Vertiv Unveils Integrated Infrastructure Solution Designed for High-Density Artificial Intelligence Workloads

In late October 2026, Vertiv Holdings Co introduced a new infrastructure solution dubbed SmartIT, specifically engineered to handle the intensive requirements of high-density artificial intelligence environments. This platform utilizes a pre-engineered design that undergoes validation at the factory to ensure it is ready for immediate operation upon delivery. Developed to support architectures such as the Open Compute Project ORV3 and NVIDIA MGX, the system is intended for a diverse range of users, including research organizations, colocation firms, and large-scale enterprises. It is designed to be compatible with both newly constructed data centers and existing facilities, allowing for advanced AI integration without the need for extensive structural renovations.

Technically, the platform incorporates a comprehensive suite of hardware within a single architecture, featuring rack infrastructure, power distribution, and direct-to-chip liquid cooling systems. It is built to support high-performance hardware such as the NVIDIA GB300 NVL72 and upcoming Rubin-based systems. A key advantage of this model is its scalability, as it can function as a standalone rack or expand into multi-rack pods. The unit measures approximately 90 inches in height and 23.6 inches in width, weighing over 400 pounds. By consolidating power conversion, thermal management, and monitoring into one package, Vertiv aims to reduce the complexities typically associated with site-specific integration.

The system includes sophisticated monitoring capabilities, supporting industry-standard protocols like Modbus, SNMP, and Redfish to oversee coolant distribution and power shelves. Safety features, such as automated shunt-trip protections and proactive leak detection, are integrated into the design. According to Vertiv’s leadership, the primary goal of SmartIT is to unify disparate disciplines—specifically cooling, power, and IT systems—into a singular, validated framework. This approach is intended to help customers maintain consistency across their computing environments while streamlining the deployment process.

Beyond the hardware itself, Vertiv offers a suite of support services to manage the platform throughout its lifecycle, from initial commissioning to ongoing maintenance. The solution was recently demonstrated at the NVIDIA GTC event in Berlin, where it was paired with an HGX Rubin unit to showcase how the infrastructure accommodates evolving AI demands. The platform is currently available for global deployment across various sectors, including healthcare, finance, education, and government.

---

Read original at Unite.AI.