Brand: HPE | Category: Servers
SKU: HPE-R9G65A | Part #: R9G65A | MPN: R9G65A
Contact for Pricing — Request a Quote
The HPE Cray XD675 AI Accelerator Server from HPE is enterprise-grade Servers hardware built for data centers. High-density 4U OCP-form-factor AI accelerator server delivering up to 8x NVIDIA H200 SXM5 GPUs with NVLink fabric and direct liquid cooling for large-scale generative AI and HPC workloads. Available in brand new condition. Available from Omnixon Global — contact our team to request a quote.
| Brand | HPE |
| Category | Servers |
| SKU | HPE-R9G65A |
| Part Number | R9G65A |
| Condition | New |
| Product Line | HPE Cray XD Series |
| Model | HPE Cray XD675 |
| Form Factor | 4U, OCP Data Center Ready (DCR) form factor |
| GPU Configuration | Up to 8x NVIDIA H200 SXM5 Tensor Core GPUs |
| GPU Interconnect | NVLink 5th Generation, 900 GB/s bidirectional per GPU pair |
| GPU Memory (Total) | Up to 1.12 TB HBM3e (8x 141 GB per GPU) |
| GPU Memory Bandwidth (Aggregate) | Up to 3.35 TB/s aggregate across all 8 GPUs |
| CPU | Dual 4th Gen AMD EPYC processors (Genoa) |
| System Memory | Up to 6 TB DDR5, 12x DIMM slots per processor (24 slots total) |
| PCIe Generation | PCIe Gen 5 |
| Network Fabric | HPE Slingshot 11 (200 Gb/s per port), supports multi-rail configurations |
| Host Fabric Interface | Up to 4x HPE Slingshot NIC (200 Gb/s) for GPU-direct RDMA |
| Internal Storage | Up to 4x NVMe M.2 SSD (boot) + front-accessible NVMe U.2 drive bays |
| Cooling Technology | Direct Liquid Cooling (DLC) with integrated GPU and CPU cold plates; supplementary airflow fans |
| Power Supply | Redundant high-efficiency (Titanium) PSUs; supports high-density rack PDU integration |
| Management | HPE iLO 6 with Redfish API, IPMI 2.0, HPE GreenLake cloud integration |
| Operating System Support | Red Hat Enterprise Linux, SUSE Linux Enterprise Server, Ubuntu, CentOS Stream |
| AI Software Stack | Compatible with NVIDIA AI Enterprise, CUDA 12.x, cuDNN, NCCL, PyTorch, TensorFlow, JAX |
| Security | HPE Silicon Root of Trust, Secure Boot, TPM 2.0, encrypted storage support |
| Chassis Dimensions | 4U OCP DCR chassis; rack-optimized for high-density GPU pod configurations |
| Launch | 2025 Q1 |
The HPE Cray XD675 AI Accelerator Server 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 Cray XD675 AI Accelerator Server: new condition; manufacturer HPE; product line HPE Cray XD Series; model HPE Cray XD675; form factor 4U, OCP Data Center Ready (DCR) form factor; gpu configuration Up to 8x NVIDIA H200 SXM5 Tensor Core GPUs; gpu interconnect NVLink 5th Generation, 900 GB/s bidirectional per GPU pair. Manufacturer part number R9G65A. For the full datasheet with electrical, environmental, and compliance details, contact our pre-sales engineering team.