This presentation will describe recent advances in national severe storm forecast services and planned evolution from a few years to a decade; from frequently updated, very high resolution probabilistic severe weather outlooks, to more nimble and effective alerting messages. The main focus will be the physical science challenges that need to be addressed in order to make this evolution effective. A top priority is the need for improved scientific understanding of convective storm processes to support enhanced probabilistic prediction of severe thunderstorms and tornadoes, likely through direct numerical prediction complemented by expert human forecasters with access to enhanced observations (especially of the boundary layer) and numerically assimilated 3D analyses. Other community challenges include, but are not limited to: Mesoscale Convective System intensity forecasts and resultant specific weather hazards, severe storm threats in high shear/low CAPE environments, extraction of effective forecast information and insights from next-generation, extremely high resolution, convection allowing NWP ensembles, and visualization of these data utilizing foundational dynamical quantities to gain insight into the current structure and future evolution of severe thunderstorms. These topics will be offered to spur discussion on key community issues and ways to work toward their solution.