| Tagline | Flagship 8-GPU AI Training Platform — Up to 8x NVIDIA H200 SXM in 8U Air-Cooled | High-Density 4U AI Training Server — 8× NVIDIA HGX H100/H200 with Direct-to-Chip Liquid Cooling | Flexible 4U AMD EPYC GPU Platform — Up to 8 PCIe GPUs for AI, HPC, and Mixed Workloads |
| Key Features | 8× NVIDIA HGX H100 (80GB) or H200 (141GB) SXM GPUs · NVIDIA NVLink + NVSwitch GPU-to-GPU interconnect · Dual 5th/4th Gen Intel Xeon Scalable (LGA-4677) — up to 64C/128T each · 32× DDR5 DIMM slots — up to 8TB at 5600MT/s ECC | 8× NVIDIA HGX H100 (80GB) or H200 (141GB) SXM GPUs · Direct-to-Chip (D2C) liquid cooling — higher sustained clocks, lower PUE · Dual 5th/4th Gen Intel Xeon Scalable (LGA-4677) — up to 64C/128T each · 32× DDR5 DIMM slots — up to 8TB at 5600MT/s (1DPC) or 4TB at 5600MT/s ECC | Up to 8× double-width PCIe GPU support — wide compatibility matrix · Compatible: NVIDIA H100, H100 NVL, L40, L40S, A100, RTX 6000 Ada, AMD MI210 · Dual AMD EPYC 9005/9004 — up to 160 cores / 320 threads total system · Up to 6TB DDR5 ECC across 24 DIMM slots |
