Could a rover drive to orbital velocity? Could a rover, assuming it is on a perfectly smooth sphere in space, eventually reach orbital velocity? Would there be a limit on the possibile speed?
 A: Note that if the sphere is perfectly smooth, the rover will actually not have any grip on the surface and its wheels will just spin uselessly. But if we assume there is grip and the surface has no obstructions, then by exerting torque on the wheels it will accelerate forward. This will either end by friction forces and air drag producing enough of a counter force, or indeed that it reaches orbital velocity and starts orbiting just above ground.
Orbital speed is $\sqrt{GM/r}$. For a spherical Ceres this is 364 m/s. This is about twice the world record for an electric powered vehicle on Earth, but just an inch above the current world land speed record using a jet-propelled car. Obviously a rocket driven rover should be able to get to orbital speed. For a smaller body the speed requirements are lower. If we take Aegaeon (Saturn LIII) the velocity is around 0.29 m/s, in which case walking would give you orbital speed.
A: With no atmosphere there isn't a fundemental difference between reaching orbit by going up vs along. The reason rockets are launched vertically is to get out of the thick lower atmosphere as quickly as possible before you really start moving at high speed.
On a moon you could build a horizontal track and launch a rocket along it. Assuming it would reach escape velocity before the curvature of the ground was significant, otherwise you would need some way of holding it 'down'.
