You could defend the "per year" number by pointing out that power needs for homes vary a lot by season, so averaging over a year is the best long term measure.

But of course that means one of these reactors isn't the perfect plant to heat a city of 50k homes, since it would be underpowered part of the year and overpowered other times.