Vertiv has unveiled new configurations of its MegaMod HDX system. As demand rises for data centres that can support energy-intensive artificial intelligence and high-performance computing workloads. The prefabricated solution combines power and cooling infrastructure in modular units. That are designed to deploy faster and scale more easily than traditional builds.
The announcement comes as operators across Africa and globally face growing pressure from AI workloads. They have push rack densities far beyond earlier standards. Training and running large models now requires much higher power and far more advanced cooling than air-based systems alone can deliver.
Vertiv MegaMod HDX, AI Data Center Infrastructure, High-Performance Computing [HPC]
The updated MegaMod HDX range introduces two formats. A compact version supports up to 13 racks with power capacity of up to 1.25 megawatts. A larger “combo” configuration scales to as many as 144 racks and up to 10 megawatts. Both support rack densities from 50 kilowatts to more than 100 kilowatts per rack, levels increasingly common in GPU-heavy environments.
At the core of the design is a hybrid cooling approach. Direct-to-chip liquid cooling is paired with air cooling to manage the intense heat generated by AI processors. This setup allows operators to support high-density clusters while managing energy use and floor space more efficiently.
The systems are delivered as prefabricated pods, built and tested in the factory before deployment. This approach shortens installation timelines and offers more predictable costs, an advantage for operators planning phased expansion. Redundant power architecture allows systems to keep running even if one module is taken offline, while a thermal buffer system helps maintain stable operations during maintenance or load changes.
Vertiv says the new configurations are available globally and are supported by its wider portfolio of power, cooling and monitoring technologies. These include uninterruptible power supplies, cooling distribution units and infrastructure monitoring tools that are already used in many large data centres.
