Your first paragraph implies 8-bit bytes are a coincidence, which is not true. It was a design decision.

A byte in computer is the smallest addressable memory location and this location at that time contained a character. The way characters are encoded is called code. Early computers used 5-bits, which was not enough for alphabetics and numerals, 6 bits was not enough to encode numbers and lower and upper case characters, which eventually lead to ASCII.

ASCII was also designed(!) to make some operations simple, eg. turning text to upper or lower case only meant setting or clearing one bit if the code point was in a given range. This made some text operations much simpler and more performant, that is why pretty much everybody adopted ASCII.

Doing 7-bit ASCII operations with a 6-bit bytes is almost impossible and doing them with 18-bit words is wasteful.

When IBM was deciding on byte size a number of other options were considered, but the most advantageous was the 8-bit byte. Note that already with 8-bit bytes, this was over-provisioning space for character code, as ASCII was 7-bit. The extra bit offered quite some space for extra characters, which gave rise to character encodings. This isn't something I would expect a person living in the USA to know about, but users of other languages used upper 128 bytes for local and language specific characters.

When going with 8-bit byte, they also made the bytes individually addressable, making 32-bit integers actually 4 8-bit bytes. 8-bit byte also allowed to pack two BCD in one byte and you were able to get them out with a relatively simple operation.

Even though 8-bits was more than needed, they were deemed cost effective and "reasonably economical of storage space". And being a power of two allowed addressing a bit in a cost effective way, if a programmer needed to do so.

I think your post discounts and underestimates the amount of performance gain and cost optimisations 8-bit byte gave us at the time it mattered most, at the time computing power was low, and the fact that 8-bit bytes were just "good enough" and we didn't get anything usable from 9, 10, 12, 14 or 16 bit bytes.

On the other hand you overestimate the gains with imaginary problems, such as IPv4, which didn't even exist in 1960s (yes, we ran out of public space quite some time ago, no, not really a problem, even on pure IPv6 one has 6to4 NAT), or negative unix time - how on Earth did you get the idea that someone would use negative unix time stamps to represent historic datings, when most of the time we can't even be sure what year it was?

I think the most scary thing is having and odd-bit bytes; there would be a lot more people raging against the machine, if byte was 9 bits.

If you want to know why 8 bits, this is a good recap - https://jvns.ca/blog/2023/03/06/possible-reasons-8-bit-bytes... - along with the link to the book from the engineers who designed 8-bit bytes.

Given how many early computers had 18-bit or 36-bit words, the possibility of 9-bit bytes doesn’t seem as unrealistic as you suggest. I don’t see 8-bit bytes as being so inevitable.