https://www.youtube.com/watch?v=MKbNOmysJQo
This is something I wondered about.
The fact is that 64 bits is large enough to hold almost any number we would need. Even 32 bits is adequate. There might be special cases in science where a higher level of precision is required, for example, if you need to calculate a flight path to Pluto. Still, current 64-bit computers can do 128-bit computations if needed but less efficiently than a 128-bit processor.
With the extra circuitry required for a 128-bit processor, the space on the CPU die is better spent having more processors. So there is no reason for general-purpose CPUs to ever be 128-bit.
Graphics cards, which do a massive number of calculations every millisecond, can have processors that use more than 64 bits so that they can move data to and from memory in larger chunks. This is a special case.
