I don't think his question was understood. The word stalling is causing everyone to focus on the motors I think.
What I tried to point out to him is the rotational force being applied to the drive shaft at each module, will travel through the least resistance. It can either rotate the wheel, or turn (steer) the entire module. So the resistance is the wheel friction vs. steering torque / globe. You have to look at the assembly to make sense of this.
Crab module picture
Imagine where the 2 (green) bevel gears mesh, if the vertical gear (on the horizontal red shaft) couldn't rotate, the horizontal gear (on the vertical red shaft) would then try to make the vertical gear orbit around it. Then the only thing stopping it would be the steering system torque. Therefore the torque output of the steering system must be greater than the output of the drive system to counteract this imposed rotational force.
This drive force would be constant against the steering system, causing a range of misalignment depending on the wheel resistance. If the bot were pinned, and the throttle was pegged, the 4 wheel modules would be spinning around like tops. The only way I saw to overcome this would be to build a low ratio globe gearbox. Which will have a very low rpm = slow steering.
He sure did an excellent job on this thing!