Yeah I worked it out and I am saving ~75% vs running my inference workloads on R...

		choilive on Sept 10, 2024 \| parent \| context \| favorite \| on: Serving AI from the Basement – 192GB of VRAM Setup Yeah I worked it out and I am saving ~75% vs running my inference workloads on RunPod. $650/mo in electricity vs $2,500/mo to do the same thing on RunPod. Been in near continuous operation over 9 months, so the system has basically paid for itself with the savings.