Systems, methods, and storage media for improving fusion energy output by simulating processes of a shperical tokamak are disclosed. Exemplary implementations may: obtain sets of fusion generator operation information; train a machine-learning model with the obtained sets of component information such that the machine-learning model predicts dynamics of plasma, density of plasma, fuel cycle, magnetic fields, or material interactions in a spherical tokamak; store the trained machine-learning model; and simulate, using the stored machine-learning model, dynamics of plasma, density of plasma, fuel cycle, magnetic fields, or material interactions during operation of the spherical tokamak.