Dell PowerEdge XE9680 LC 8-GPU liquid-cooled AI server. Supports 8× NVIDIA H200/H100/B200 SXM GPUs with factory-integrated direct-to-chip warm-water liquid cooling. Dual 5th Gen Intel Xeon Scalable processors, 32× DDR5 DIMM slots (up to 8TB), and NVSwitch fabric delivering 900 GB/s GPU-to-GPU bandwidth. 6U rack-mount platform with integrated coolant distribution loop — the enterprise AI training flagship from Dell.
6U Rack-Mount
8× SXM (H200/H100/B200)
Dual 5th Gen Intel Xeon
32× DDR5 DIMM (8TB max)
Direct-to-Chip Warm-Water LC
900 GB/s GPU Interconnect
6× 3300W Titanium PSU
8× 400GbE / IB NDR400
8× EDSFF E3.S NVMe
iDRAC9 Enterprise
32°C (W32 warm water)
Up to 12 kW per node
The Dell PowerEdge XE9680 LC is the enterprise AI training platform of choice for Fortune 500 companies and sovereign AI cloud builders. Dell's factory-integrated direct-to-chip liquid cooling uses 32°C warm water (W32 specification), eliminating the need for chillers and enabling PUE below 1.05 in colocation deployments. Each node handles up to 12 kW of thermal load — 8× H200 GPUs plus dual CPUs — while maintaining GPU junction temperatures below 75°C at sustained full load. The iDRAC9 Enterprise management controller provides out-of-band thermal telemetry, leak detection, and automated flow-rate adjustment via Redfish API. QS Compute offers the XE9680 LC as a turnkey rack-scale solution: 32 nodes (256 GPUs) integrated with a CoolIT CHx750 CDU, plumbed, pressure-tested, and commissioned on-site by certified Dell and CoolIT field engineers. Typical deployment delivers >97% scaling efficiency on GPT-175B training across 256 GPUs with 800 Gbps InfiniBand fabric.
Need Dell PowerEdge XE9680 LC? Enterprise liquid-cooled AI servers.
Configured, tested, factory-cooled — deployed with full rack integration and on-site commissioning. Contact us for pricing and solution architecture.
Request Quote