Very true. However certain NP problems (problems not computationally feasible to solve on a classical computer) might just be BQP (solvable in polynomial time on a quantum computer.) The question is whether SHA-256 (or SHA-128) is NP-complete: if so, it is probably not BQP.
*If* BQP = P then classical computers / Turing machines can run Grover's quadratic (square root time) search. After a few hundred thousand mined the hardness catches up and process is slow again. If additionally, GP constant time search [ http://arxiv.org/abs/1303.0371 ] is in BQP then all remaining coins are mined/minted near instantaneously (no SHA hardness is sufficient to slow the search). Further, double spending is possible if one can search in constant time.
QKD is a possible solution to keeping transaction integrity --- http://en.wikipedia.org/wiki/Quantum_key_distribution
No matter how 'unlikely' that all might sound by running standard software, just keep in mind that strictly speaking no theorems (including Grover's optimality for linear QC) would be violated.