The rapid transition to a decarbonized energy economy is widely believed to hinge on the rate of cost improvements for certain clean energy technologies, in particular renewable power and energy storage. This paper adopts the classical learning-by-doing framework of Wright (1936), which predicts cost (price) to fall as a function of the cumulative volume of past deployments. We examine the learning rates for key clean energy system components (e.g., solar photovoltaic modules) and the life-cycle cost of generating clean energy (e.g., wind energy and hydrogen obtained through electrolysis). Our calculations point to significant and sustained learning rates, which, in some contexts, are much faster than the traditional 20% learning rate observed in other industries. Finally, we argue that the observed learning rates for individual technologies reinforce each other in advancing the transition to a decarbonized energy economy.