1 1

G7868 X6 AI GPU Server for Large-Scale Training and HPC

G7868 X6 AI GPU Server for Large-Scale Training and HPC

$999.00
$999.00
Sale Sold out
Shipping calculated at checkout.

Overview

Multi-Accelerator Module Compatibility

Compatible with HGX 8-GPU modules

Supports OAM 1.5–2.0 accelerator modules

Designed for high-density AI computing environments

Flexible Topology Architecture

Flexible GPU topology switching

Optimized resource allocation for different compute workloads

PCIe module configuration adaptable to specific business requirements

Modular System Architecture

Modular design for compute, control, cooling, and PCIe units

Independent module maintenance for easier servicing

Supports liquid cooling modules for high-performance deployments

High-Performance System Architecture

Dual CPU support with up to 64 cores per processor

DDR5 memory up to 5600MT/s

PCIe 5.0 expansion architecture for high-bandwidth devices

Scalable AI Infrastructure

Up to 29 NVMe/SATA storage drives

High-density GPU deployment for AI training clusters

Optimized for large-scale data center environments

View full details
  • Overview
  • Description
  • Specification
  • Datasheet

The G7868 X6 Server is a high-performance AI computing platform designed for large-scale training workloads, high-performance computing (HPC), and modern data center environments.

The system is compatible with HGX 8-GPU modules or standard OAM 1.5–2.0 accelerator modules, enabling powerful AI computing capabilities and supporting diverse accelerator architectures. Its modular system design significantly improves thermal performance and serviceability, ensuring the full performance potential of high-end GPUs can be effectively utilized.

The control unit supports up to two CPUs with up to 64 cores each, combined with 8-channel DDR5 memory running at speeds up to 5600MT/s, delivering up to 50% higher memory bandwidth for data-intensive workloads. With PCIe 5.0 expansion architecture and support for up to 29 storage drives, the platform offers excellent scalability for advanced AI and HPC infrastructure deployments.

Multi-Accelerator Module Compatibility

Compatible with HGX 8-GPU modules

Supports OAM 1.5–2.0 accelerator modules

Designed for high-density AI computing environments

Flexible Topology Architecture

Flexible GPU topology switching

Optimized resource allocation for different compute workloads

PCIe module configuration adaptable to specific business requirements

Modular System Architecture

Modular design for compute, control, cooling, and PCIe units

Independent module maintenance for easier servicing

Supports liquid cooling modules for high-performance deployments

High-Performance System Architecture

Dual CPU support with up to 64 cores per processor

DDR5 memory up to 5600MT/s

PCIe 5.0 expansion architecture for high-bandwidth devices

Scalable AI Infrastructure

Up to 29 NVMe/SATA storage drives

High-density GPU deployment for AI training clusters

Optimized for large-scale data center environments

Category Specification
Form Factor 8U Rack Server
Dimensions (H × W × D) 352 mm × 447 mm × 882 mm
Processor Up to 2 CPUs
CPU Cores Up to 64 cores per CPU
Max CPU Power 350W
GPU Architecture Supports HGX 8-GPU modules or OAM 1.5–2.0 modules
Memory Slots Up to 32 DDR5 DIMM slots
Memory Speed Up to 5600 MT/s
Max Memory Capacity Up to 8 TB
Memory Technology Supports CXL 1.1 and HBM technology
Storage Controller Integrated 6Gb/s SATA controller; optional 12Gb/s SAS HBA or RAID controller
Storage Bays Up to 29 × 2.5" drives
NVMe Support Up to 12 NVMe drives
Boot Storage 2 × 2280 / 22110 SATA or 1 × NVMe M.2 SSD
PCIe Expansion Up to 15 PCIe 5.0 expansion slots
PCIe Slots Up to 14 × PCIe 5.0 x16 slots
OCP Slot 1 × OCP 3.0 onboard slot
Management AST2600 BMC with IPMI 2.0, Redfish, SOL, KVM, Virtual Media
Networking OCP 3.0 NIC support, optional 1/10/25/50/100/200/400GE adapters
Security Chassis intrusion detection, PFR 3.0, SGX 2.0, TPM/TCM
Cooling 15 high-performance hot-swappable fans, N+1 redundancy
Power Supply 6 × 3500W 54V platinum PSUs (future support for 4200W / 5200W); optional additional 12V platinum PSUs
Power Redundancy N+N / N+1 redundancy
Operating Temperature 0°C – 35°C

WHY CHOOSE US

Why Choose Portower Products?

Global Certification Icon

Strong Comprehensive Strength

End-to-end capabilities covering R&D, manufacturing, quality control, and global supply, ensuring stable delivery and long-term partnership reliability.
Advanced R&D and Manufacturing Icon

Advanced R&D and Manufacturing

In-house engineering teams and modern production facilities enable fast iteration, strict testing, and consistent performance at scale.
Innovation and Intellectual Property Icon

Innovation and Intellectual Property

Continuous investment in product innovation with proprietary designs and core technologies to stay ahead of evolving network demands.
International Quality Recognition Icon

International Quality Recognition

Products certified to major global standards, widely deployed across international markets and trusted by professional customers worldwide.
FAQ Image

"Do you offer international shipping?"

4.8 ★
12K+

Have A Question? We've got The Answers

Yes. Our products are designed for stable, continuous deployment and are widely used in commercial and infrastructure projects.
Absolutely. We support volume pricing, stable production scheduling, and long-term supply agreements.
Yes. We offer flexible OEM/ODM services to support branding, packaging, and customized project requirements.
All products go through strict quality control and standardized testing to ensure consistent performance across batches.
We provide professional pre-sales consultation and responsive after-sales technical support for integration and deployment.

FACTORY SHOW

Our Factory Show

Portower operates a modern factory with advanced R&D labs...

ABOUT US