I would assume that any efficient hypersonic airliner would need to use variable geometry or powered manipulation of airflows to attain a reasonable efficiency in the transition zones of its flight envelope, so while you obviously would not get all of the acceleration energy back, it would transition into efficient lower velocity flight and reduction in engine thrust needed during the deceleration phase. But if we really knew how to do all of this we would already be flying transatlantic in an hour.

For the same reasons that theoretically possible doesn’t mean necessarily practical, theoretically difficult doesn’t mean practically impossible. We just won’t know until we significantly advance the state of the art. We might find out that atmospheric flight is a waste of time and that exoatmopheric ballistic trajectories are where it’s at. Or we might find out that zeppelins were the way all along.

We are still a long way off from having engines that can operate any where near theoretical efficiencies except within very narrow parameters and altitudes. Until we sort that we will have difficulty traversing a wide speed envelope without waste.