I am never sure why it's thought that QR codes are somehow "fluid" or "dynamic" or dependent on server/net data.
A QR code is simply a barcode, a complex one that can encode up to several hundred characters of data while still remaining "coarse" enough to be read easily. (When the data reaches its limit, the QR pattern can be so fine that any small printing or display flaw makes it unreadable.)
Once generated, all the QR code has, forevermore, is the data that was encoded in it. It will remain readable as long as it is clear enough for a suitable reader/decoder to read and decode it. It does not matter where or how it was generated.
The only thing that is "dynamic" about QR codes is when they actually encode an address or URL that can be redirected. Those destination addresses have to remain available, or scanning the code won't do anything useful... just as if you typed in a 'dead' web address.
But the codes are permanent and fixed at the moment of creation, and can only be changed by re-generating them with different embedded information.