--- title: Maximizing AI Factory Performance per Watt with NVIDIA... description: NVIDIA DSX MaxLPS is a suite of chip, thermal, system, and software technologies designed to maximize AI factory throughput within a fixed power budget,... canonical: https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS | daily.dev og:description: NVIDIA DSX MaxLPS is a suite of chip, thermal, system, and software technologies designed to maximize AI factory throughput within a fixed power budget,... og:url: https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq og:image: https://api.daily.dev/og/posts/iAMH3xeqq.png og:image:alt: Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS **[NVIDIA Developer](https://daily.dev/sources/nvidiadev)** · 12 min read · 0 upvotes · 0 comments ## Summary NVIDIA DSX MaxLPS is a suite of chip, thermal, system, and software technologies designed to maximize AI factory throughput within a fixed power budget, addressing the fact that static rack power provisioning strands unused headroom. Key components include Dynamic Power Software (DPS), which dynamically reallocates unused power across racks; workload profile power solutions for inference and training; and 45°C liquid cooling that reduces cooling overhead via free cooling. For Vera Rubin NVL72 factories, NVIDIA projects MaxLPS can enable up to 40% more GPU capacity within the same power budget, with measured performance-per-watt gains of roughly 1.3-1.5x on GB200 NVL72 and Vera Rubin NVL72 using DeepSeek-R1 and Kimi-K2.5 workloads. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much more GPU capacity can NVIDIA DSX MaxLPS enable for Vera Rubin NVL72 AI factories within the same power budget? NVIDIA DSX MaxLPS can enable up to 40% more Vera Rubin NVL72 GPU capacity within the same fixed power budget. This comes from combining dynamic power allocation, workload profile power solutions, and 45°C liquid cooling to reduce provisioned rack power from 136 kW to 101 kW, enabling 35% more racks while preserving throughput, with performance-per-watt improving roughly 1.3-1.4x. _daily.dev helps teams planning AI factory capacity keep up with infrastructure efficiency techniques like these._ ### What is NVIDIA Dynamic Power Software (DPS) and how does it reduce stranded power in a data center? NVIDIA Dynamic Power Software (DPS), currently in Developer Preview, is a power management system that models data center topology from the utility level down to racks, nodes, and GPUs, letting operators define resource groups, power budgets, and policies. It continuously compares allocated power against actual consumption and reallocates unused headroom from underutilized GPUs or racks to others in the same group, without changing the overall site power envelope. _engineers evaluating dynamic power management tools can track releases like this via daily.dev._ ### Why do NVIDIA Vera Rubin NVL72 racks use 45°C liquid cooling inlet temperatures instead of colder coolant? Vera Rubin NVL72 racks are designed for 45°C liquid-cooling inlet operation because warmer coolant lets facilities rely more on free cooling using outside air or water, cutting reliance on energy-intensive chillers or water-intensive evaporative coolers. This reduces average annual power usage effectiveness (PUE), freeing power budgeted for worst-case cooling scenarios to be redirected toward compute during much of the year. _data center engineers weighing cooling design tradeoffs follow infrastructure efficiency updates on daily.dev._ ## Similar posts on daily.dev - [How NVIDIA DSX MaxLPS Maximizes AI Factory Throughput and Efficiency](https://daily.dev/posts/how-nvidia-dsx-maxlps-maximizes-ai-factory-throughput-and-efficiency-8wqurs3ys) · NVIDIA Developer · 0 upvotes · 0 comments - [Maximize AI Factory Energy Efficiency Through Full-Stack Inference and Training Optimizations](https://daily.dev/posts/maximize-ai-factory-energy-efficiency-through-full-stack-inference-and-training-optimizations-lwusm4jtd) · NVIDIA Developer · 0 upvotes · 0 comments - [NVIDIA Launches DSX Ready to Qualify Power and Cooling Products for AI Factories](https://daily.dev/posts/nvidia-launches-dsx-ready-to-qualify-power-and-cooling-products-for-ai-factories-h9hyxrocg) · NVIDIA · 0 upvotes · 0 comments - [Scaling Token Factory Revenue and AI Efficiency by Maximizing Performance per Watt](https://daily.dev/posts/scaling-token-factory-revenue-and-ai-efficiency-by-maximizing-performance-per-watt-2qzgn7uql) · NVIDIA Developer · 0 upvotes · 0 comments --- Tags: [#nvidia](https://daily.dev/tags/nvidia), [#ai-infrastructure](https://daily.dev/tags/ai-infrastructure) [View this post on daily.dev](https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS","url":"https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq"},"datePublished":"2026-08-21T15:04:33.004Z","dateModified":"2026-08-21T15:24:13.251Z","description":"NVIDIA DSX MaxLPS is a suite of chip, thermal, system, and software technologies designed to maximize AI factory throughput within a fixed power budget,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/d801fbaac9eb3cde944d4284954d47df?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/d801fbaac9eb3cde944d4284954d47df?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"NVIDIA Developer","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"NVIDIA Developer","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/86e45aab42ba48ce83103d01b1119910","url":"https://daily.dev/sources/nvidiadev"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"nvidia,ai-infrastructure","timeRequired":"PT12M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"NVIDIA Developer","item":"https://daily.dev/sources/nvidiadev"},{"@type":"ListItem","position":3,"name":"Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps-iamh3xeqq#faq","mainEntity":[{"@type":"Question","name":"How much more GPU capacity can NVIDIA DSX MaxLPS enable for Vera Rubin NVL72 AI factories within the same power budget?","acceptedAnswer":{"@type":"Answer","text":"NVIDIA DSX MaxLPS can enable up to 40% more Vera Rubin NVL72 GPU capacity within the same fixed power budget. This comes from combining dynamic power allocation, workload profile power solutions, and 45°C liquid cooling to reduce provisioned rack power from 136 kW to 101 kW, enabling 35% more racks while preserving throughput, with performance-per-watt improving roughly 1.3-1.4x. daily.dev helps teams planning AI factory capacity keep up with infrastructure efficiency techniques like these."}},{"@type":"Question","name":"What is NVIDIA Dynamic Power Software (DPS) and how does it reduce stranded power in a data center?","acceptedAnswer":{"@type":"Answer","text":"NVIDIA Dynamic Power Software (DPS), currently in Developer Preview, is a power management system that models data center topology from the utility level down to racks, nodes, and GPUs, letting operators define resource groups, power budgets, and policies. It continuously compares allocated power against actual consumption and reallocates unused headroom from underutilized GPUs or racks to others in the same group, without changing the overall site power envelope. engineers evaluating dynamic power management tools can track releases like this via daily.dev."}},{"@type":"Question","name":"Why do NVIDIA Vera Rubin NVL72 racks use 45°C liquid cooling inlet temperatures instead of colder coolant?","acceptedAnswer":{"@type":"Answer","text":"Vera Rubin NVL72 racks are designed for 45°C liquid-cooling inlet operation because warmer coolant lets facilities rely more on free cooling using outside air or water, cutting reliance on energy-intensive chillers or water-intensive evaporative coolers. This reduces average annual power usage effectiveness (PUE), freeing power budgeted for worst-case cooling scenarios to be redirected toward compute during much of the year. data center engineers weighing cooling design tradeoffs follow infrastructure efficiency updates on daily.dev."}}]} ```