Brand: HPE | Category: Servers
SKU: HPE-XD685-B200 | Part #: XD685-B200 | MPN: XD685-B200
Contact for Pricing — Request a Quote
HPE Compute XD685 — 5U HGX B200 platform, HPE engineering for next-gen Blackwell AI workloads.
The HPE ProLiant Compute XD685 8x B200 SXM6 AI System (part number XD685-B200) is a flagship enterprise AI server distributed worldwide by Omnixon Global. Built around the NVIDIA B200 accelerator family and engineered for production AI infrastructure, it is one of the most-requested platforms for foundation-model training, fine-tuning, and large-scale inference in 2025–2026.
Yes. Every unit ships, serial registration, and the official software/firmware stack.
Yes. We routinely deliver 8-node, 16-node, and 32-node B200 clusters with NVIDIA Quantum-2 InfiniBand or Spectrum-X Ethernet fabrics, parallel storage (DDN, VAST, Weka, Pure FlashBlade//E), and full rack/power/cooling design. Send your training-cluster BOM and we'll respond with a complete priced design.
Omnixon Global engineers configure full AI clusters — networking, storage, cooling, racks, deployment.
Request a Quote → Or talk to an engineerOmnixon Global supplies enterprise IT hardware. Request a quote or contact our team for project design and Bill-of-Materials review.
| Brand | HPE |
| Category | Servers |
| SKU | HPE-XD685-B200 |
| Part Number | XD685-B200 |
| Condition | New |
| Form Factor | 5U |
| Interface | NVMe |
| GPU Support | 8× NVIDIA B200 SXM6 192GB HBM3e (1.5TB aggregate HBM3e) |
| GPU Count | 8x |
| AI Optimized | Yes |
| Processor | 2× Intel Xeon 5th Gen Emerald Rapids |
| Max Memory | 32× DDR5-5600 DIMM slots, max 8TB |
| Drive Bays | 8× E3.S NVMe + 2× M.2 boot |
| GPU Interconnect | NVLink 5.0, 1800 GB/s |
| AI Performance | Up to 144 petaFLOPS FP8 |
| Network | 8× ConnectX-7/8 400GbE supported + BlueField-3 DPU |
| Power Supply | 6× 5500W Titanium redundant |
| Cooling | Direct liquid cooling (DLC) standard |
| Form Factor | 5U rackmount |
| Management | HPE iLO 6 |
The HPE ProLiant Compute XD685 8x B200 SXM6 AI System is built for enterprise data-center workloads — virtualization (VMware, Proxmox, Nutanix), private cloud, database hosting, and AI/ML training. Its 5U Rack Server form factor 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 XD685 8x B200 SXM6 AI System: form factor 5U Rack Server; new condition; gpu 8× NVIDIA B200 SXM6 192GB HBM3e (1.5TB aggregate HBM3e); cpu 2× Intel Xeon 5th Gen Emerald Rapids; memory 32× DDR5-5600 DIMM slots, max 8TB; storage 8× E3.S NVMe + 2× M.2 boot; gpu interconnect NVLink 5.0, 1800 GB/s; ai performance Up to 144 petaFLOPS FP8. Manufacturer part number XD685-B200. For the full datasheet with electrical, environmental, and compliance details, contact our pre-sales engineering team.
