In this page (http://mathworld.wolfram.com/FermatPrime.html) we have the following result:
$2^{2^n}+1$ is a Fermat prime if and only if the period length of $1/(2^{2^n}+1)$ is equal to $2^{2^n}$. In other words, Fermat primes are full reptend primes.
Thus, we get:
$2^{2^n}+1$ is not a Fermat prime if and only if the period length of $1/(2^{2^n}+1)$ is strictly less then $2^{2^n}$. See: Period of the decimal expression for the rational number $\frac{1}{n}$ is at most $n-1$
My question is: Can we deduce that there is infinitely many indices $n$ such that the period length of $1/(2^{2^n}+1)$ is strictly less then $2^{2^n}$.