If you transition now, the incompatiblities for most apps are small to nonexistent.
If, say, you transition 3 years from now when GCC can't build a working binary for a new iOS ABI version that the App Store rules require all new submissions support, and the syntax clang supports has diverged from what GCC was accepting, that's more time needed and dollars lost for every minute the porting takes you.
But that's just a guess at the future. Clang/LLVM provides significant benefits today. Better error messages, better warnings, a seriously great static analyzer (given the constraints imposed by C's semantics), and Automated Reference Counting are just a few of them.
One would hope Clang isn't going to drop support for syntax that it supports today, because that would be a backwards-compatibility issue. However, I do agree that switching earlier rather than later is better. At the very least, the longer you put off the switch, the more code you'll have written that may need to be adjusted to work properly under Clang.
I ask because I'm an iOS dev and I don't know anything about this issue.
Why would the transition be much costlier later? More imperative I understand, but why more costly?