NVIDIA Spectrum-X networking for HPE
Ethernet AI fabric switching up to 800GbE, supported within the HPE portfolio.
- Who it's for
- Enterprises with an established Ethernet networking team
- Scale
- Level 5
A complete rack server, not a graphics card. It holds eight AMD Instinct accelerators and two AMD EPYC processors, and is built for organisations training or serving large models on AMD silicon.
No payment is taken online. Every request is an enquiry answered by an engineer.
Where does this sit?
Enterprise. Four to eight enterprise accelerators in a rack chassis, sized for production load rather than experiments. Typically bought by an enterprise ai or platform team. Datacentre or a well-provisioned rack.
Who it's for
What it runs
When it is the wrong answer
Every figure below is explained in plain English. Switch to the technical view for the bare specification.
8U rack, air-cooled eight-way GPU server
2x AMD EPYC 9004 / 9005 series
MeansThe general-purpose CPUs that feed and coordinate the accelerators.
Up to 8x AMD Instinct MI300X or MI325X (OAM)
MeansOAM is a server-only accelerator module. These cannot go in a desktop.
MattersAccelerator count and memory decide which models fit.
DDR5, 24 DIMM slots per QuickSpecs — maximum capacity awaiting verification
NVMe / SAS / SATA SFF bays — exact count awaiting verification
OCP 3.0 slots plus PCIe Gen5 NIC options — exact port configuration awaiting verification
Air-cooled
Multiple redundant high-wattage PSUs — exact rating awaiting verification
Verification
Manufacturer documented — HPE ProLiant Compute XD685 QuickSpecs a00073553enw. Reviewed 2026-08-31.
Manufacturer sourceThe things that catch people out after the hardware has already been ordered.
What else will I need?
Relationships documented by the manufacturer, or by us during a deployment.
Ethernet AI fabric switching up to 800GbE, supported within the HPE portfolio.
200Gb/s InfiniBand switching and adapters for HPC and AI clusters.
Scale-out object storage built for AI data pipelines and RAG.