How do markets economize on scarce natural resources? With an application to fossil energy, we emphasize technological change aimed at saving on the scarce resource. We develop quantitative macroeconomic theory as a tool for interpreting the past and thinking about the future. We argue, first, that aggregate U.S. data calls for a short-run substitution elasticity between energy and the capital/labor inputs that is near Leontief. Given this fact and an aggregate CES function, we note that energy-saving technical change took off right as the oil shocks hit in the 1970s. We rationalize this observation using a theory that views technical change as directed: it can be used to save on different inputs and, hence, the long-run substitutability between inputs becomes higher than Leontief. For our application, we estimate long-run dependence on fossil energy - measured by its factor share - to climb to a little below 10%; absent endogenous technical change directed toward energy-saving, it would go to 100%.