NVIDIA · GPU Servers
NVIDIA HGX H100 8-GPU Baseboard
8× H100 80GB SXM5 · NVLink 900 GB/s · 3.2TB/s Memory BW · HGX Baseboard
$290,000 $320,000
The NVIDIA DGX H100 is the reference architecture for enterprise AI at scale. Built around 8× NVIDIA H100 80GB SXM5 GPUs interconnected via NVLink at 900 GB/s, delivering 32 petaFLOPS of FP8 AI performance. Dual Intel Xeon Platinum 8480C (56C each), 2TB DDR5-4800 ECC, and 30TB NVMe storage. Includes NVIDIA AI Enterprise software stack, DGX OS, and 3-year enterprise support. Ideal for training GPT-scale LLMs, computer vision, and HPC.
| Processor | Intel Xeon Platinum 8480C (2.0 GHz, 105MB) |
| Generation | 4th Gen Sapphire Rapids |
| Cores / Threads | 2× 56 cores / 112 threads (112C / 224T total) |
| RAM Installed | 2TB DDR5-4800 ECC |
| RAM Type | DDR5 ECC Registered |
| RAM Slots | 32 DIMM slots |
| Max RAM | 2TB |
| Storage | 30TB NVMe (2× 1.92TB OS + 8× 3.84TB data) |
| Storage Type | PCIe 5.0 NVMe |
| Max Storage | 30TB |
| Graphics Card | 8× NVIDIA H100 80GB SXM5 (NVLink 900 GB/s) |
| Operating System | DGX OS (Ubuntu-based) |
| Weight | 130 kg |
| Ports | 8× OSFP 400Gb/s InfiniBand · 2× 10GbE · BMC |
| Warranty | 3 Years International |
No reviews yet. Be the first to review this product!
NVIDIA · GPU Servers
8× H100 80GB SXM5 · NVLink 900 GB/s · 3.2TB/s Memory BW · HGX Baseboard
$290,000 $320,000
NVIDIA · GPU Servers
8× H200 141GB HBM3e · NVLink 900 GB/s · 4.8TB/s Memory BW · HGX Baseboard
$420,000 $470,000
Dell · GPU Servers
4U · 2× Xeon Platinum 8480+ · 1TB DDR5 · 4× H100 80GB SXM5 · 4× 3.84TB NVMe
$215,000 $245,000
Dell · GPU Servers
6U · 2× Xeon Platinum 8480+ · 2TB DDR5 · 8× H100 80GB SXM5 · 8× 3.84TB NVMe
$335,000 $375,000