2019s how AI scales with physics, not power.\u201d<\/span><\/span><\/p>\n<\/span><span class=\"quotee\">&#8211; Elham Heidari, Ph.D., assistant professor in the UF Department of Electrical and Computer Engineering<\/span><\/span><\/blockquote><\/div><\/div><\/div><\/div><\/section><\/div>\n\n\n\n<h2 class=\"wp-block-heading\">A neural network that thinks in light&nbsp;<\/h2>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"alignleft size-medium\"><img loading=\"lazy\" decoding=\"async\" width=\"500\" height=\"369\" src=\"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/dalir-ye-500x369.jpg\" alt=\"Hamed Dalir, associate professor of Electrical and Computer Engineering and director of UF\u2019s Center for Heterogeneous Integration and Photonics (CHIP) Lab, is shown with UF ECE alumnus Jiachi Ye. \" class=\"wp-image-42027\" srcset=\"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/dalir-ye-500x369.jpg 500w, https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/dalir-ye-768x567.jpg 768w, https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/dalir-ye.jpg 1000w\" sizes=\"auto, (max-width: 500px) 100vw, 500px\" \/><figcaption class=\"wp-element-caption\">Hamed Dalir, associate professor of Electrical and Computer Engineering and director of UF\u2019s Center for Heterogeneous Integration and Photonics (CHIP) Lab, is shown with UF ECE alumnus Jiachi Ye. <\/figcaption><\/figure>\n<\/div>\n\n\n<p>Architecture challenges a long-standing assumption that meaningful computation must wait until after digitization. Instead, the UF system performs the front-end transformations in the optical domain.&nbsp;&nbsp;<\/p>\n\n\n\n<p>A Fourier-domain wavefront processor, shaped by programmable phase control, performs massively parallel feature extraction in light long before the signal reaches electronics. This inversion of computational order plays to the natural strengths of each domain. Photons perform large parallel transforms with almost no energy. Electrons excel at programmable logic and compact memory.&nbsp;&nbsp;<\/p>\n\n\n\n<p>Together they form a hybrid engine that uses physics not power to unlock scale.&nbsp;<\/p>\n\n\n\n<p>As AI models grow, it is these front-end transformations that dominate energy consumption. By executing them in optics, UF\u2019s approach sidesteps the thermal and power crises that define next-generation accelerators.&nbsp;<\/p>\n\n\n\n<p>\u201cThe future of AI accelerators is hybrid,\u201d said Jiachi Ye, UF ECE alumnus and first author of the paper. \u201cPhotons handle the high-throughput parallelism, and electrons perform the precise logic. That\u2019s how you scale performance without scaling heat or power.\u201d&nbsp;<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">The road ahead: From free-space prototypes to multi-GHz integrated photonics&nbsp;<\/h2>\n\n\n<div class=\"wp-block-image\">\n<figure class=\"alignright size-medium\"><img loading=\"lazy\" decoding=\"async\" width=\"500\" height=\"571\" src=\"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/wu-cai-500x571.jpg\" alt=\"UF ECE graduate students Tongyao Wu and Quian Cai working in the lab.  \" class=\"wp-image-42029\" srcset=\"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/wu-cai-500x571.jpg 500w, https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/wu-cai.jpg 700w\" sizes=\"auto, (max-width: 500px) 100vw, 500px\" \/><figcaption class=\"wp-element-caption\">UF ECE graduate students Tongyao Wu and Quian Cai working in the lab.  <\/figcaption><\/figure>\n<\/div>\n\n\n<p>The present demonstration uses a 60 Hz liquid-crystal spatial light modulator (SLM). Its speed ceiling has nothing to do with the architecture and everything to do with the component. Replacing that SLM with a photonic-integrated circuit built around indium\u2013tin-oxide (ITO) modulators moves the design from the world of human-visible updates to radio-frequency operation.&nbsp;<\/p>\n\n\n\n<p>\u201cJumping from 60 Hz to tens of gigabits per second is not an upgrade but a new physics regime, what the free-space SLM accomplishes in a full second, an ITO photonic chip completes in a couple of nanoseconds,\u201d said Tongyao Wu and Abdulaziz Bazammul, both current ECE graduate students.