Only if you're caught unaware. It turns out that things like GUIDs (for example) have this nice even distribution across the entire domain of rows in the table or index and it's something you can take huge advantage of that you can't even do with ever-increasing "append only" rows and that's go WEEKS and even months without ANY page splits, not even the "good" page splits that plague append-only indexes.
Unless there's a concentrated campaign, telephone numbers are frequently very close to the even distribution of GUIDs and can take that same advantage.
The trouble is, people keep screwing up such wonders by using current "Best Practice" methods of index maintenance which, ironically, creates a nasty drug-habit-like Catch-22 that creates the need for defragmentation.
And, no... updates alone don't cause page splits unless unless you have a page that's 100% full or you have an "expAnsive".
It all comes down to Fill Factor, when to defrag, and whether to use Reorg or Rebuild... and the current "Best Practice" methods aren't the right combination for any of that. Heh... I've got a couple of hundred hours in testing to prove all of this and am going to try to summarize it in a 2 hour presentation at the Pittsburgh SQL Saturday this year.