- ARM is not Apple IP, it's owned by SoftBank and licensed to Apple at rates low enough that it's impossible to undercut them.
- TSMC's 5nm manufacturing capacity was entirely bought-out for Apple Silicon, blocking other OEMs from competing on equal footing.
- The SOC team, who arguably did the most innovation of all, was gutted with the founding of Nuvia immediately after Apple Silicon's launch.
They look good on specific workflows because they are focused on specific accelerators and code path optimisation that they know is more beneficial to the way most of their customers use their hardware.
But their lead is hardly uniform and there are still plenty of things where they are barely doing better than intel, if at all. And their GPUs that are supposedly good don't perform that well outside of benchmarks, efficient, surely but not exactly something to rave about.
When you look at it from a high vantage point, there have been a lot of compromises to be made to get mostly efficiency/battery life, the transition wasn't cheap and some stuff will forever stay stuck.
With the progress of Intel/AMD it doesn't feel like that big of a win. And this is also why ARM isn't working so well on the Windows side. So much to give up for not that significant gains.
At least we know it was mostly about architecture optimisation and access to better lithography processes. Eventually they will all converge to similar performances and we will regret the code compatibility. Some big software houses (like Autodesk) basically got vindicated for not building software for macOS, it's not a really good sentiment for the Mac long term.
It's not clear if any of it matters because clearly, they would rather sell iPads with a software distribution monopoly in the form of their App Store. Maybe with EU ruling they will change their mind, but I doubt it, as long as they can maintain their marketing...