I personally tend to put too much time into looking for a domain name too early in the project (because it's fun to do, honestly). It's much better to at the very least do serious research first (competitors, understanding and talking to possible customers, etc.), and have done proof-of-concept work for technical challenges, and done some design on your minimum-viable-product first version before worrying about a domain name, because it's so frequent to find some serious problem during that process that will cause you to abandon the project (or perhaps realize that you don't find it as interesting as you thought initially -- if you don't have some serious passion for solving this particular problem, it's not likely you'll have the motivation to get through the hard parts of building a business around it!).
If you've already gotten past those steps -- sure, go for it. Upper 3-figures isn't a high number if this feels like a very good candidate.
I don't recommend buying it as an investment, though... There was an HN thread a few weeks back on the subject of assessing domain names, and in general they aren't very valuable. Yes, many names are taken, but there are so many businesses that are succeeding in spite of quite odd domain names (hell, thinking of domain names, what about "GoDaddy.com" for a completely random domain name?).