Lancium and NVIDIA Partner to Build Gigawatt-Scale AI Factories

The partnership aims to integrate NVIDIA’s full-stack AI platform—covering accelerated computing, networking, and software—directly into Lancium’s massive campus footprint. By utilizing NVIDIA’s DSX reference designs, the firms intend to optimize performance metrics, specifically targeting faster time-to-first-token and improved cost-efficiency. Lancium plans to deploy NVIDIA’s DSX MaxLPS technology to increase GPU density by up to 40% within existing power constraints, while DSX Flex will allow the campuses to dynamically adjust consumption based on grid conditions.

Michael McNamara, CEO of Lancium, emphasized that the company has spent years securing the land and energy infrastructure required to deliver compute capacity at unprecedented scales. Nico Caprez, NVIDIA’s vice president of global AI infrastructure growth, described the move as a critical step in building the essential infrastructure for a new industrial era. Blackstone, which holds a significant stake in Lancium, views the deal as a convergence of its two primary investment pillars: energy transition and AI infrastructure.