Most plants are hermaphrodites, producing both pollen and seeds, but there are many species in which some individuals are morphodites and others are purely female. Often this femaleness (male sterility) is caused by a mitochondrial mutation.
I once heard Bob Trivers explain this: it’s simple and interesting. Hermaphroditic individuals often self-fertilize, which is a gift and a curse. It’s a gift because seeds from self-fertilization have two copies of the plant’s genome, rather than one: fitness is increased, all else equal. It’s a curse because of inbreeding depression: lots of homozygosity makes one weak. From the plant’s point of view, having two sexes and selfing is a good thing as long as the extent of inbreeding depression is less than one-half. But mitochondria are only transmitted maternally, and so have nothing to gain from inbreeding- no extra copies get transmitted. While they suffer from being inside inbred, gap-toothed seeds. So, any mitochondrial mutation that prevents selfing (by eliminating pollen production) increases mitochondrial fitness, while generally reducing nuclear gene fitness.
Also, energy spent on pollen production is now available for making more seeds, but that’s a secondary effect, usually.