Previously implemented in collaboration with NVIDIA to support the NVIDIA DGX™ GB300 system, the highest-performing and highest-scale AI system available for enterprises, the Vertiv™ SmartIT solution integrates rack infrastructure, DC power conversion and distribution, direct-to-chip liquid cooling, in-cabinet cooling distribution, and monitoring into a factory-validated design. Integrating and validating these systems before shipment reduces on-site installation complexity and simplifies integration with existing power and cooling infrastructure. The solution is now available for AI deployments worldwide.

“AI is radically transforming the demands placed on data center infrastructure, requiring closer coordination between power, cooling, and computing systems to support increasingly complex workloads,” said Scott Armul, Director of Product and Technology at Vertiv. “Vertiv™ SmartIT reflects our approach to solving this challenge by bringing these disciplines together in a factory-validated architecture. By leveraging our mission-critical infrastructure expertise, we help customers simplify deployment, achieve greater consistency across their AI environments, and establish a foundation that can evolve with their computing needs.”.

Vertiv™ SmartIT helps customers address key deployment challenges, including:
Reduced deployment times: Factory validation replaces on-site integration, limiting commissioning to facility connections.
Simplified vendor coordination: Customers benefit from a single order, a single partner, and an architecture that encompasses cabinets, power, cooling, and monitoring.
Adapting existing facilities for AI: Existing enterprise facilities and Class A commercial buildings can support AI-like densities without a complete rebuild.
Scalability through a replicable architecture: The same modular building block allows growth from a single rack to multi-rack deployments without infrastructure redesign between phases.
Enhanced operational visibility: Integrated leak detection, automated response, and continuous monitoring improve operational visibility from the rack to the entire facility. These capabilities are built into the solution from the outset, eliminating the need to combine separate tools from different vendors.

Vertiv™ SmartIT is a scalable, pre-engineered infrastructure platform for high-density AI deployments that supports NVIDIA GB300 NVL72, NVIDIA Vera Rubin NVL72, NVIDIA DGX™ Rubin NVL8, and NVIDIA HGX™ Rubin NVL8, based on the NVIDIA MGX™ and Open Compute Project (OCP) ORV3 architectures.

Designed to scale from a single rack to multi-rack deployments, it enables customers to deploy and expand AI infrastructure within a common architecture, while integrating power, cooling, monitoring, and controls. Integrated monitoring, leak detection, and support for Redfish, SNMP, and Modbus simplify operations, while Vertiv™ Global Services provides support for implementation, commissioning, and lifecycle management.

Vertiv™ SmartIT can be deployed as part of a broader Vertiv infrastructure architecture, with complementary solutions that extend management and connectivity capabilities beyond the rack. Vertiv™ SmartRun extends integration across the row, while Vertiv™ Rack Power Distribution Units (rPDUs) provide power distribution and monitoring. Vertiv™ Unify offers integrated monitoring, analytics, and controls across the entire electrical and cooling infrastructure, including leak detection and coordinated response control. Together, these solutions help customers scale from single Vertiv™ SmartIT deployments to larger AI environments.

Vertiv showcased a Vertiv™ SmartIT demo unit with NVIDIA HGX™ Rubin NVL8 at NVIDIA GTC Berlin in October, demonstrating an integrated infrastructure architecture designed to support NVIDIA HGX™ Rubin NVL8 deployments. This allows customers to start with a configuration tailored to their needs and progressively scale their installations within a common architecture as AI demands grow.