Large Language Models (LLMs) demand immense amounts of memory, and the more people use them, the more memory is required.