This paper investigates the long-run effects of immigration on wages and welfare in a model with endogenous technology choice (ETC) where firms are allowed to choose their optimal skill intensity from a menu of available technologies. I embed the ETC framework into the Auerbach and Kotlikoff model (1987) that features a large set of overlapping generations, a rich collection of population dynamics, and a social security system. I calibrate the model to match with the U.S. data and evaluate the effect of ETC with the help of two experiments. In the first experiment, I increase the share of high-skilled immigrants and compare the wage and welfare predictions of the model with ETC to a standard model where the skill intensities in production technology are fixed. In the standard model, since the skill intensities are constant, increase in the supply of high-skilled labor leads to a decrease in high-skilled wages and an increase in low-skilled wages. On the other hand, in the model with ETC, negative supply-side effects are counterbalanced by an increase in the intensity of the more abundant high-skilled labor, leading to a smaller decrease in their wages. The discrepancy between wage predictions of these two models is also reflected in the welfare: while the model with ETC predicts an increase in both high- and low-skilled natives’ welfare, the standard model would predict a decrease in the welfare of the high skilled and a larger increase in the welfare of the low skilled. In the second experiment, I examine the effects of an increase in low-skilled immigration and find that in this case, since the initial production technology is low-skilled intensive, the ETC effects are smaller. These results imply that if ETC is ignored, both in the short run and long run, wage and welfare analyses of immigration will be incomplete, and even misleading.