Local AI infrastructure

GX10 Fleet Control

Connecting
Power figures are NVIDIA-reported GPU/device power, not per-outlet wall draw. Unified memory is measured from Linux MemAvailable.

Inference pairs

Nodes