"Regenerative cooling" is a rocket term. It came to mind because I used to be an aerospace engineer. There's no requirement in the definition of regenerative cooling for the coolant to be dumped.
> Presumably active cooling of a hyperloop tube would use a closed refrigerant cycle
I don't think one can just presume that. You've already got lots of pumps for pumping air. Given (a) the seal on the tube will be periodically broken (for entry, exit and maintenance) and (b) a safety factor, you'll have more pump capacity than you'll need. Filtering and then compressing atmosphere, running it through a heat exchanger, and then letting it expand through the cooling channels before dumping it doesn't seem obviously worse than having kilometers of refrigerant running around.
> which has little to do with regenerative cooling, besides the idea of cooling channels
See above. Also, I assumed if you did this you'd use it to boost SpaceX's nozzles' economies of scale.