\u00a0<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Vertical integration: Photons and electrons stop cooperating and start unifying&nbsp;<\/h2>\n\n\n\n<p>The long-term vision pushes beyond even planar photonic integration. UF\u2019s roadmap leads toward true three-dimensional heterogeneous integration a stack where multilayer photonic computing planes sit directly beneath sub-10 nm electronic logic.&nbsp;&nbsp;<\/p>\n\n\n\n<p>Through vertical bonding and chip-scale interconnects, the photonic and electronic domains stop being separate technologies and become a single computational system.&nbsp;<\/p>\n\n\n\n<p>\u201cWe have been beginning to work with a major foundry on 7 nm nodes, and shortly we will move to 5 nm, depending on what the system demands,\u201d said Hamed Dalir, Ph.D., ECE associate professor and director of UF\u2019s Center for Heterogeneous Integration and Photonics (CHIP) Lab.&nbsp;<\/p>\n\n\n\n<p>\u201cOur goal is to have the first vertically integrated photonic\u2013electronic prototype ready as early as 2027, showing this architecture not just as an idea, but as a deployable accelerator capable of setting new limits in speed, efficiency, and density,\u201d said Dalir.&nbsp;&nbsp;<\/p>\n","protected":false},"excerpt":{"rendered":"<p>AI has entered a phase where the limits are no longer set by algorithms or transistor counts but by energy and speed.&nbsp;&nbsp; For years, the industry relied on a simple strategy: Deploy more Graphics Processing Units (GPUs). That approach is now colliding with physics. Datacenters push megawatts of power, airborne and orbital systems cannot dissipate [&hellip;]<\/p>\n","protected":false},"author":9,"featured_media":42025,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"single-templates\/single-sidebar-none.php","format":"standard","meta":{"_acf_changed":false,"inline_featured_image":false,"featured_post":"off","footnotes":"","_links_to":"","_links_to_target":""},"categories":[15,61,57],"tags":[499,501],"class_list":["post-42021","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ece","category-research-innovation","category-stories","tag-elham-heidari","tag-hamed-dalir"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.5 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019 - News from Herbert Wertheim College of Engineering<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019 - News from Herbert Wertheim College of Engineering\" \/>\n<meta property=\"og:description\" content=\"AI has entered a phase where the limits are no longer set by algorithms or transistor counts but by energy and speed.&nbsp;&nbsp; For years, the industry relied on a simple strategy: Deploy more Graphics Processing Units (GPUs). That approach is now colliding with physics. Datacenters push megawatts of power, airborne and orbital systems cannot dissipate [&hellip;]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/\" \/>\n<meta property=\"og:site_name\" content=\"News from Herbert Wertheim College of Engineering\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/UFWertheim\/\" \/>\n<meta property=\"article:published_time\" content=\"2025-12-04T19:35:43+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-12-04T21:06:51+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/Elham-Heidari-lab.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"2000\" \/>\n\t<meta property=\"og:image:height\" content=\"1300\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Zachrich Jeng, Sarah\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@ufwertheim\" \/>\n<meta name=\"twitter:site\" content=\"@ufwertheim\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Zachrich Jeng, Sarah\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/\"},\"author\":{\"name\":\"Zachrich Jeng, Sarah\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#\\\/schema\\\/person\\\/0b1e439eb197ec59b8b5e64c8abcf7a7\"},\"headline\":\"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019\",\"datePublished\":\"2025-12-04T19:35:43+00:00\",\"dateModified\":\"2025-12-04T21:06:51+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/\"},\"wordCount\":737,\"publisher\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/wp-content\\\/uploads\\\/sites\\\/249\\\/2025\\\/12\\\/Elham-Heidari-lab.jpg\",\"keywords\":[\"Elham Heidari\",\"Hamed Dalir\"],\"articleSection\":[\"Department