TBH, I don't think the project could sustain itself as well if I did that. Having a single player version ready and out there to keep my own motivation up and attract funds to the project is, in my mind, more important. Additionally, I don't know much about networking or how it should work. The way a client and server will communicate will be largely based on how the game is written, something that logically would need to happen first.
Good choice IMO. While retrofitting the networking part later on most probably WILL be more work, I also think that omitting the networking layer in the beginning on a hobby project with limited resources is a good idea.
As a side note, be aware that doing local multiplayer as in Splitscreen multi player is pretty easy, as long as you can scale the graphics performance with mutliple players and game cameras on the same box.
This way, you can still test your games Multiplayer-readiness from a game design perspective on a smaller scale, and to some people, the fact more and more games omit local multiplayer, is a real let down. Nothing better than to sit in arms reach of your buddy while gaming, so you can smack him if he is again playing unfair ;)
