Rosetta 2 expects certain capabilities from ARM CPU, which exists solely for the purpose of running x86 simulation. Apple will drop that in future versions of their SoCs and Rosetta will go as well.
Based on decades of historical precedent. Apple has migrated from 68k to PowerPC to Intel to ARM. You can no longer run PowerPC applications on current MacOS, as Rosetta v1 was dropped five years after migrating to Intel. Likewise 68k emulation was dropped some time after transitioning to PowerPC. Rosetta v2 is another temporary solution until x86 support is completely gone a few years from now, unless they intend to continue offering the Game Porting Toolkit.
Except that those situations were materially different from the current situation in two huge ways:
1. The architecture they're migrating away from is the dominant architecture in the market (both now and for the past few decades), meaning that rather than simply losing support for older MacOS software, they'd be losing access to easy virtualization of everything Windows and Linux can do.
2. The architecture they're migrating to is their own, 100% in their control, meaning that they can maintain hardware-level support for this translation layer without needing to convince any other company to put that effort in on each subsequent chip generation.
Now, does this mean they're guaranteed to maintain Rosetta 2 forever? No, of course not. I don't know what Apple's going to do any more than you or viktorcode do.
But it does mean that seeing this transition as being exactly a mirror of the past 2 architecture transitions they did is dangerous, at best.
Agreed that this is a different situation. If they drop Rosetta then Crossover is gone and Game Porting Toolkit is finished. Also Docker will be limited without x86. I'm using that to run SQL Server on Linux on MacOS for work.
This is an unprecedented situation given their new push into Mac gaming, seemingly trying to convince AAA studios that there is a market. If successful they can increase their user base and steal some thunder from Steam, and it probably bodes well for their Vision plans.
I'd say they have an incentive to not drop Rosetta until they reach a critical nexus of support from the studios... enough that it will cause the laggards to quickly add support once it's taken away. Maybe that will be there in a couple years, but likely it will take a bit longer.
The historical precedent of dropping ARMv6 support from Apple SoCs, for instance. No matter what the secrecy level of a corporation is, no one will waste their CPU area for backwards compatibility when it's no longer demanded by the majority of the user base. It runs agains the economics of hardware production.
