The top answer to this question: Power spectral density vs Energy spectral density states:
However if random process is stationary, then it is for sure that it has some finite power over some band of frequencies
Why is that? Also, is stationarity a necessary condition for a process to have finite power over a band of frequencies?