Normally this sort of issue is solved with a capacitor network. It's actually common to get them removed during a "cost optimization" stage because the device will still appear to work. Just with grately reduced reliability.

Any capacitor network will need some power to stay charged waiting for use. If peaks are rare enough, they won't be a reasonable choice.

The root of the issue is that the internal resistance of the battery is too high to supply current under peak load to the device. Anyone creating an electronic device has to perform these calculations and design a power supply that is sufficient. Its basic electrical engineering.

For extremely short duration peaks up to a few microseconds capacitors are a great workaround. They are necessary for pretty much every digital integrated circuit. A billion transistors all switching at the same time on the rising edge of a clock cycle has a massive instantaneous current draw.

> Anyone creating an electronic device has to perform these calculations and design a power supply that is sufficient.

Yes. And it does look like a case of undersized batteries. If I had to guess, I'd say it was because of too strict product design specifications meeting a lower than expected quality in manufacturing.

CPU power peaks usually last for a (large) fraction of a millisecond or more, because that's how the OS scheduler works. Added to that, peaks may not happen for entire minutes or hours. Both of those work against the usefulness of large capacitor banks.