I’m not claiming insider knowledge, just going by publicly documents on Apple’s architecture and operations.
Private Cloud Compute runs on Apple silicon servers derived from the same silicon used in devices, with custom secure OS infrastructure built around them. See Apple’s security documentation and reporting on PCC servers by finding the link somewhere in the comments of this thread.
As for their econ model, just see Cook’s methodical style with fast turnover to avoid stockpiling [1], i.e., the opposite of what is going on with AI servers where margins are cut at every step of the way. So sharing an assembly facility doesn’t imply chips are interchangeable, it may well just be assembly efficiency, which I guess is common.
I’m not claiming insider knowledge, just going by publicly documents on Apple’s architecture and operations.
Private Cloud Compute runs on Apple silicon servers derived from the same silicon used in devices, with custom secure OS infrastructure built around them. See Apple’s security documentation and reporting on PCC servers by finding the link somewhere in the comments of this thread.
As for their econ model, just see Cook’s methodical style with fast turnover to avoid stockpiling [1], i.e., the opposite of what is going on with AI servers where margins are cut at every step of the way. So sharing an assembly facility doesn’t imply chips are interchangeable, it may well just be assembly efficiency, which I guess is common.
[1] https://mondays.supernegotiate.com/post/inside-tim-cook-s-ma...