The increase in bedrock channel width and the decline in bedrock channel slope with increasing drainage area are fundamental characteristics of mountainous landscapes. Compared with slope, little is known about controls on steady-state bedrock channel width. We model steady-state bedrock channel width by iteratively solving models for bed and bank erosion by impacting bedload, using measurable physical parameters, including uplift rate, water discharge, sediment supply, grain size, rock strength, and bank angle. The results indicate that width is largely controlled by sediment flux, rather than water discharge. The commonly used width-discharge scaling relation is an artifact of the covariance of sediment flux and water discharge. Scaling up from cross-section to drainage basin scales, our model reproduces the width-drainage area scaling relation and suggests that the scaling exponent is largely controlled by the downstream change in the fraction of total sediment supply transported as bedload. Bedrock rivers become wider and less steep as water flows from upstream to downstream in mountainous landscapes. The reasons why bedrock channels widen are poorly understood. Here, we develop a method to predict typical channel width based on the mechanism of bedrock erosion by impacting sediment particles. Our method can be tested in natural bedrock rivers from measurable parameters. Our results show that channel width is largely controlled by the amount of sediment supplied from upstream, rather than the amount of water flow. We show the widely observed increase in bedrock river width with drainage area is due to the increase in sediment supply not water discharge. A new model is developed for steady-state bedrock channel width based on bed and bank erosion by bedload particle impactsModel results indicate that the width-drainage area scaling relation exponent is largely set by the coarse sediment fractionThe width-discharge scaling relation is an artifact of the covariance of sediment flux and water discharge