---
title: "Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS"
description: "Power-constrained AI factories measure efficiency by AI output per watt, because facility overhead, rack losses, and operational inefficiency reduce the power available to compute. In an illustrative 100 MW budget, 60% reaches AI load, while static rack provisioning can strand headroom that another rack could use. NVIDIA DSX MaxLPS addresses this with dynamic power allocation, workload-profile power optimization, and 45° C thermal and site design, including warm-water liquid cooling. Its Dynamic Power Software, currently in Developer Preview, models topology from utility to GPUs and reallocates unused capacity within policy-defined budgets, with optional DSX Exchange integration. For Vera Rubin NVL72, NVIDIA projects up to 40% more Rubin GPU capacity within the same power budget, while validation compares unmanaged and managed runs using throughput, latency, errors, power, utilization, and compliance."
---

# Maximizing AI Factory Performance per Watt with NVIDIA DSX MaxLPS

[NVIDIA](https://yomu.fyi/company/nvidia-developer-blog) · Tanya Lenz · Aug 24, 2026

**Type:** Explainer

## Summary

Power-constrained AI factories measure efficiency by AI output per watt, because facility overhead, rack losses, and operational inefficiency reduce the power available to compute. In an illustrative 100 MW budget, 60% reaches AI load, while static rack provisioning can strand headroom that another rack could use. NVIDIA DSX MaxLPS addresses this with dynamic power allocation, workload-profile power optimization, and 45° C thermal and site design, including warm-water liquid cooling. Its Dynamic Power Software, currently in Developer Preview, models topology from utility to GPUs and reallocates unused capacity within policy-defined budgets, with optional DSX Exchange integration. For Vera Rubin NVL72, NVIDIA projects up to 40% more Rubin GPU capacity within the same power budget, while validation compares unmanaged and managed runs using throughput, latency, errors, power, utilization, and compliance.

## Context

AI factories are power-constrained, and only part of delivered site power reaches AI compute after facility overhead, rack losses, cooling, networking, storage, backup, failures, restarts, checkpointing, and other inefficiencies. Static rack provisioning reserves maximum rack power for peak demand, leaving unused rack-level headroom stranded rather than available to workloads elsewhere.

## Approach / What changed

NVIDIA DSX MaxLPS combines dynamic power allocation, workload profile power solutions, application and GPU configuration, and 45° C thermal and site design. Dynamic Power Software models the data-center topology, monitors power telemetry, reallocates unused capacity within defined policies, and adapts to power events. Warm-water liquid cooling and optional DSX Exchange integration address cooling efficiency and facility signals.

## Takeaways

- In the illustrative 100 MW power-budget waterfall, 20 MW goes to facility overhead, 10 MW to rack losses, and 10 MW to operational inefficiency, leaving 60 MW for AI load.
- Dynamic Power Software continuously compares allocated power with actual consumption and can reallocate unused headroom within managed groups without changing the site power envelope.
- NVIDIA projects that MaxLPS combined with data-center power planning can fit up to 40% more Rubin GPU capacity in the same power budget for Vera Rubin NVL72 AI factories.

**Tags:** [Architecture](https://yomu.fyi/topic/architecture), [Monitoring](https://yomu.fyi/topic/monitoring), [Performance](https://yomu.fyi/topic/performance), [Scalability](https://yomu.fyi/topic/scalability)

- Source: [NVIDIA](https://developer.nvidia.com/blog/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps)
- Source URL: https://developer.nvidia.com/blog/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps
- Ingested by Yomu: 2026-08-29T01:38:39.003Z

[Read original post](https://developer.nvidia.com/blog/maximizing-ai-factory-performance-per-watt-with-nvidia-dsx-maxlps)