This dataset contains molecular dynamics (MD) trajectories of aqueous zinc chloride (ZnCl₂) solutions at concentrations ranging from 1 mol/kg to 30 mol/kg, with 30 mol/kg representing the "water-in-salt" regime. Simulations were performed using deep neural network potential trained with DFT data. Details about the neural network potential can be found in the publication: PRX Energy 4, 023004 (10.1103/PRXEnergy.4.023004). The simulation used a 0.5 fs timestep. For each concentration, the dataset includes three 100-ns trajectories for a smaller system (about 1600 atoms) and one 20-ns trajectory for a larger system (about 15,000 atoms). These trajectories can be used to study ion solvation, structural organization, and dynamical behavior as a function of concentration.