Zambian farmers have long experience with maize hybrids and input subsidies. The successful development and diffusion of improved maize seed in Zambia during the 1970s–80s was supported by government commitment to parastatal grain and seed marketing and subsidized provision of services to maize growers. When this system was dismantled under fiscal duress, production of the nation’s staple food—maize—declined. In 2002, concerned that national food security might be jeopardized, the government reinstated subsidies for fertilizers and maize seed with the stated goal of building the resource base of smallholders. We test the hypothesis that the subsidy on hybrid seed use in Zambia is selectively biased both by the delivery mechanism and through self-selection of farmers who choose to exercise their claim. We find that the subsidy is a recursive determinant of seed demand but its recipients have more land, more income, and lower poverty rates. In 2010, we estimate that 14% of smallholders had a high predicted demand for hybrid seed but did not grow it—and were not reached by the program. This paper contributes to an emerging body of literature that documents the effects of the new generation of “smart” input subsidies in Africa, with a focus on seed (as compared to fertilizer), and a detailed exploration of demand segments.