SYS-A22GA-NBRT 8-GPU with HGX B200

DP Intel 10U System with NVIDIA HGX B200 8-GPU
  1. Dual Intel Xeon 6960P: dual-socket, up to 72 cores, 432MB cache, up to 3.1 GHz.
  2. 24x 128GB DDR5-6400 2Rx4 (32Gb) ECC RDIMM
  3. NVIDIA HGX B200 (Umbriel): 8-GPU baseboard with 180GB HBM3e
  4. 4x Micron 7450 PRO 3.8TB NVMe PCIe 4.0 3D TLC U.3 7mm,1DWPD
  5. Mellanox ConnectX-4 LX EN: dual-port 25G SFP28
  6. 2x NVIDIA BlueField-3 B3140H: 400GbE DPU in HHHL 
  7. ConnectX-6 Dx: dual-port 100GbE QSFP56 adapter

 

🧩 Product Overview: Supermicro SYS-A22GA-NBRT

 

The SYS-A22GA-NBRT is a next-generation 10U GPU server built to deliver extreme compute density, massive memory bandwidth, and advanced AI acceleration. It is ideal for enterprises and research institutions working on large language models (LLMs), deep learning, scientific simulations, and autonomous systems.

 

🔑 Key Features

 

The SYS-A22GA-NBRT is engineered for extreme performance and scalability:

  • Dual Intel® Xeon® 6900 series processors with up to 128 cores and 504MB cache
  • 8x NVIDIA SXM B200 GPUs with 180GB HBM3e total memory and NVLink/NVSwitch interconnect
  • 24x DIMM slots supporting DDR5 ECC RDIMM/MRDIMM up to 8800MT/s and 6TB capacity
  • 10x PCIe 5.0 x16 LP + 2x PCIe 5.0 x16 FHHL slots for expansion
  • 10x front hot-swap 2.5" NVMe U.2 + 2x front hot-swap NVMe M.2
  • 2x 10GbE RJ-45 LAN ports + 1x dedicated BMC/IPMI port
  • Advanced security: TPM 2.0, Silicon Root of Trust (NIST 800-193), Secure Boot, Firmware Recovery
  • Comprehensive management tools: SuperCloud Composer®, Supermicro Server Manager (SSM), Super Diagnostics Offline (SDO), and more

 

🚀 Key Applications

 

This system is purpose-built for the most demanding AI and compute workloads, including:

  • AI/Deep Learning Training and Inference
  • Large Language Models (LLMs) and Generative AI
  • High Performance Computing (HPC)
  • Scientific Research
  • Conversational AI
  • Business Intelligence & Analytics
  • Drug Discovery
  • Financial Services and Fraud Detection
  • Autonomous Vehicle Technologies

 

🖥️ Chassis

 

  • 10U X14 HGX 8GPU (Rear I/O), X14DBG-DAP, CSE-GP1001TS
    A high-density 10U chassis designed for 8 GPUs with rear I/O layout. Optimized for airflow, power, and thermal management in AI and HPC environments.

 


 

🧠 Processor

 

  • Dual Intel Xeon 6960P: dual-socket, up to 72 cores, 432MB cache, up to 3.1 GHz
    Dual high-performance processors with up to 72 cores each, ideal for compute-intensive AI and data workloads.

 


 

🧬 Memory

 

  • 128GB DDR5-6400 2Rx4 (32Gb) ECC RDIMM
    Ultra-fast ECC memory modules designed for high-throughput AI training and inference tasks.

 


 

💽 Storage

 

  • Micron 7450 PRO 3.8TB NVMe PCIe 4.0 3D TLC U.3 7mm, 1DWPD
    High-endurance enterprise SSDs for fast data access and reliable performance under heavy workloads.

 


 

🎮 GPU & Accelerator

 

  • NVIDIA Umbriel GPU Baseboard, HGX B200 8-GPU 180GB HBM3e
    Cutting-edge GPUs with massive memory bandwidth and compute power for next-gen AI models.

 


 

🔌 DPU (Data Processing Unit)

 

  • Nvidia BlueField-3 B3140H E-series HHHL DPU, 400GbE
    Offloads networking and security tasks, enabling efficient in-network computing and data movement.

 


 

🌐 Networking

 

  • ConnectX-6 Dx 100GbE Ethernet Adapter Card, dual port QSFP56
    High-speed interconnect for GPU-to-GPU and cluster communication.
  • Std LP 2-port 25G SFP28, Mellanox ConnectX-4 LX EN
    Mid-tier networking for storage and node communication.

 


 

💼 Software

 

  • NVIDIA ENT Business STD Support Services for ADPTR-BF-SW, 3Y
    Enterprise-grade support for NVIDIA software, ensuring updates and technical assistance.

 


 

🛡️ Support & Warranty

 

  • 3 YRS LABOR, 3 YRS PARTS, 1 YR CRS UNDER LIMITED WRNTY
    Comprehensive warranty coverage for peace of mind and hardware assurance.

 


 

🧩 Accessories

 

  • Air Shroud Mylar & Add-on Shroud Set
    Ensures optimal airflow and cooling efficiency for high-density GPU configurations.

 


 

NTS AI Software Stacks

Purpose-Built for High-Performance AI Infrastructure

ForgeKit is engineered for environments demanding maximum control, security, and compliance. With a minimal install—just the driver and CUDA—it’s perfect for air-gapped, regulated, or mission-critical deployments. ForgeKit is customizable by design, making it the go-to stack for enterprises prioritizing data sovereignty and infrastructure governance.

$400,761.20