AI Factories That Use Less Power and Water, Built in Days | NVIDIA DSX
NVIDIA DSX Pod ships factory-built AI infrastructure deployable on-site in days, not months
“Once site utilities are ready, estimated on-site deployment goes from several months to days.”
NVIDIA announced DSX Pod, a factory-integrated AI factory unit built around Vera Rubin NVL72 that arrives pre-cabled and tested in transportable containers, cutting on-site deployment from months to days. The system pairs a dynamic power control loop (MaxLPS) that cuts energy use 30% with waterless dry-cooler technology projected to save 54 million liters of water daily at 2 GW scale. This matters because it addresses the two biggest constraints on AI capacity expansion—build time and resource consumption—while lowering the barrier for enterprises and hyperscalers to stand up new compute.