Vertiv, a global leader in critical digital infrastructure, today announced the deployment of integrated power, liquid cooling, rack infrastructure and installation services supporting the Naval Postgraduate School’s (NPS) new NVIDIA DGX™ GB300 system. The project establishes an advanced, locally operated AI environment for education, research, digital engineering, modeling and simulation. It also demonstrates how an existing facility can be transformed to support the electrical, thermal and operational requirements of next-generation accelerated computing.
NVIDIA provided the DGX GB300 system to NPS, while Vertiv provided and deployed the supporting physical infrastructure, engineering expertise and installation services required to bring the high-density, liquid-cooled environment online.
“Advanced AI systems are redefining the physical requirements of the data center,” said Gio Albertazzi, CEO of Vertiv. “The NPS deployment demonstrates what is required to turn rack-scale accelerated computing into operational AI capability. By coordinating compute requirements with power, liquid cooling, installation and lifecycle readiness, organizations can deploy AI faster, reduce integration risk and create a repeatable foundation for future growth.”
Turning AI compute into operational capability
Rack-scale AI systems are changing how organizations plan and deploy digital infrastructure. Enterprise AI requires a complete physical configuration that coordinates compute, power, liquid cooling, racks, controls, installation, commissioning and lifecycle support.
The NPS project provides a practical example of how an existing facility can be modernized to support advanced AI capability without waiting for an entirely new purpose-built environment. For enterprise, government, research and institutional customers, this approach can reduce deployment complexity, accelerate time to productive AI capacity and establish a scalable foundation for future expansion.
High-density AI systems require substantially more power and cooling capacity than traditional IT environments. Deployments can involve power distribution upgrades, liquid-cooling integration, facility modifications, fluid management and specialized commissioning. Designing these elements as a coordinated system can reduce integration risk and improve operational readiness.
A repeatable physical configuration for Enterprise AI
The Vertiv™ SmartIT pre-integrated, rack-level AI infrastructure portfolio is designed to accelerate AI deployment across a range of physical infrastructure configurations, from standalone enterprise systems to fully integrated whitespace solutions. For larger-scale environments, Vertiv SmartIT can be integrated with Vertiv™ SmartRun to extend the same pre-engineered architecture converging the whitespace, coordinating rack infrastructure, power, cooling, controls and services as one deployable system. This modular approach provides customers with a repeatable path from an initial AI deployment to broader capacity expansion in both existing facilities and new builds.
Working with NVIDIA, NPS, the NPS Foundation and a broad ecosystem of partners, Vertiv designed and delivered a three-rack Vertiv SmartIT integrated AI infrastructure configuration to support the NVIDIA system. The Vertiv SmartIT configuration integrates:
- Power protection and distribution
- AI infrastructure racks
- Liquid-cooling technologies
- Fluid management and monitoring
- Facility and system interfaces
- Installation sequencing
- Testing and commissioning support
“Deploying an AI factory requires more than accelerated computing — it requires power, cooling and physical infrastructure engineered to work together,” said Charlie Boyle, vice president of DGX systems at NVIDIA. “Together with Vertiv, we’re bringing the NVIDIA DGX GB300 system online at the Naval Postgraduate School, giving researchers and faculty the infrastructure to advance AI research, digital engineering and education.”
Read Also:
