There need be no event horizon for the curvature to be non-negligible.

Go on, fix it, I'm gonna go read up on string-theory. What good is modern physics if you can't prove that anything someone says about physics is just an asymptotic approximation?