Uvation MarketplaceMarketplace

NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 and Quantum-X800 InfiniBand

NVIDIA DGX SuperPOD combines DGX Vera Rubin NVL72 racks with Quantum-X800 InfiniBand networking. Each rack integrates 72 Rubin GPUs, 36 Vera CPUs, and 75 TB of fast memory, with liquid cooling and high-density power for large-scale AI training and inference.

72x
NVIDIA Rubin GPUs per rack
36x
NVIDIA Vera CPUs per rack
75 TB
total fast memory per rack

Rack-Scale AI,
Connected for Scaling

Combine GPU-dense compute, high-bandwidth networking, and integrated rack infrastructure to support large AI workloads across an eight-rack scalable unit.

Rack-Scale Compute

Each DGX Vera Rubin NVL72 rack combines 72 Rubin GPUs and 36 Vera CPUs for demanding
AI compute workloads.

High-Capacity Memory and Storage

75 TB of fast memory per rack, paired with E1.S NVMe OS and local storage drives.

Integrated Networking and Cooling

ConnectX-9 adapters, BlueField-4 DPUs, Quantum-X800 InfiniBand networking, and liquid-cooled rack infrastructure support scalable deployment.

Unified Compute, Memory, and Fabric

Bring GPU compute, CPU processing, memory, storage, and high-speed interconnects together in a rack-scale AI platform designed for multi-rack deployment.

NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 and Quantum-X800 InfiniBand architecture
8x
NVIDIA Rubin GPUs per system
2x
NVIDIA Vera CPUs
2.3 TB
TB high-bandwidth GPU memory (HBM4) per system
800 Gb/s
Quantum-X800 InfiniBand compute fabric

Workloads

Run compute-intensive AI workloads on rack-scale infrastructure designed for large models, high-throughput processing, and multi-rack expansion.

576x

Interconnected NVIDIA Rubin GPUs across 72 systems - unified AI factory pooling massive compute and fast memory for frontier model training without communication delays.

72x

Distributed computing systems synchronized via dual processors, ultra-fast network adapters, management controllers, and NVIDIA Mission Control software for seamless cluster coordination.

167k

Response tokens per second generated at scale - reading billions of prompt tokens per second across the cluster to deliver instantaneous AI inference for millions of simultaneous users.

Engineered to dominate

Built for Rack-Scale AI Performance

Combine dense GPU compute, high-capacity memory, dedicated networking, and liquid-cooled infrastructure in a scalable platform for demanding AI deployments.

Exploded view of the NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 and Quantum-X800 InfiniBand
8 DGX Vera Rubin NVL72
racks per scalable unit
4
power shelves per DGX rack
Up to 110 kW
per power shelf
ConnectX-9
compute and BlueField-4 storage/management connections
Exploded view of the NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 and Quantum-X800 InfiniBand
8 DGX Vera Rubin NVL72
racks per scalable unit
4
power shelves per DGX rack
Up to 110 kW
per power shelf
ConnectX-9
compute and BlueField-4 storage/management connections

The rack, itemized

SUBSYSTEM

SPECIFICATION / CONFIGURATION DETAILS

Processor

36× Vera CPUs per DGX rack

Operating System

DGX OS; Ubuntu supported per DGX rack

Chipset

N/A

Accelerators / GPU

72× Rubin GPUs per rack; 8 racks per scalable unit

Memory

75 TB fast memory per DGX rack

Primary Storage

1× 2 TB E1.S NVMe OS drive per compute tray

Secondary Storage

4× 7.68 TB E1.S NVMe local drives per compute tray; certified partner shared storage

Security

Vera Rubin NVL72 confidential computing

Chassis

Multi-rack, liquid-cooled deployment

Power Supply

4× power shelves per DGX rack; up to 110 kW per shelf

Ports

ConnectX-9 compute and BlueField-4 storage/management connections

Slots / PCIe Architecture

N/A

Bezel

N/A

Form Factor

8 DGX Vera Rubin NVL72 racks per scalable unit

Network Card

NVIDIA ConnectX-9 adapters and NVIDIA BlueField-4 DPUs

Warranty

N/A

NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 and Quantum-X800 InfiniBand power and networking

Power Fabric

~24 Kw

power usage per liquid-cooled DGX Rubin NVL8 server

72x

DGX Rubin NVL8 systems organized into a single SuperPOD scalable unit

DC

Busbar compute racks supported by rack power shelves, busbars, and PDUs across the deployment.

Nuclear-Ready, Plug and Play AI Factories

Bypass grid constraints with self-contained, high-density inference ready infrastructure directly adjacent to clean, base load nuclear power.

Global Deployment Engine

Any Infra, Anywhere in the world.

Global Shipping and Deployment Capability treats advanced infrastructure as mobile, modular freight.

Packaged complex computing nodes, cooling manifolds, and power distribution units into certified, standard form factors.

360° Service Ecosystem

Unifying Legacy with Next-Gen Compute.

Vertically integrated service ecosystem that serves as an operational bridge between legacy environments and next-generation, AI-native infrastructure.

Finance and Leasing

Capital-as-a-Service for High-Density Compute

Convert upfront capital expenditures (CapEx) into predictable operating expenses (OpEx).

Uvation allows enterprises and sovereign entities to deploy multi-megawatt AI factories and advanced GPU clusters without liquidity constraint.