Product Overview
The Supermicro SYS-421GE-TNHR2-LCC-G1 Gold Series is a high-performance, liquid-cooled AI server designed for enterprise AI infrastructure, large language model (LLM) training, AI inference, deep learning, and high-performance computing (HPC).
Powered by eight NVIDIA H200 Tensor Core GPUs, dual Intel Xeon Platinum 8570 processors, and 3TB of DDR5 ECC system memory, this platform delivers substantial GPU computing capability for demanding AI workloads.
Featuring NVIDIA NVLink and NVSwitch, high-speed 400G networking, and direct-to-chip liquid cooling, the system is designed for intensive computing environments and large-scale AI clusters.
Technical Specifications
Component
Specification
Manufacturer
Supermicro
Series
Gold Series G1
Model
SYS-421GE-TNHR2-LCC-G1
Chassis
4U CSE-GP403TS-R000NFP
Motherboard
Supermicro X13DEG-M
Processor
2× Intel Xeon Platinum 8570
CPU Cores
112 total cores (56 per processor)
CPU Frequency
2.1 GHz
CPU Cache
300MB per processor
Installed System RAM
3TB DDR5-5600 ECC RDIMM
Memory Capacity
Up to 8TB, configuration-dependent
GPU Platform
NVIDIA HGX H200
GPU Configuration
8× NVIDIA H200 Tensor Core GPUs
Total GPU Memory
1,128GB HBM3e (141GB per GPU)
GPU Memory Bandwidth
Up to 4.8TB/s per GPU
GPU Interconnect
NVIDIA NVLink + NVSwitch, up to 900GB/s
Boot Storage
2× 960GB M.2 NVMe SSD
Drive Bays
8× front hot-swap 2.5-inch U.2 NVMe
High-Speed Networking
8× 400G NDR InfiniBand/Ethernet OSFP
Additional Networking
1× dual-port 200G NDR200/Ethernet QSFP112
Onboard LAN
2× 10GbE Intel X710-AT2
PCIe Expansion
8× PCIe 5.0 x16 low-profile + 2× FHHL
Cooling
Direct-to-chip liquid cooling for CPUs and GPUs
Power Supply
4× 5250W 80 PLUS Titanium, redundant (2+2)
Operating Temperature
10–35°C
Dimensions
174 × 449 × 895.35 mm
Net Weight
80 kg
Form Factor
4U Rackmount
Key Features
- 8× NVIDIA H200 GPUs: High-performance acceleration with 1.128TB aggregate HBM3e GPU memory.
- 112 CPU Cores: Dual Intel Xeon Platinum 8570 processors.
- 3TB Installed DDR5 Memory: Enterprise-grade ECC RAM for memory-intensive workloads.
- High-Speed GPU Networking: 400G NDR connectivity for distributed AI training and HPC clusters.
- Advanced Liquid Cooling: Direct-to-chip cooling designed for high-density computing environments.
- Factory-Integrated Gold Series: Fixed configuration with manufacturer integration and burn-in testing.
Applications
- Large Language Model (LLM) Training and Fine-Tuning
- Generative AI and Enterprise AI Infrastructure
- High-Performance Computing (HPC)
- Deep Learning and AI Inference
- AI Data Centers and GPU Computing Clusters
- Scientific Research and Advanced Simulation
Gold Series Configuration
The Supermicro Gold Series G1 is offered as a fixed-configuration, factory-integrated system.
The configuration includes eight NVIDIA H200 GPUs, dual Intel Xeon Platinum 8570 processors, 3TB DDR5 ECC RAM, high-speed networking and integrated liquid-cooling hardware.
Important: The server requires a compatible facility coolant distribution unit (CDU), liquid-cooling connections, and sufficient electrical infrastructure. Factory integration does not eliminate these deployment requirements.
Availability & Ordering
Billion Bison Technologies | Enterprise AI & HPC Solutions
- Price applies to the specified fixed Gold Series G1 configuration.
- Pricing and availability are confirmed upon quotation.
- International shipping options are available, subject to destination and freight requirements.
- Warranty, delivery terms and included accessories are confirmed before ordering.
- End-user verification and export compliance documentation are required. Export licences may be necessary depending on the destination, end user and intended use.
- Prices are subject to change without prior notice.
Request a Quotation
info@billionbison.com
Global Supplier of NVIDIA GPU Servers, AI Infrastructure & High-Performance Computing Solutions.