


Dell PowerEdge XE9685L
4U Liquid-Cooled GPU Powerhouse for Next-Gen AI & HPC
Why It Stands Out
Blazing Compute Density – Dual AMD EPYC™ 9005 Series CPUs (up to 192 cores each) paired with 8× NVIDIA HGX B200 GPUs (180 GB HBM3e) interconnected via NVLink, delivering extreme performance in a compact 4U chassis.
High-Speed Memory & Expansion – Supports up to 3 TB of DDR5 across 24 DIMM slots, plus 12 PCIe Gen5 slots for flexible I/O scalability.
Efficient Thermal Design – Direct liquid cooling captures 70% of system heat, enabling sustained GPU performance, quieter operation, and lower energy consumption.
Rapid AI Training & Inference – Up to 2.5x faster training than prior H100-based systems, making it ideal for large language models, diffusion, and recommendation workloads.
Enterprise Resilience – Redundant 3000 W Titanium power supplies, robust hardware security, and full lifecycle management with Dell iDRAC and OpenManage.
Summary Table
FeatureSpecificationChassis4U rackmount with direct liquid coolingProcessorsDual AMD EPYC™ 9005 Series, up to 192 cores eachGPUs8× NVIDIA HGX B200 (180 GB HBM3e, NVLink)Memory24× DDR5 DIMMs, up to 3 TBExpansion12× PCIe Gen5 slotsCooling SystemDirect liquid cooling (70% heat capture)Power6× 3000 W Titanium redundant PSUsManagementDell iDRAC9, OpenManage, secure boot, TPM 2.0
Attension: this item may be under US border control depends on configuration, please contact us for more detail.
4U Liquid-Cooled GPU Powerhouse for Next-Gen AI & HPC
Why It Stands Out
Blazing Compute Density – Dual AMD EPYC™ 9005 Series CPUs (up to 192 cores each) paired with 8× NVIDIA HGX B200 GPUs (180 GB HBM3e) interconnected via NVLink, delivering extreme performance in a compact 4U chassis.
High-Speed Memory & Expansion – Supports up to 3 TB of DDR5 across 24 DIMM slots, plus 12 PCIe Gen5 slots for flexible I/O scalability.
Efficient Thermal Design – Direct liquid cooling captures 70% of system heat, enabling sustained GPU performance, quieter operation, and lower energy consumption.
Rapid AI Training & Inference – Up to 2.5x faster training than prior H100-based systems, making it ideal for large language models, diffusion, and recommendation workloads.
Enterprise Resilience – Redundant 3000 W Titanium power supplies, robust hardware security, and full lifecycle management with Dell iDRAC and OpenManage.
Summary Table
FeatureSpecificationChassis4U rackmount with direct liquid coolingProcessorsDual AMD EPYC™ 9005 Series, up to 192 cores eachGPUs8× NVIDIA HGX B200 (180 GB HBM3e, NVLink)Memory24× DDR5 DIMMs, up to 3 TBExpansion12× PCIe Gen5 slotsCooling SystemDirect liquid cooling (70% heat capture)Power6× 3000 W Titanium redundant PSUsManagementDell iDRAC9, OpenManage, secure boot, TPM 2.0
Attension: this item may be under US border control depends on configuration, please contact us for more detail.