Research institutions are at the forefront of the convergence between traditional high-performance computing and modern AI workloads. Leaders from the University of Utah and Cineca detail their experiences deploying and scaling VAST in their parallel computing environments, supporting some of the biggest supercomputers in research and education. This session provides practical comparisons of VAST against traditional parallel file systems, discusses applications using parallel I/O on VAST clusters, and examines how the surge in AI research is shifting priorities for HPC centers. We will explore which VAST data services are most valuable to research institutions and provide practical guidance for operators looking to deploy, scale, and simplify HPC infrastructure. Learn how these leading institutions are modernizing their data foundations to support both traditional simulation workloads and the next generation of AI-driven scientific discovery.