By any chance was this VM created under Workstation 16.2, and used the "experimental" vTPM implementation?
If so, I'm thinking that the oh-so-broken-and-should-never-be-used "experimental" TPM implementation of that release is at the root of this issue, Two of the major deficiencies in the 16.2 experimental vTPM implementation were that
I'm beginning to suspect that that auto-generated encryption password is based on the BIOS UUID (or something similar in the system's firmware). Changing the BIOS may have updated whatever that vTPM implementation is basing its encryption key on (perhaps a BIOS UUID), which would be the equivalent of moving the VM to another machine.