Grin clamping torque arms take the torque arm fit from millimeters worth of looseness & some damage on the bike at >4000w to 0-1mm of looseness, and i don't have long term results but i bet there would be little to no damage from both heavy regen and forward power with a pair of those in a high power hub motor.
Keep in mind that any looseness, over time, equates to rocking of the axle and wear on the contact surfaces of arm, axle, dropout, etc. If the axle and arm are identical hardness, ductility, etc., then the damage will be slower, but if one of them is harder than the other, the softer one will get chewed up.
Anything that clamps down on a larger surface area of the axle flat will disallow this rocking.
If only one side's axle is secured, the rest of the motor core / axle will still rock back and forth, twisting the axle up to the secured point, and can eventually shear thru the axle itself once it has stressed the metal beyond it's limits.
That can happen even if *both* axles are secured if the forces on the axle are great enough, such as with some of the motorcycle-class builds where a lower power motor is used at way higher power levels than it was designed for. (I have an example of that, sent to me after this happened, in a QS205 run at tens of kW vs the three or so it was meant for).
What actually happens on any specific setup depends on the materials and forces in that setup, and the time involved. The longer it's used, the more likely a failure becomes, once material limits are exceeded.