Cultural transmission in nonhuman animals is often sex biased, with females more frequently or efficiently learning cultural behaviors than males. The evolutionary origins of sex-biased cultural transmission have been a mystery, though it has been proposed that female offspring may gain greater reproductive benefit from cultural traits than sons—the “disparate benefits” hypothesis. I propose a different, “uniparental teaching,” hypothesis where sex-biased transmission evolves in uniparental species if mothers teach, that is, invest in their offsprings’ learning. I show, with theoretical models, that mothers evolve to invest more in teaching daughters than sons because teaching daughters results in greater inclusive fitness benefits. Teaching a son gives him a reproductive benefit for one generation. However, I show that because daughters may teach future generations, teaching a daughter can be a better long-term investment. I also model the disparate benefits hypothesis and show that the uniparental teaching hypothesis better fits the empirical patterns of sex-biased transmission in the well-studied example of “sponging” in bottlenose dolphins. Uniparental teaching may also explain sex-biased transmission in other species, including chimpanzees. My findings suggest that controversial mechanisms of cultural transmission in nonhumans, such as teaching, may be inferred from population-level patterns of transmission even when it is difficult to observe transmission directly in the field or laboratory.