of Electrical and Computer Engineering\",\"Research &amp; Innovation\",\"Stories\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/\",\"url\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/\",\"name\":\"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019 - News from Herbert Wertheim College of Engineering\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/wp-content\\\/uploads\\\/sites\\\/249\\\/2025\\\/12\\\/Elham-Heidari-lab.jpg\",\"datePublished\":\"2025-12-04T19:35:43+00:00\",\"dateModified\":\"2025-12-04T21:06:51+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#primaryimage\",\"url\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/wp-content\\\/uploads\\\/sites\\\/249\\\/2025\\\/12\\\/Elham-Heidari-lab.jpg\",\"contentUrl\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/wp-content\\\/uploads\\\/sites\\\/249\\\/2025\\\/12\\\/Elham-Heidari-lab.jpg\",\"width\":2000,\"height\":1300,\"caption\":\"UF students work in Elham Heidari\u2019s lab.\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/research-innovation\\\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#website\",\"url\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/\",\"name\":\"News from Herbert Wertheim College of Engineering\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#organization\",\"name\":\"Herbert Wertheim College of Engineering\",\"url\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/wp-content\\\/uploads\\\/sites\\\/249\\\/2025\\\/07\\\/hwcoe-avatar.png\",\"contentUrl\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/wp-content\\\/uploads\\\/sites\\\/249\\\/2025\\\/07\\\/hwcoe-avatar.png\",\"width\":720,\"height\":720,\"caption\":\"Herbert Wertheim College of Engineering\"},\"image\":{\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/UFWertheim\\\/\",\"https:\\\/\\\/x.com\\\/ufwertheim\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/#\\\/schema\\\/person\\\/0b1e439eb197ec59b8b5e64c8abcf7a7\",\"name\":\"Zachrich Jeng, Sarah\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/d7888e076c232465255e89642610ba5d3face1a9e94d7d151624fde5e97ea8fb?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/d7888e076c232465255e89642610ba5d3face1a9e94d7d151624fde5e97ea8fb?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/d7888e076c232465255e89642610ba5d3face1a9e94d7d151624fde5e97ea8fb?s=96&d=mm&r=g\",\"caption\":\"Zachrich Jeng, Sarah\"},\"url\":\"https:\\\/\\\/www.eng.ufl.edu\\\/news\\\/author\\\/sarahz\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019 - News from Herbert Wertheim College of Engineering","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/","og_locale":"en_US","og_type":"article","og_title":"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019 - News from Herbert Wertheim College of Engineering","og_description":"AI has entered a phase where the limits are no longer set by algorithms or transistor counts but by energy and speed.&nbsp;&nbsp; For years, the industry relied on a simple strategy: Deploy more Graphics Processing Units (GPUs). That approach is now colliding with physics. Datacenters push megawatts of power, airborne and orbital systems cannot dissipate [&hellip;]","og_url":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/","og_site_name":"News from Herbert Wertheim College of Engineering","article_publisher":"https:\/\/www.facebook.com\/UFWertheim\/","article_published_time":"2025-12-04T19:35:43+00:00","article_modified_time":"2025-12-04T21:06:51+00:00","og_image":[{"width":2000,"height":1300,"url":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/Elham-Heidari-lab.jpg","type":"image\/jpeg"}],"author":"Zachrich Jeng, Sarah","twitter_card":"summary_large_image","twitter_creator":"@ufwertheim","twitter_site":"@ufwertheim","twitter_misc":{"Written by":"Zachrich Jeng, Sarah","Est. reading time":"4 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#article","isPartOf":{"@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/"},"author":{"name":"Zachrich