GPU Server

HIGH PERFORMANCE COMPUTING

AS-A126GS-TNBR

Highlights

CPU:
2x AMD EPYC 9005/9004 Series
RAM:
Up to 6TB
STORAGE:
PCIe4 M.2 NVMe & PCIe5 NVMe 2.5"
GPU SUPPORT:
8x NVIDIA B200

AS-A126GS-TNBR Specs

FORM FACTOR 10U Rackmount
PROCESSORS Up To 384 Cores 2x AMD EPYC 9005/9004 Series Processors
MEMORY Up to 6TB DDR5
STORAGE Up to 2x M.2, 8x NVMe 2.5”, 2x SATA 2.5” OR Up to 8x E1.S NVMe
NETWORK Dual Port 10GbE RJ45
PCIE EXPANSION 2x PCIe 5.0 x16 FHHL Slots, 8x PCIe 5.0 x16 LP Slots
WARRANTY 5 Year Aspen Systems Warranty
DIMENSIONS 17.6 Width x 17.2 Height 33.2 Length/Depth
MANUFACTURER SuperMicro

B200 Specs

The HGX B200 includes 8x B200 GPUs in a single system. Below are performance specifications. Real word performance may vary based on workload.

Total HGX B200 Specifications
Model
Architecture
Total FP4 Tensor Core
Total FP8/FP6 Tensor Core
Total Memory Bandwidth
HGX B200
Blackwell
144 PFLOPS | 72 PFLOPS
72 PFLOPS
62 TB/s
Individual B200 GPU Specifications:
FP64/FP64 Tensor Core
FP32
GPU Memory | Bandwidth
Max TDP
Interconnect
37 TFLOPS
75 TFLOPS
180GB | 7.7 TB/s
Configurable up to 1,100 W
5th Gen NVLINK: 1.8TB/s

Turnkey GPU Server With Your Personalized Software, Ready To Go

Build, optimize, and deploy accelerated workloads with the software, frameworks, libraries, and development tools your GPU infrastructure depends on.

40+ Supported Software Technologies
GPU Optimized Computing Stack

AI SOFTWARE STACK

Aspen Systems delivers turnkey AI solutions with expert installation, optimization, and support for PyTorch, TensorFlow, Keras, NVIDIA CUDA, and NVIDIA AI Enterprise, providing a production-ready platform for AI training and inference. Choose from Red Hat Enterprise Linux, Rocky Linux, Ubuntu, SUSE Linux, and other operating systems to create a solution tailored to your environment.

NVIDIA B200 Servers Built for
AI at Scale

High-performance NVIDIA B200 server platforms engineered for large-scale AI training, inference, reasoning, and accelerated computing.

B200 GPU Server

B200 Rack-Scale Systems

Dense server platforms built around NVIDIA B200 GPUs for high compute density, memory bandwidth, and scalability.

Get a Quote for a Custom Enterprise AI

Tell us about your project and our team will get in touch to help you find the right solution.

    Are you human?

    captcha

    *Aspen Systems Inc. will never share or sell your personal information for any reason. Read our policy on data privacy for more information.

    Frequently
    Asked
    Questions

    What is an NVIDIA B200 GPU server?
    An NVIDIA B200 GPU server is a high-performance, GPU-accelerated system designed for demanding AI workloads. Powered by NVIDIA Blackwell GPUs, it delivers the compute power and memory capacity needed for large-scale AI training, inference, and high-performance computing applications.
    What workloads are NVIDIA B200 GPU servers designed for?
    NVIDIA B200 GPU servers are designed for large language models (LLMs), generative AI, deep learning, model training and inference, scientific simulations, and other compute-intensive HPC workloads.
    How many NVIDIA B200 GPUs can a server support?
    Supported GPU configurations depend on the server platform. NVIDIA HGX B200 servers can house eight NVIDIA B200 GPUs connected through NVIDIA NVLink to support demanding multi-GPU AI workloads.
    What are the benefits of NVIDIA Blackwell architecture?
    NVIDIA Blackwell architecture is engineered to accelerate advanced AI workloads with powerful Tensor Core processing, high-bandwidth GPU memory, and optimized performance for large-scale model training and inference.
    Can NVIDIA B200 GPU servers be customized?
    Yes. Aspen Systems can help configure GPU servers around your workload requirements, including processor selection, system memory, storage, networking, and GPU configuration. Available options depend on the selected server platform.
    How can I request pricing for an NVIDIA B200 GPU server?
    You can request a customized quote from Aspen Systems. Our team can help evaluate your AI or HPC workload requirements and recommend an appropriate B200 GPU server configuration. Request a Quote
    Trusted partnership by known organizations