I'm a Professional engineer, but I do not design lenses. I do know a little about how they work, and as lot about how electronic and mechanical design works.
Image Stabilization is accomplished by moving a lens element or group in order to keep the projected image stable on the camera sensor. This means that motion sensing elements in the lens must detect any tiny movement and acceleration, send a signal to the voice coil type motor that moves that element just the right amount to offset movement in the lens. If the signal isn't just perfect, the movement will overshoot, and the image will jump, so a feedback circuit determines when the lens element is moving too much, and a correction is introduced. If the gain of the circuit doing this is set too high, the IS lens group will just keep moving, so it has to damp out and stop very quickly.
In order to add a additional stop of IS, the motion detector must be more sensitive, the motor must be more powerful, and the feedback control must have more gain. This gets difficult and expensive, its not a linear function, a additional stop might up the cost 100 times or more.
I suspect that the sensors are already pushing the limits of sensitivity for low cost commercial sensors, a larger motor needs to be bigger and takes more power, and the feedback circuits fall in the same category as the sensor.
I am certain that 5 stop IS is possible, but the price might be breathtaking and out of proportion to the value of the lens. The designer has to balance all the factors and give the customer the best bang for his buck. Adding $1000(or whatever the cost) for 1 stop better IS does not deliver enough value to the buyer for his expense.