Come on. Let's get a serious discussion here.
Yes, you can still virtualize systems and run Rosetta to run Intel apps. However, as far as we know, virtualization will only be available for ARM, and Rosetta will not run 32-bit software. And we have PLENTY of it to run, especially old games.
First, Apple starting notifying folks that 32-bit macOS apps were being phased out 3-4 years ago. Developers knew that Carbon ( a decent chunk of 32-bit legacy apps) was a dead end long before that. 32-bit apps were dead whether Apple moved to ARM or not. Same way the 32-bit kernel booting went 5-7 years ago.
Second, Virtualization isn't only availabe for ARM. Apple has had hypervisor support in macOS since 2017
"...
April 17, 2017By
Veertu Team
Today, Veertu Labs is thrilled to pull back the curtain on three Anka toolsets, our latest endeavors geared ..."
Build on-premise macOS private cloud for iOS DevOps. macOS VM for iOS CI and iOS DevOps and real device cloud.
veertu.com
This has been deeply in wide scale, production instances for years.
https://www.macstadium.com/anka
macOS 11 is going to add "more viritualization" support, but the notion that this is brand new or only Apple Silicon only is on super thin ice. Apple Silicon for the Mac is going to probably going to add either features or more performance to IOMMU ( VT-d in Intel terms ) to the standard macOS hypervisior API. Kernel extensions that don't depend upon IOMMU mapping as Systems Extensions are being kicked out of the kernel on x86-64 also. They are going to be just as "dead" as 32-bit are. Nothing particularly due to the instruction set of the CPU at all. It is Apple not being permanently saddled with legacy baggage they don't want to do deal with in the future.
What is "new" is Apple punting on any options other than virtualization on Apple Silicon. Decent chance there is no "open" boot environment (i.e., No BIOS clone , no EFI , no UEFI ). That is it is straight iBoot ( which was present on T2 Macs to some small shim to boot macOS on internal or external APFS volume and that is it).
So, it's not that Apple is lying: the problem is that they are overly optimistic, thinking users won't miss x86 / Windows software. But they are really underestimating how some software might be wanted and/or necessary.
It isn't really a matter of whether there is some ( any) users that will miss. It is a matter of how many. Look hard enough can probably find someone who'd like to run Mac Paint or version one HyperCard.
Apple knows probably better than most end users how many folks download the Windows drivers for Apple hardware. The primary location for those are their servers so number of downloads isn't that hard to get. Also not hard to get how many folks download the macOS system updates and upgrades.
Underestimating? Probably not. Assigning some outsized proportion of importance to them? Again probably not.
2020 isn't 2005. The Mac user base is much larger now. Up around 100M versus something with one less digit 15 years ago. The inflow momentum from iPhone is also not even in the same zipcode as what flowed in from "iPod". For the x86 transition Apple was a "follower". For the ARM transition (even though Windows has had 2 (or more ) cracks at Windows on ARM there isn't much there). In this case, Apple is more so a leader into this zone. There is far larger base of iOS (on arm) software ecosystem than there is a Windows on ARM software ecosystem. There inertia on ARM is either iOS or Android . It is not Windows.
Another issue is: do you really think Nvidia,AMD and Intel will really sit down and wait for Apple to steal their share? Of course not. As we speak, they are probably moving towards faster processors and graphics cards. So, Apple's advantage (if any) might as well diminish on the long run.
Yeah but already said there aren't looking for "max performance". they are looking for perf per watt. They are
far , far more interested in moving "laptops" up into the "desktops" performance space than they are looking for some huge leap over current desktop performance. There is a slide from WWDC that made that quite explicit. ( not really guessing there. )
They aren't trying to win the biggest ever super-duper FLOPS race. They are just trying to make Macs "better" ( probably in the case design direction that Apple already has been driving too. e.g. one port wonder MacBook . )
For the laptops Apple most assuredly out to steal all of those other players shares. Period. And since they get the make the parts selection call for those laptops , they are extremely probably going to select their all own stuff in 2-3 years from now (if not sooner) .
There are some desktop corner cases I don't think Apple is going to drive too deep into because the unit volume is too low. There are some Thunderbolt external PCI-e enclosure cases that are the same way. GPUs on a removable 'card' Apple will probably leave for AMD (and Intel if they get their act together, don't burn bridges with Apple , and put in the work. ). The embedded ( soldered to the logic board) work is going to be harder to get, but probably possible in a couple of product instances.
To bring it back to this thread's topics though. in order to keep up with the Nvidia 30xx and 40xxx and 50xx Apple probably will need to lean on AMD and/or intel to cover that. The upper end of that range the Mac GPU card volume is way too lower for Apple to mess with their own stuff. Lots of work to do and not much pay off for the work.