A user spent £200 to install a datacenter GPU into a consumer gaming rig. The setup requires custom cooling and specific driver workarounds to function outside a server rack. This experiment highlights the accessibility of high-VRAM hardware for local LLM inference. It proves that budget-conscious developers can bypass expensive consumer cards for AI workloads.