When I was in the early teens (so quite a bit older than the kid in your case), I got my hands on a book about set theory, and it absolutely blew my mind. The concept of countability, axiomatic definition of functions and so on really gave me a completely different perception of math. Up until then, math seemed to be something that follows nature, so to speak, three plus two makes five because if I have three eggs and two eggs its five eggs or whatever primary school taught me. I remember back then that I'd wish that some teacher would have made me aware earlier of a more formal, axiomatic approach to math and all that, that there is a more fundamental basis to it than "that's just what it is". It really furthered my interest in math, and while I ultimately eventually moved over to CS, it definitely had quite a fundamental impact on me back then.
The particular book I've read was in German (I still have it) and unlikely to be ideal for a 5 year old; just wanted to give this little personal anecdote somewhat related to your question.
Science-fiction can be stimulating in that regard, someone who likes science in addition to math might be the one to turn some of it into science-reality someday :)
Eventually I came to the conclusion that the better math textbooks had more than just math. One of the most enlightening math books I had was a late 19th century antique text "Progressive Higher Arithmetic" which was never in one of my courses, but was given to me when I was an old teenager and it could be an interesting reference in any century. Not only the differences in the way teaching has been done in the past, but a nice leather-bound volume with numerous blank pages among those adjacent to the front & back covers, covered with poetry written by the original owner. With a pen and common inkwell, in beautiful handwriting there is one that I can not forget:
Remember me when death shall close
My eyelids in the last repose
And evening breezes gently wave
The grass above your school-mates grave.
If you could find a collection of the Mathematical Games articles from Scientific American, those were entertaining and required little background knowledge, but conveyed the fun of math.
Geometry is a good subject because it doesn't have very many prerequisites other than basic arithmetic, and the older textbooks were mostly about proofs. (Note I'm biased, proofs are what made math come alive for me).