I think it's really hard and the industry over relies on third party controllers for almost everything, writing unified control take a large team lots of time.
Look at how VW is struggling with CARIAD.
We still dont really know whats under the hood of this thing.
It could be marketing hype and not all what it purports to be, or it could be really slick, low latency, and tuned better than anything.
Controller development requires a lot of testing. Take the silicon for example, typically the industry would look at something and see if it has already been done and deployed, and known bugs fixed, and then license the block (i.e. your USB port, your ethernet port, your bluetooth, your can bus, etc) to other companies for a royalty payment instead of "we must build our own so we own it". If you look at the layout of a lot of the chips, you will see a rectangle in the middle, corner, edge, etc and then something build around it, then that whole thing will turn into another rectangle in another generation and be put into some other project and be the "rectangle" within. The most reliable design is something that has been around for a long time, and they are only updated / replaced if they can't scale up, otherwise they will do incremental improvement.
For a complete redesign, you will be trading this stability for something you want. It could be because the older design won't scale up, it could be because older design is in conflict with something new, it could also mean your older components are obsoleted and no longer available. This is something that one company will have a hard time doing if they do not have the scale big enough for the volume, and a lot of smaller companies go out of business because the redesign failed and they have no money to recover from it (i.e. older products can't sell, new products can't sell without a stable redesign).
Maybe VW / BMW / Mercedes are big enough to do it all by themselves, but if Bosch already has a good deal and a good design to buy from, they would be wise to just buy it or license it as the foundation for their own design (they can also buy from Denso if they are not loyal to German, or use that to leverage the negotiation). Sure you have to add extra stuff to support a new valve control mechanism or a new electric motor power regulator, etc, but you can focus your development on those instead of spending a lot to reinvent the wheel.
If Honda and Mazda have to do everything themselves instead of buying from Denso and Aisin, they would be bankrupt a long time ago. Toyota is smart to spin them off and keep them separate so they can sell to others as well to keep the economy of scale bigger.
.
