G6400L, Foredge 6U GPU Server with 8 x NVIDIA RTX 5090 GPUs

Foredge G6400L is a 6U rack-mounted GPU server with two Intel Xeon Gold 6430 processors, eight GeForce RTX 5090 graphics cards, and a CPU-GPU direct-connect design. It provides a high-density platform for heterogeneous computing with PCIe 5.0 x16 GPU expansion, DDR5 memory expansion, SATA/SAS/NVMe storage support, and redundant power and cooling.

Best Fit Applications

High-Density GPU Compute

Designed for high-density GPU compute deployments that require eight dual-width GPUs in a 6U rack-mounted platform.

Heterogeneous Computing Workloads

Built for heterogeneous computing workloads that benefit from direct CPU-GPU connectivity and PCIe 5.0 x16 GPU expansion.

GPU Deployments with Local Storage Expansion

Suitable for GPU deployments requiring local SATA, SAS, or NVMe storage expansion across up to 12 drive bays.

Why G6400L

Eight RTX 5090 GPUs in a 6U Platform

The chassis is specified for eight dual-width GeForce RTX 5090 graphics cards, with each GPU position supporting PCIe 5.0 x16.

CPU-GPU Direct Connect Design

The official product description identifies a CPU-GPU direct-connect architecture designed to improve data-transfer efficiency compared with PCIe switch architectures.

Serviceable Redundant Hardware

Hot-swappable fans and four hot-swappable power supplies support N+1 fan redundancy and 3+1 power-module redundancy.

Large Expansion Envelope

The platform supports up to 32 DDR5 memory slots, up to 12 SATA/SAS/NVMe drives, and optional OCP 3.0 or PCIe network cards.

Compare Related Models

	G6400L	SYS-221GE-TNHT-LCC	G530 V5	SR675 V3
Form Factor	6U rack-mounted server	GPU server	GPU server	ThinkSystem server
Processor / Accelerator	2 × Intel Xeon Gold 6430 processors; base frequency 2.10GHz; 8 × GeForce RTX 5090 turbo graphics cards	Dual 4th Gen Intel Xeon; 8x PCIe 5.0 GPU	2x Intel Xeon Scalable; 8x NVIDIA GPU	2x AMD EPYC; up to 8x GPU
Memory / Storage Summary	Official product page: 8 × Samsung 64GB DDR5 ECC RDIMM 2Rx4; supports up to 32 memory slots; 1 × Samsung 960G SATA enterprise-grade solid-state drive; 1 × Samsung 3.84T NVMe U.2 enterprise-grade solid-state drive	32 DDR5 DIMMs	PCIe 4.0	Up to 2TB RAM
Best Fit Scenario	GPU deployments using the Xeon Gold 6430 processor configuration and eight RTX 5090 GPUs	GPU server deployment with PCIe 5.0 expansion	Multi-GPU AI and accelerated computing	High-memory multi-GPU server workloads
Not Suitable For	Deployments without 6U rack space or without power and cooling capacity for a four-PSU GPU server	General-purpose storage-only deployments	Deployments without GPU acceleration requirements	Deployments requiring a confirmed Intel Xeon platform