Liquid hydrogen cooled superconductors (20K) would be good especially if we manage to use some materials that don't suffer hydrogen embrittlement processed/machined economically enough to really play out the cost benefits of not relying on the rare noble gas.