Having computers that are fast enough that we can ray-trace everything, would be nice, yes... Though for an initial batch of camera or light rays that are all uniformly distributed and not bouncing around, rasterization still might be more efficient

You haven't given us much to discuss though, so googling turned up this paper that actually gives some big-O figures on the algorithmic improvements that are up for grabs when we have qubits available to us:

http://www12.tuiasi.ro/users/103/f4_2012_2_Caraiman.pdf

The TL;DR big-O parts:

The complexity of the quantum Z-Buffering algorithm is O(P*sqrt(d*N)) while the classical algorithm needs O(P*(d+N))) steps. Here P represents the total number of pixels, d is the average number of polygons intersecting one pixel and N is the number of polygons in the scene. The main advantages of the quantum variant of the algorithm are the fact that it has a much better scalability with respect to the number N of polygons than its classical counterpart and that it copes with arbitrarily shaped objects and it is not limited to simple polygons
...
The main advantage of this quantum [ray-tracing] solution over the classical case is the optimal time complexity for general object queries, with linear space complexity. In the classical case, polygons are used because computing the intersection is easier than in the case of generic objects
...
Thus, the first step of the quantum photon mapping algorithm has O(sqrt(N)) time complexity, unlike O(N) in the classical case