> You'll have to explain why a dedicated hardware company wouldn't be able to come up with such a chip given sufficient competition. If many people use language X, then it makes sense for an IC design company to optimise for that language (unless there is a near-monopoly situation and they don't care about what the market wants).

They gave you an example. Apple (software part) don't need to run around and ask hardware companies to implement optimisations in their hardware, and hope that implementation will be as promised (see: Microsoft and Qualcomm, and debacle called "custom ARM chip for Windows"). Apple (software part) can plan their releases many years in advance, knowing that Apple (hardware part) will fulfil their promise (see: Intel's roadmap recently, NVidia and Apple in the past). Apple (software part) can choose programming language that suits them and be sure that the hardware will follow. Apple (hardware part) can plan many years in advance and make obviously superior product knowing that there will be enough software support (Rosetta 2) to overcome initial platform switching pain. The whole M1 Mac story is all about the company that is so integrated that they can pull something that nobody else can.

You choose to ignore the example.