An independently developed open-source Linux kernel module called GreenBoost has been reported today. It aims to augment the dedicated video memory (vRAM) on NVIDIA discrete GPUs by utilizing both system memory and NVMe storage. The primary intent of GreenBoost is to provide a CUDA caching layer, allowing for the execution of larger AI models, specifically Large Language Models (LLMs), that might otherwise exceed the capacity of a graphics card's dedicated vRAM. This development seeks to make more extensive AI workloads accessible on NVIDIA hardware.