Supermicro AS -A126GS-TNBR 10U Rackmount H14 DP GPU Server Pre-Configured Edition
- Key Features/Applications: Scientific Research, AI/Deep Learning Training, Industrial Automation, Healthcare, Data Center, Conversational AI, Business Intelligence & Analytics, Drug Discovery, Climate and Weather Modeling, Finance & Economics, High Performance Computing (HPC)
- CPU: Dual AMD EPYC™ Turin 9575F Processors, up to 128 Cores
- Chassis: 10U
- Drive: 2x 960GB M.2 PCIe 4.0 NVMe + 4x 3.8TB U.2 PCIe 4.0 NVMe SSD
- RAM: 24x 96GB DDR5-6400 RDIMM Memory
- GPU: 1x Onboard NVIDIA SXM HGX B200 (8-GPU) Board
- Network Ports: 2x 25GbE SFP28 + 8x 400G Ethernet/NDR IB Ports
- Power Supply: 6x 5250W Redundant (3 + 3) Titanium Level Power Supplies
- Highly Recommended for Data Center Environments Only
* Please note that 3-Year (or above) Onsite Service is required, and pricing may vary based on your service location. Please contact us before purchasing this server.
* Orders are subject to review and require verification prior to shipping. We reserve the right to cancel any order at our discretion.
* Estimated lead time is currently 14-16 weeks post-verification. Actual delivery schedule will depend on quantity and availability.

10U Multi-GPU Server with NVIDIA HGX B200
The Supermicro AS -A126GS-TNBR is an air-cooled 10U system featuring 8 NVIDIA HGX B200 GPUs and dual AMD EPYC™ Turin 9575F CPUs for AI training and HPC.
Extreme Multi-GPU Performance for Large-Scale AI Training
Engineered for deep learning training, conversational AI, and drug discovery, the Supermicro AS -A126GS-TNBR packs massive compute power into a 10U chassis. Driven by dual AMD EPYC™ Turin 9575F processors and an air-cooled 8-GPU NVIDIA SXM HGX B200 board, it accelerates complex multi-node parallel computing for enterprise data centers.
Massive Memory Bandwidth
To eliminate bottlenecks in high-performance computing (HPC) and climate modeling, this pre-configured system supports total 2.25TB of high-speed DDR5-6400 memory.
Data Center Class Infrastructure and High-Draw Redundant Power
Purpose-built strictly for high-density data center environments, this high-draw 10U platform requires dedicated rack PDU infrastructure. It includes 2 960GB M.2 NVMe and 4 3.8TB U.2 NVMe SSDs. Power delivery is backed by six 5250W Titanium Level redundant power supplies (3+3), ensuring absolute system stability under full compute loads.
Key Features
- Pre-configured, ready-to-ship system. Estimated lead time is 2-3 weeks post-verification.
- Dual AMD EPYC™ Turin 9575F processors providing high-density CPU compute power
- 8 onboard NVIDIA SXM HGX B200 GPUs (air-cooled) for massive AI training acceleration
- Pre-installed 2.25TB DDR5-6400 memory for memory-intensive workloads
- Pre-installed 2 960GB M.2 NVMe and 4 3.8TB U.2 NVMe SSDs
- Dual onboard 25GbE SFP28 and 8x 400G Ethernet/NDR IB Ports
- Data center deployment only 10U chassis optimized for heavy rack-scale AI integration
- 6x 5250W Titanium level (3+3) requiring high-draw rack PDU infrastructure
| System Family | GPU |
|---|---|
| CPU | 5th Gen AMD EPYC™ 9005 Series |
| Specification | |
| Search Terms | AS-A126GS-TNBR, A126GS, A126GS-TNBR, 10U GPU, 10U Rackmount, AMD 10U, AMD EPYC 9005, AMD EPYC 9004, amd cpu, dual amd, dual processor, B200, B200 system, NVIDIA HGX B200, onboard GPU, NVIDIA Certified, OCP accepted, Scientific Research, Deep Learning, Industrial Automation, Healthcare, Conversational AI, Business Intelligence, Drug Discovery, Climate, Weather Modeling, Finance, DDR5 6400, DDR5 4800 |
| We're sorry, an error has occurred while generating this content. |

