The power is the same (actually opposite sign) on both sides only in the ideal stationary case. In the ideal case there is no extra power needed to change the kinetic energy as the speeds are constant. In the non ideal case the power on both side does not have to be the same. If needed to overcome friction or to accelerate the vehicle the power on both sides can be different (e.g. more power going in to accelecate, or more power coming out to slow down). The system has plenty of power (force exchanged through the vehicles times paltform speed) - so there is no real arguing about not enough power to increase the kinetic energy. The vehicle could move and still produce extra power, e.g. for lights.
---------------------------
To calculate the stationary speed we don't even need to look at forces or power, just look at the velocities. The calculation was shown in a similar way before and is still correct:
The gear ratio (k) connects the speeds of the two wheels and thus the speed of the vehicle relative to the 2 platforms (V1 and V2). So as an equation V2 = k * V1. The 2 plattforms move relative to each other at a speed V0. So V2 = V1 + V0.
Do the math with the 2 equations, and one gets V1 = V0 / (k-1). This gives a valid solution for all gear ratios except k = 1. So the 1:1 gear ratio does not work as we have seen before.
With a suitable value for the gear ratio you can have V1 at nearly any speed you want: both directions ( sign of V1) and also faster or slower than the relative morement of the platforms. In a reall life situation very high speed ratios may not work well due to friction, but gear rations of -1, 0.5, 1.5 or 3 are not a problem. This would give you half and twice the relative speed of the platforms in both directions.
The equations show the the vehicle could move at the calculated speed to avoid slip or tearing the vehicle appart. The platforms will provide the necessary forces and power to make the vehicle move this way, as this is the path of least restance (only friction).
If you remember I posted here a motor connected to a generator and it had also a light bulb (free energy generator or something like that was the text describing that) And what you say about the vehicle is exactly the same thing.
Obviously if vehicle is stationary in ideal case the kinetic energy will remain zero as vehicle is not moving relative to the ground (red box).
And also yes the power will not be equal at G and M in real world but power at M will always be lower than power at G meaning the vehicle can only move from right to left so the Kinetic energy vehicle will have will be in a certain direction.
We are talking about power available at the wheel not motor or generator power as in a motor or engine powered vehicle.
Say transmission efficiency from G wheel to M wheel is 90%
And say breaking power is 100W say that is 10N at 10m/s then say 90W of this is available at M wheel (due to 10% loss) then you can select any gear ratio you want and say you want to increase force to 20N then speed will be 90W/20N = 4.5m/s
So vehicle will move backwards (right to left) and is only because of the power difference and there is no need to know the forces and speeds as they are irrelevant to know what direction the vehicle will move.
So in this particular example if you store energy for 1 second and then apply that stored energy for another second you can see best what happens.
first you will store the 100W from the generator and while doing so the vehicle moved backwards 10m since M wheel is not powered in this first second.
Then the stored energy 100Ws is used to M wheel but transfer and motor efficiency is just 90% so 90Ws are available.
The vehicle has higher force available but travels slower so with the energy it will be able to travel back 9m thus vehicle is now 1m to the left compared to when it started.
So if you ignore the storage that was just to understand the vehicle moves from right to left at 1m/s speed.
You can calculate this much easier just from the power delta.
As mentioned many time before. If you are not convinced by my correct calculations you just need to test. And I promise if you (any of you) can demonstrate vehicle moving from left to right I will pay you back for the experiment.