The purpose of this study was to determine the quantitative variability of apparent diffusion coefficient values across a large fleet of MR systems. Using a NIST traceable magnetic resonance imaging diffusion phantom, imaging was reproducible and the measurements were quantitatively compared to known values. Significant differences in identical phantoms were not observed, but uncertainty in the measurements was seen at low apparent diffusion coefficient values. The same trend was observed when the diffusion phantoms were imaged in 20 different MR systems. The characterization of ADC variability for these systems provides an improved quality control for quantitative diffusion weighted imaging.