Jeng, Sarah","@id":"https:\/\/www.eng.ufl.edu\/news\/#\/schema\/person\/0b1e439eb197ec59b8b5e64c8abcf7a7"},"headline":"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019","datePublished":"2025-12-04T19:35:43+00:00","dateModified":"2025-12-04T21:06:51+00:00","mainEntityOfPage":{"@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/"},"wordCount":737,"publisher":{"@id":"https:\/\/www.eng.ufl.edu\/news\/#organization"},"image":{"@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#primaryimage"},"thumbnailUrl":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/Elham-Heidari-lab.jpg","keywords":["Elham Heidari","Hamed Dalir"],"articleSection":["Department of Electrical and Computer Engineering","Research &amp; Innovation","Stories"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/","url":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/","name":"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019 - News from Herbert Wertheim College of Engineering","isPartOf":{"@id":"https:\/\/www.eng.ufl.edu\/news\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#primaryimage"},"image":{"@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#primaryimage"},"thumbnailUrl":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/Elham-Heidari-lab.jpg","datePublished":"2025-12-04T19:35:43+00:00","dateModified":"2025-12-04T21:06:51+00:00","breadcrumb":{"@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#primaryimage","url":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/Elham-Heidari-lab.jpg","contentUrl":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/12\/Elham-Heidari-lab.jpg","width":2000,"height":1300,"caption":"UF students work in Elham Heidari\u2019s lab."},{"@type":"BreadcrumbList","@id":"https:\/\/www.eng.ufl.edu\/news\/research-innovation\/americas-next-ai-accelerator-constraint-heat-power-and-the-end-of-just-add-gpus\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.eng.ufl.edu\/news\/"},{"@type":"ListItem","position":2,"name":"America\u2019s next AI accelerator constraint: Heat, power and the end of \u2018just add GPUs\u2019"}]},{"@type":"WebSite","@id":"https:\/\/www.eng.ufl.edu\/news\/#website","url":"https:\/\/www.eng.ufl.edu\/news\/","name":"News from Herbert Wertheim College of Engineering","description":"","publisher":{"@id":"https:\/\/www.eng.ufl.edu\/news\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.eng.ufl.edu\/news\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/www.eng.ufl.edu\/news\/#organization","name":"Herbert Wertheim College of Engineering","url":"https:\/\/www.eng.ufl.edu\/news\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.eng.ufl.edu\/news\/#\/schema\/logo\/image\/","url":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/07\/hwcoe-avatar.png","contentUrl":"https:\/\/www.eng.ufl.edu\/news\/wp-content\/uploads\/sites\/249\/2025\/07\/hwcoe-avatar.png","width":720,"height":720,"caption":"Herbert Wertheim College of Engineering"},"image":{"@id":"https:\/\/www.eng.ufl.edu\/news\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/UFWertheim\/","https:\/\/x.com\/ufwertheim"]},{"@type":"Person","@id":"https:\/\/www.eng.ufl.edu\/news\/#\/schema\/person\/0b1e439eb197ec59b8b5e64c8abcf7a7","name":"Zachrich Jeng, Sarah","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/d7888e076c232465255e89642610ba5d3face1a9e94d7d151624fde5e97ea8fb?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/d7888e076c232465255e89642610ba5d3face1a9e94d7d151624fde5e97ea8fb?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/d7888e076c232465255e89642610ba5d3face1a9e94d7d151624fde5e97ea8fb?s=96&d=mm&r=g","caption":"Zachrich Jeng, Sarah"},"url":"https:\/\/www.eng.ufl.edu\/news\/author\/sarahz\/"}]}},"_links":{"self":[{"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/posts\/42021","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/users\/9"}],"replies":[{"embeddable":true,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/comments?post=42021"}],"version-history":[{"count":8,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/posts\/42021\/revisions"}],"predecessor-version":[{"id":42047,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/posts\/42021\/revisions\/42047"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/media\/42025"}],"wp:attachment":[{"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/media?parent=42021"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/categories?post=42021"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.eng.ufl.edu\/news\/wp-json\/wp\/v2\/tags?post=42021"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}