Hypersonic missiles are still in the prototype/research stage, the short/medium-term nuclear threat remains non-hypersonic. Regardless, right now there are zero systems that can hope to stop a hypersonic missile, reaction time limitations alone are a huge hurdle there.
That said, effective missile defense is theoretically possible even given current tech, it's just extremely hard. In that sense it's not different than the various AI initiatives. Lots of wasted money on goose chases, but also lots of advantage for those who get it right, and some legitimately useful innovations along the way.