Contracts like that don't make any real promises. The vendor agrees to deliver product A under conditions X or product B under conditions Y, but it doesn't promise it will be able to deliver A. If the customer relies on getting A, it's obviously taking a significant risk.
Yes, they make promises. Like how many cores, expected performance, TDP, pin-compatibility, etc. things that are important to vendors in order for them to be able to have systems ready to be manufactured and ready to ship as soon as possible after the new chips are available. It also avoids them the hassle of prototyping a system with LPDDR4 or Thunderbolt 4 when those aren’t even on Intel’s roadmap!
No one is signing anything with Intel that states, “Give us money up front to reserve a bunch of the first run of chips. When will we have them? I dunno, don’t want to make any promises. Before 3021? Nah, like I said, we don’t want to make promises. We’ll do something or another, more than likely but don’t hold us to it! Thanks for the money though”
Your arguments seem to imply that stupidity and incompetence are the key hiring criteria at Apple.
Of course Apple is making many prototypes it will never sell. That does not require any significant effort even at a company 1000x smaller than Apple. If most of your R&D efforts don't fail, you are not particularly ambitious and your products will most likely be mediocre and uninnovative.
Prototypes based on the narrow range of expectations that Intel and their internal architects have defined? Absolutely. Which curve at which angle, do the fans need to be a half of a millimeter shorter, how about those pads on the bottom, rubbery enough? Can we model the batteries to fit more capacity? Where does that leave us for total weight?
No one was prototyping for something that exceeded projected TDP. They WOULD have if Intel had told them up front that they would likely exceed the projected TDP by a staggeringly impressive amount. But, as we saw, that’s not what Intel did.
