HPE Apollo 6500 Gen11 (XD685) AI/HPC Server (8x AMD MI300X / NVIDIA H100)

The HPE Apollo 6500 Gen11 (XD685) is a high-density accelerator server engineered for demanding AI training, inference, and HPC workloads. The platform supports up to eight AMD Instinct MI300X APUs or NVIDIA H100 SXM5 GPUs, delivering exceptional parallel compute density within a 5U chassis. Built on AMD EPYC 9004 series (Genoa) processors, the system leverages PCIe Gen 5 interconnects and high-bandwidth NVLink or AMD Infinity Fabric architectures depending on accelerator configuration, enabling the sustained memory bandwidth and low-latency GPU-to-GPU communication essential for large-scale model training.

The XD685 chassis integrates HPE's high-efficiency direct liquid cooling (DLC) capability alongside traditional air cooling, addressing the substantial thermal demands generated by eight fully loaded accelerator cards. The server supports dual AMD EPYC 9004 series CPUs, high-capacity DDR5 memory via 24 DIMM slots, and multiple NVMe SSD bays for fast local storage staging. Front-accessible PCIe expansion and OCP 3.0 network adapter slots provide flexible high-speed fabric connectivity, supporting 200GbE, InfiniBand NDR, and other low-latency interconnects required for distributed AI cluster deployments.

Targeted at enterprises, hyperscalers, and national research organizations building out GPU-dense AI infrastructure, the Apollo 6500 Gen11 platform is optimized for generative AI model training, large language model (LLM) fine-tuning, computational fluid dynamics (CFD), molecular dynamics simulation, and seismic processing. The system ships under HPE part number P65010-B21 and integrates into HPE's broader Apollo HPC ecosystem, compatible with HPE Cray supercomputing software stacks, HPE GreenLake management, and standard data center rack and power infrastructure.

Ideal for

• Large language model (LLM) and generative AI training using distributed GPU clusters with NVLink or Infinity Fabric high-bandwidth GPU interconnects
• Enterprise AI inference at scale for real-time natural language processing, image recognition, and recommendation engine workloads requiring sustained GPU throughput
• High-performance computing simulations including computational fluid dynamics (CFD), finite element analysis (FEA), and molecular dynamics for life sciences and engineering
• Seismic data processing and subsurface imaging in energy sector workflows requiring dense floating-point compute over large datasets
• Multi-node deep learning research and model development in academic, government, and national laboratory environments integrated with HPE Cray software ecosystems
• AI-accelerated drug discovery and genomics pipelines where AMD MI300X unified memory architecture supports exceptionally large in-memory datasets without CPU offload

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.