SupermicroGPU & AI Servers

Supermicro 4U Liquid-Cooled HGX B300 System

Model SYS-422GS-NB3RT-LCC

DLC-2 liquid-cooled 8-GPU node capturing up to 98% of heat for dense, efficient AI factories.

  • GPUsNVIDIA HGX B300 8-GPU
  • DensityUp to 64 GPUs per rack
  • CoolingDLC-2 direct liquid cooling, ~98% heat capture
  • NetworkingIntegrated ConnectX-8, 800 Gb/s
Illustration of an 8-GPU AI server
Overview

About the 4U Liquid-Cooled HGX B300 System

This 4U Supermicro system packs a liquid-cooled NVIDIA HGX B300 8-GPU board into half the height of the air-cooled version. Direct-to-chip cold plates remove most of the heat, so racks can hold far more GPUs.

It is part of Supermicro's DLC-2 stack, which runs on 45 °C warm water to cut chiller use, and ships as rack-level (L11) or cluster-level (L12) validated solutions.

HGX B300 8-GPU

2.3 TB HBM3e per system with 1.8 TB/s NVLink.

DLC-2 liquid cooling

Direct-to-chip cold plates; up to 40% data center power savings.

45 °C warm water

Reduces water use and can eliminate chillers and compressors.

8× ConnectX-8 SuperNICs

Up to 800 Gb/s per GPU, plus 4 PCIe Gen 5 x16 slots.

Dense racks

4U per 8 GPUs for high GPU counts per rack.

L11/L12 validated

Racks and clusters are tested before shipping.

Specifications

Technical Specifications

Processor & memory

CPUsDual Intel Xeon 6700 series with P-cores, up to 350 W
Memory32 DIMMs; up to 8 TB at 5200 MT/s or 4 TB at 6400 MT/s DDR5 RDIMM

GPU

GPUsNVIDIA HGX B300 8-GPU
GPU memory2.3 TB HBM3e per system
InterconnectFifth-generation NVLink, 1.8 TB/s

Storage & networking

Drive bays8× hot-swap E1.S NVMe
Boot2× M.2 NVMe
Compute network8× NVIDIA ConnectX-8 SuperNICs, up to 800 Gb/s
Expansion4× PCIe Gen 5 x16

Chassis, power & cooling

Form factor4U rackmount
Dimensions (W × H × D)448 × 174 × 896.8 mm
Weight84.6 kg net
Power4× 6.6 kW redundant (2+2) Titanium
CoolingDirect-to-chip (D2C) cold plates, Supermicro DLC-2
Source: official Supermicro product information. Specifications are set by the manufacturer and may change. Plan Man confirms the final configuration, availability and lead time in your quote.
Use cases

What it's built for

  • Dense AI training clusters
  • High-throughput inference
  • Energy-efficient AI factories
  • HPC
Plan Man services

Supplied and supported in the UAE

  • Requirement analysis, sizing and bill of materials
  • Genuine supply with manufacturer warranty
  • Rack integration, cabling and onsite deployment
  • Power and liquid cooling planning for dense GPU racks
  • After-sales support, RMA handling and upgrades from Dubai

Get a Quote

Tell us your workload and timeline for the Supermicro 4U Liquid-Cooled HGX B300 System. We reply within one business day with a configuration, pricing and lead time.

Request a Quote