Ah. Yeah, that's not a sure thing, but progress has been very, very encouraging lately, and there's no obvious fundamental limit, either. Advancement in the error correcting codes has even happened, which was unexpected when the big push to build a QC began.

(FWIW, Grover's search algorithm is more of a toy than anything. You're getting a quadratic speedup on exponentially hard problems, which leaves them still exponentially hard)