Brand: HPE | Category: Servers
SKU: HPE-P65820B21 | Part #: P65820-B21 | MPN: P65820-B21
Contact for Pricing — Request a Quote
The HPE ProLiant Compute DL384 Gen12 with NVIDIA GH200 from HPE is enterprise-grade Servers hardware built for data centers. HPE ProLiant DL384 Gen12 pairs NVIDIA GH200 Grace Hopper Superchips with a GPU-dense 4U architecture engineered for large-scale AI inference and accelerated HPC training workloads. Available in brand new condition. Available from Omnixon Global — contact our team to request a quote.
| Brand | HPE |
| Category | Servers |
| SKU | HPE-P65820B21 |
| Part Number | P65820-B21 |
| Condition | New |
| Product Line | HPE ProLiant Compute DL384 Gen12 |
| Form Factor | 4U Rack Server |
| Accelerator Architecture | NVIDIA GH200 Grace Hopper Superchip |
| GPU Architecture | NVIDIA Hopper (H100 GPU die) |
| CPU Architecture per Superchip | NVIDIA Grace, 72-core Arm Neoverse V2 |
| CPU-GPU Interconnect | NVLink-C2C, 900 GB/s bidirectional bandwidth per Superchip |
| GPU Memory per Superchip | 96 GB HBM3 |
| CPU Memory per Superchip | Up to 480 GB LPDDR5X |
| GPU Memory Bandwidth | Up to 4.0 TB/s HBM3 per GH200 Superchip |
| FP8 Tensor Core Performance | Up to 2,000 TFLOPS per GH200 Superchip |
| FP16 / BF16 Tensor Core Performance | Up to 1,000 TFLOPS per GH200 Superchip |
| FP64 Performance | Up to 67 TFLOPS per GH200 Superchip |
| Inter-GPU Connectivity | NVLink 4.0 / NVLink Switch support for multi-node scale-out |
| PCIe Generation | PCIe 5.0 |
| Network I/O | Up to 400 GbE or NDR InfiniBand via OCP 3.0 and PCIe 5.0 slots; NVIDIA ConnectX-7 or BlueField-3 DPU options |
| Storage Options | NVMe Gen5 U.2/E3.S front-accessible drive bays; support for HPE NS204i-u Boot Device |
| Cooling Options | High-performance fans (air-cooled); Direct Liquid Cooling (DLC) rear-door heat exchanger and warm-water direct-to-chip loop support |
| Power Supply | Redundant high-efficiency Titanium-rated PSUs; per-chassis wattage scales with GH200 module count and TDP |
| Systems Management | HPE iLO 7 BMC with NVIDIA GPU Telemetry integration; HPE GreenLake Compute Ops Management |
| Operating System Support | Ubuntu 22.04/24.04 LTS, Red Hat Enterprise Linux 9.x, SUSE Linux Enterprise Server 15 SP5+ |
| Software Ecosystem | NVIDIA AI Enterprise, CUDA 12.x, cuDNN, TensorRT-LLM, NCCL, OpenMPI, HPE Performance Cluster Manager |
| Chassis Dimensions | 4U; standard 19-inch EIA rack compatible |
| Launch Generation | Gen12 (2025-Q3) |
The HPE ProLiant Compute DL384 Gen12 with NVIDIA GH200 is built for enterprise data-center workloads — virtualization (VMware, Proxmox, Nutanix), private cloud, database hosting, and AI/ML training. It fits standard EIA-310 server racks and supports redundant PSUs and hot-swap drives common in production environments.
Key specifications for the HPE ProLiant Compute DL384 Gen12 with NVIDIA GH200: new condition; manufacturer HPE; product line HPE ProLiant Compute DL384 Gen12; form factor 4U Rack Server; accelerator architecture NVIDIA GH200 Grace Hopper Superchip; gpu architecture NVIDIA Hopper (H100 GPU die); cpu architecture per superchip NVIDIA Grace, 72-core Arm Neoverse V2. Manufacturer part number P65820-B21. For the full datasheet with electrical, environmental, and compliance details, contact our pre-sales engineering team.