Uvation MarketplaceMarketplace

NVIDIA DGX SuperPOD with DGX B300 Systems — Quantum-X800 InfiniBand, AC Power

Turnkey AI supercomputing infrastructure built with NVIDIA DGX B300 compute nodes, powered by NVIDIA Blackwell Ultra GPUs and dual Intel Xeon Platinum 6776P processors per node.

576
NVIDIA Blackwell Ultra GPUs per reference Scalable Unit
800 Gb/s
NVIDIA Quantum-X800 InfiniBand fabric connectivity
2–4 TB
Standard to maximum system memory per DGX B300 compute node

An AI Factory Built for Enterprise Scale

Built on NVIDIA Blackwell Ultra, DGX SuperPOD links 576 GPUs and 144 Intel Xeon CPUs across 72 nodes, scaling reasoning, inference and training efficiently.

576 GPUs Linked in One Scalable Unit

Seventy-two DGX B300 nodes share ultra-high-speed Quantum-X800 InfiniBand bandwidth inside standard AC-powered EIA compute racks.

Massive Memory and Bandwidth for Giant Models

Up to 4 TB of system memory per node and 8× ConnectX-8 adapters keep data close to compute for larger models.

Built for the Era of AI Reasoning

Eight Blackwell Ultra GPUs per system and high-throughput InfiniBand fabrics accelerate reasoning, long-context inference and training.

576 GPUs. One InfiniBand Fabric.
One SuperPOD.

NVIDIA Quantum-X800 InfiniBand connects all 576 GPUs and 144 Intel Xeon CPUs across 72 DGX B300 nodes into one unified compute cluster.

NVIDIA DGX SuperPOD with DGX B300 Systems rack architecture
576×
NVIDIA Blackwell Ultra GPUs
144×
Intel Xeon Platinum 6776P processors
Up to 288 TB
System memory across scalable unit
800 Gb/s
InfiniBand fabric with ConnectX-8

Workloads

From frontier training to real-time reasoning, DGX SuperPOD with DGX B300 Systems runs demanding AI workloads at data-center scale.

AI Reasoning and Real-Time Inference

576 Blackwell Ultra GPUs across a 72-system scalable unit deliver ultra-low-latency prefill and sustained decode throughput to keep reasoning and inference responses fast.

Large Language Model Training at Scale

An 800 Gb/s Quantum-X800 InfiniBand fabric with ConnectX-8 adapters links 72 nodes to help teams train frontier models in less time.

Higher AI Factory Output at Scale

Up to 288 TB of high-speed system memory and BlueField-3 DPUs keep large AI training pipelines and complex datasets moving efficiently across the cluster.

Engineered to dominate

Cluster-Scale AI Power

DGX SuperPOD with DGX B300 Systems combines Blackwell Ultra compute, dual Intel Xeon processors, and Quantum-X800 InfiniBand networking to deliver peak performance for AI factories.

NVIDIA DGX SuperPOD with DGX B300 Systems racks
10,368 PFLOPS
FP4 Tensor Core Compute (Sparse) (144 PFLOPS per compute node)
5,184 PFLOPS
FP8 / FP6 Tensor Core Compute (Sparse) (72 PFLOPS per compute node)
7,776 PFLOPS
FP4 Tensor Core Compute (Dense) (108 PFLOPS per compute node)
800 Gb/s
Scale-Out Networking per GPU (8× ConnectX-8 interfaces per node)
NVIDIA DGX SuperPOD with DGX B300 Systems racks
10,368 PFLOPS
FP4 Tensor Core Compute (Sparse) (144 PFLOPS per compute node)
5,184 PFLOPS
FP8 / FP6 Tensor Core Compute (Sparse) (72 PFLOPS per compute node)
7,776 PFLOPS
FP4 Tensor Core Compute (Dense) (108 PFLOPS per compute node)
800 Gb/s
Scale-Out Networking per GPU (8× ConnectX-8 interfaces per node)

The rack, itemized

SUBSYSTEM

SPECIFICATION / CONFIGURATION DETAILS

Processor

2 × Intel Xeon 6776P per node

Operating System

NVIDIA DGX OS

Chipset

NA

Accelerators / GPU

8× Blackwell Ultra GPUs per node; reference SU: 576 GPUs

Memory

Per compute node: 2 TB standard, up to 4 TB system memory

Primary Storage

Per node: 8 × 3.84 TB E1.S NVMe

Secondary Storage

Per DGX B300: 2x 1.92 TB M.2 NVMe boot; 8x 3.84 TB E1.S NVMe cache

Security

DGX B300 BMC; cluster security configuration dependent

Chassis

EIA racks; reference layout: 4 DGX B300 nodes per compute rack

Power Supply

12 AC PSUs per DGX B300 (200-240 VAC, N+N); three 2U rack PDUs per rack; about 56 kW per rack

Ports

Per DGX B300 node: 8 × OSFP, 2 × dual-port QSFP112, 1GbE management; Quantum-X800 InfiniBand switching

Slots / PCIe Architecture

Per node: 2 NVLink switches; network fabric determined by reference architecture

Bezel

DGX B300 front bezel

Form Factor

Reference SU: 72 DGX B300 nodes

Network Card

Per compute node: 8× ConnectX-8; 2× dual-port BlueField-3

Warranty

DGX nodes: 3-year business-standard hardware and software support; complete solution coverage per contract.

NVIDIA DGX SuperPOD with DGX B300 Systems power and networking

Memory, Fabric and
Networking at Scale.

Up to 288 TB

of fast memory per scalable unit

800 Gb/s compute ports

via ConnectX-8 SuperNICs

Quantum-X800 InfiniBand

fabric across 576 GPUs

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.