Enhancing Generative AI Storage and Performance: Collaborating with NVIDIA