The word size of most modern CPUs is 32 or 64 bits. Both are a power of 2.
This seems "intuitively reasonable", because it has a bit of 2 states, but I do not really have a hard argument to explain why this is the case. In addition, in the past, the word sizes of the processors have been multiples of other numbers, such as multiples of 6 (because the characters were encoded in 6 bits).
What is the reason why chip designers have chosen to make word sizes a multiple of 2?