There really is no benefit to creating a 128-bit operating system. A 128-bit (or higher) architecture is useful for certain mathematically intensive operations, such as graphics, cryptography, and/or complex system modeling, but the question was specifically about operating systems.
The biggest problem with 32-bit operating systems was the 4 gigabyte limit on addressable memory space. But we didn’t really need to go to 64-bit architecture and OS to solve this problem… in fact, a 40-bit architecture would have allowed 1 terabyte of addressable memory space, which would have been more than enough for the foreseeable future.
You see, for every bit you add to the architecture, you are doubling the number of addresses that are available. The addresses are basically the number of combinations that can be formed given a certain number of bits. For example:
1 bit = 0 or 1, which is a total of 2 combinations
2 bits = 00, 01 , 10, or 11, which is a total of 4 combinations
3 bits = 000, 001, 010, 011, 100, 101, 110, or 111, which is a total of 8 combinations
So going from 32 bits (which is a total of 4,294,967,296 combinations) to 64 bits (which is a total of 18,446,744,073,709,551,616 combinations) is already overkill. And it’s not just the addressable space that has been drastically increased, take a look at this chart (this is for Windows OSes):
Ok, so the 64-bit architecture has us covered for QUITE some time. Now, just for fun, let’s look at how many combinations we would get from 128 bits:
340,282,366,920,938,463,463,374,607,431,768,211,456 (corrected)
vs 64 bits:
18,446,744,073,709,551,616
So, yeah, that would be like killing a fly with a tactical nuke—overkill doesn’t even begin to describe it!
Sources: Internet