If it saved up enough power to enable start up, yes no problem. (but according to their own graphs they charge when the power is switched on)
If its supposed to reduce the amount of power used by the motor (by any apreciable amount), snake oil comes to mind, its just not possible. The savings would be miniscule and the capacitors would be long dead before you recouped the costs.
A capacitor can smooth out the power, but there is no way that it can dramatically reduce the power consumed by a motor (smoothing the power can obviously cause a small gain and I do mean tiny)
Power factor (the power "that seems" to be required but not actually used) is always a strange one to explain Its the amount of power thats needed to be available for a motor to work, but its not the actual amount power consumed.
Reactive loads such as inductors (motors) and capacitors dissipate zero power, yet they drop the voltage and draw current giving the deceptive impression that they actually do dissipate that amount of power.
This "phantom power" is called reactive power, and it is measured in a unit called Volt-Amps-Reactive (VAR), rather than ACTUAL watts.
Reactive power is symbolised with the letter Q (look at the chart mentioned in the original link) the only column with any obvious difference is the Q column or VAR, Then you have apparent power Symbolised by the letter S this is volts and amps without regard for the Phase angle, but even this isnt the TRUE power, true power is a combination of the two, or a measurement of volts and amps accurate, you do not get billed on VAR, only on the P or watts (true power) column is your billing amount note the difference there....
So although the power company give you 110 Watts because you only borrow the last 10 and give it back you only get billed for 100 Watts.