QKD provides a method of sending a session key with the ability to detect unauthorized interception, i.e., potential attack. This in no way removes the possibility of side-channel or brute-force attacks. In addition, while we currently cannot detect the state of a photon and send it along unchanged or detect and reproduce it, that doesn't mean we never will. That would allow a man-in-the-middle attack and defeat the theoretical assurance of invulnerability of key transfers.

Either way, QKD is very technically difficult (typical line-of-sight transmission and no possibility of repeaters) and has not been used widely outside laboratory conditions. I believe it never will achieve widespread application, and especially will not withstand breaks developed in the long-term. This will happen through large-scale quantum computer development or relevant advances in quantum theory.