- Colosseum will support highly regulated sectors with up to 115 exaflops of AI compute power.
- Vertiv’s 360AI platform will deliver scalable power and cooling infrastructure for the system.
What happened: Vertiv, Nvidia, and iGenius to deploy Colosseum AI supercomputer in Italy
Vertiv, Nvidia, and iGenius have announced a strategic collaboration to deploy Colosseum, a large-scale AI supercomputer, in southern Italy. The project will leverage Nvidia’s DGX platform, featuring Grace Blackwell Superchips, and is expected to be operational in 2025.
The supercomputer will run on Vertiv’s 360AI reference architecture, a scalable solution designed to handle power and thermal requirements for Nvidia GB200 NVL72 systems. The 7MW deployment can support up to 132kW per rack, addressing the needs of demanding AI workloads.
Although the companies haven’t disclosed the full compute specifications, earlier reports suggest the system may include 80 Nvidia GB200 NVL72 units and could deliver up to 115 exaflops of computing power. The infrastructure will serve sectors such as healthcare, finance, and public administration — industries that require secure, high-performance AI capabilities.
Also read: xAI expands colossus supercomputer in Memphis
Also read: UK unveils $12.6M Nvidia Grace-Based supercomputer
Why it’s important
This partnership highlights Europe’s increasing focus on AI sovereignty and infrastructure development. Deploying Colosseum in Italy represents a significant investment in localised AI computing power, reducing reliance on global cloud providers and improving data compliance.
The supercomputer is expected to enhance innovation and operational efficiency in critical sectors while supporting Italy’s and Europe’s digital transformation strategies. It also signals Vertiv’s ambition to provide the backbone for future AI infrastructure deployments worldwide.
