Can you provide a reference about the fact that 100 qubits would enable qualitative advances in cryptanalysis? I'm also not sure what you mean by `qualitative'.
Gover's Algorithm[1] halves the time to search an unsorted list. Which basically means the time to pick a random number 200 bits long, is now equal to the time to pick a random number 100 bits long. Keys as large as 60 bits are currently vulnerable to attack[2], so logically post grover's algorithm we could attack 120 bit symmetric encryption.
If that doesn't make sense Bruce Schneier explains it also. [3]
