REIMAGINED FACILITY & EXPERT TECHNOLOGY DESIGN
LEDG led the design strategy and provided technical oversight for one of the first large-scale deployments of NVIDIA’s GB300 technology in the United States. This landmark project represents a major step forward in next-generation AI infrastructure, supporting the rapid expansion of high-density GPU workloads at unprecedented scale.
The data center campus is being developed to support a total capacity of 100MW. In the initial 24MW phase, LEDG’s customer is deploying more than 10,000 NVIDIA GB300 GPUs, with thousands more planned across subsequent phases. From the start, the project demanded expertise and a forward-looking design approach capable of supporting aggressive growth while maintaining reliability, performance, and operational resilience.
Delivering the first GB300 deployment at scale would present challenges for any project team, but the complexity was heightened by the choice of site. Rather than new construction, the customer selected an existing building that required full repurposing to support an AI cloud environment, compounded with an accelerated design and construction schedule. This transformation required extensive modifications across nearly every building system, including the building envelope, structural support, site layout, equipment planning, and on-site electrical infrastructure and distribution.
LEDG played a pivotal role in reimagining the facility to meet the extreme power density and thermal requirements of GB300 systems. The team developed innovative solutions to efficiently utilize limited underfloor and overhead clearances, coordinating daily with NVIDIA, Supermicro, and the customer to maximize GPU workload capacity without compromising redundancy, security, or system performance.
Beyond infrastructure design, LEDG worked closely with the tenant’s operations organization to understand how they, the tenant, would manage and operate GPU workloads, identify technical details for ongoing monitoring, and develop detailed testing procedures to commission critical equipment.
By combining technical leadership and design expertise, cross-disciplinary coordination, and a strong understanding of AI workload operations, LEDG helped enable a first-of-its-kind GB300 deployment, laying the foundation for future growth and establishing a new benchmark for AI data center design in the United States.