Because the difference between a 64 bit integer (which would also have been future-proofed for the current IP service model) and a 128 bit integer is not simply 8 more bytes, but also the fact that all modern non-MCU computers can treat a 64 bit integer as a scalar, but are effectively forced to handle a 128 bit integer as a string or a structure of some sort.
This can be the difference between a 1-line patch to a C program and a 30 line patch.
Of course, the standards committees don't care about that cost (it is an externality to them), because "rough consensus and working code" stopped being the code of the IETF more than a decade ago.
This can be the difference between a 1-line patch to a C program and a 30 line patch.
Of course, the standards committees don't care about that cost (it is an externality to them), because "rough consensus and working code" stopped being the code of the IETF more than a decade ago.