Grapefruit didn't evolve naturally. In fact, it's only a couple of hundred years old. It's the way it is because of human intervention.
Grapefruit are a cross, probably, of pomelo and a sweet orange - as you say, not all that long ago. While some grapefruit varieties are quite pithy, pomelos (that I've seen) are very pithy - hence the 'this can only have been even more expensive a protective layer in earlier versions of this plant'.
The question still stands - why is this adaptation useful, given it's probably expensive. (That assumption of mine may be entirely wrong.)
Early species would have been shorter, and all species in the history of this would have fruited long before the individual trees reached 10 metres (I'm guessing even in ideal conditions that implies 10+ years growth). Most would-be consumers of the fruit would be able to reach fruit within the first metre or so (if unable to climb or fly) and anywhere in the tree otherwise.
What kind of animals live, or lived, in wild citrus' native range? Now I'm curious. [EDIT: Himalayan foothills... Macaques? Elephants?]
Or it's selected for particular animals as spreaders? In wild grapes, for example, most are adapted for bird spread (small dark acidic berry, high up in a tree on dangling shoots) but there are a handful of species (vitis labrusca "fox grape" for example, and vitis rotundifolia aka muscadines) that have adapted for mammals and they are quite different: larger berry, lighter colours, strong smell (think Concord), lower acids, slip skin, and a tendency to "shell" (fall off the vine when ripe). They also tend to grow wild in shadier moister areas, e.g. the underbrush where foxes and skunks and racoons etc. will grab them, not birds.
I understand the 'travel far and wide', though of course evolution's not directed, and is there significant difference in fitness & success over the long term for a variation that allows a plant to produce offspring 1km away in one generation (say 5y), rather than taking 2 generations?
(And, of course, it's worse than that. The 10m drop potential isn't realised until the plant is probably 15 years old or more -- at which point the distance / propagation calculation is almost irrelevant, as that specimen would have produced fruit for at least 5 years whose seeds would have travelled the same distances, regardless of this (future) robustness capability.)
Sadly, though, I suspect plant archaeologists have more pressing concerns than this question.