Note that it is not just the capacitors that must handle the higher battery voltage.
It is also the FETs, and the low-voltage-power supply (LVPS), which converts battery voltage to the 15/12/5/etc volts that are used within the brain, sensors, and gate drive electronics of the system.
If there is a display on the EV, it usually has it's own internal LVPS, which also must handle that higher voltage, and it's own capacitors, and often it's own FET (or other transistor) that is used to switch battery voltage on and off to the controller's LVPS by the display's power button.
Any part that does not handle that higher voltage may fail, either immediately or over time from overheating or other degradation.
One example of a catastrophic chain failure that could happen (even though it's probably pretty rare):
Many cheap controllers use a large resistor to drop the battery voltage into the LVPS, and this resistor is typically already sized to barely handle the heat it has to dissipate. When overvolted, it will run hotter, and may also allow higher voltage to the LVPS input than the LVPS can handle (and if it fails at a lower-than-original resistance, which sometimes happens instead of failing open), the LVPS will definitely see higher voltage than it should, and then the LVPS itself may fail.
If the LVPS is designed in a common way of using parts in the path that can fail shorted (or partially shorted) internally, then higher voltage than is tolerable by parts powered by it can be passed thru to those parts, and then *they* can also fail. This can damage not only the controller MCU, gate drivers, etc., but also motor hall sensors, PAS sensors, throttles, or anything else powered by the LVPS.
At that point so many things are damaged that it may require an entire new system (except the battery) to repair the EV.
An SMPS-based LVPS is probalby not designed in the same way and probably wouldn't let the same chain failure happen, but it's not a guarantee without knowing the design.