Yeah, I think, like I wrote, Apple’s decisions appear to be “Marketing-Message Driven,” and from at least since November 10, 2020, when the first Apple Silicon based Mac was unveiled, the Marketing Message was that Apple Silicon (CPU and GPU) was the future and Intel or X86 was the past. And Apple’s M1 did “beat the pants off” anything Intel or AMD had out there in terms of performance, thermals and energy use. I think at the time there were PCs that beat M1 Macs, but these PCs were large, power hungry desktops with thermal issues that caused fans to spin up in no time, resulting in loud fan noise from CPU cooling as well as the fans onboard AMD and Nvidia graphics cards.
...
Very suddenly, Nvidia and its proprietary CUDA technology and CUDA API has been the go-to technology undergirding most if not all of the rapidly advancing A.I. models that seem to be making major news on a practically daily basis.
Does Apple want to yield to Nvidia or stick with Apple M[#] GPU cores for graphics, GPGPU and A.I.?
I think (as I think you do) that Apple will stick with Apple Silicon for high end Macs (Mac Pro) no matter what.