There are 2 sides of every story. Tesla's "software defined" architecture can ..... lead to ECU failure because they screw up their logging and wears out their flash chips (not enough volume to test them, unlike reusing the same boring architecture for 20 years within Toyota, for example).
In large scale design and production, volume means test data and stability. Toyota wait and use proven designs so they are relatively boring and reliable. Anyone can do it, but not many "cutting edge" companies like BMW, Mercedes, etc like to do it. Tesla being way out there means they can fix a lot of their own problems with software updates, but that also means they may (and they did) cut a lot of corners and ship half baked products (they have to because they never have the same volume). Plus, you need more chips and you will get more supply shortage than your competitors with fewer chips if managed correctly.
Denso probably sell the same ECU architecture to every Toyota, Nissan, Honda, etc. You don't need too many unique spare if you flash the right firmware to it right before installation. Of course, that also means if shortage hits they run out real fast (that's the other side of the same coin).
BTW, it is easier to stuff a bunch of excessive transistors into an expensive car vs a cheap one. Just because Tesla probably put in $1000 worth of chips in every car doesn't mean it is a good idea to put in $1000 worth of chips in a Yaris. Porsche can even afford to buy a brand new Miele washer to take it apart, and scavenge a chip to finish the car, and leave the washer in a warehouse and repair it when the chip shortage is over. You can't afford that on a Golf.