In previous studies, the asset growth anomaly is found to be driven by less profitable firms or firms with losses. However, we provide contrary evidence that the risk-adjusted return on the low-minus-high asset growth portfolio is statistically and economically significant among more profitable firms, when controlling for financing constraints. In fact, the asset growth effect is most pronounced in firms with both high profitability and high financing constraints. We present a theoretical framework demonstrating that these results are consistent with q-theory, supporting the hypothesis that the discount-rate channel underlies the asset growth effect. In our analysis, we employ the most up-to-date, machine-learning-based indices developed by Linn and Weagley (2023), to enhance accuracy of measuring financing constraints.