Chip shortage say no, I'd assume that the two year roadmap was to give enough lead time to resolve a few outstanding issue:
Apple's 2 year window was probably composed years before the pandemic ever showed up. It is probably largely independent of the current chip shortage ( which Apple probably didn't anticipate the depth and breadth of the impact. ) The chip shortage probably will have Apple hand waving as to why Dec 2022 is still "2 years" later. Decent chance the end of their transition slides into 2023 in pragmatic terms (i.e., shipping in volume).
If Apple had previously attached the MBP 14" and 16" SoCs to the A15 iteration they plan was very highly likely laid down before the pandemic even surfaced. Which would be indicative that it was going to take more than 12 months to get the Macs transitioned. If more than 12 then might as well give themselves 2 calendar years wiggle room.
If Monday's SoC are just larger grouping of A14/M1 cores with a better internal interconnect then the pandemic probably played a minor role there timeshifting the roll out.
1. RAM limitations of AppleSOC, as we have seen thus far all M1 Mac's have on die memory. While I'm sure Apple could use normal slotted DDR RAM, what would that do performance wise.
On package memory; not "on die". Apple's die is mounted on a substrate that the RAM module packages are also mounted to the same substrate. Apple is focused more on Performance/watt than myopically on Performance (at any cost ). For the laptops (and 24" iMac and probably at least one Mini variant ) they are also focused on minimal z-height.
Slotted RAM isn't going to necessarily bring more performance (given the specific features Apple has weaved into their SoC. ). It is more capacity (e.g., getting to 200-500 GB of RAM). Also helps with more effective use of 3D space (volume). For triple/quad digit GBs of RAM also probably should add ECC also which Apple isn't covering at the moment either.
Slotted RAM would require memory controllers that consume more power. If you incrementally heat up the Apple dies that probably won't be a good outcome in performance. The performance/watt is 'baked' into the other components on the die. Apple isn't out to build the best user overclocking die implementation possible.
Depends upon where there MBP 16" SoC lands in performance with what the larger screen iMac would require. If put two of those together then I'm not sure Apple is really going to be pressed about doing a "desktop exclusive" SoC.
If Apple can cover the ground with laptop oriented SoC designs then they probably will stretch that over as many Mac products as they can.
Depends upon where the iGPUs land.
2. Also the number of PCI-E lanes needed for PCI-E and Thounderbolt4. Unless Apple wants to go with a Mac mini Pro( Cube 3.0 ). I highly doubt that as they sold the 2013 Mac Pro for six years, and had to admit it was a failure.
Apple's "war" on discrete GPUs is probably going to delay something getting out for the upper "half" of the desktop line up. If they end up removing all dGPUs from the entire laptop line up then they'll need some different die(s) for provisioning a decent amount of PCI-e lanes. That will take more time.
I don't think Apple thought the MP 2013 was a total failure that some customers want to label it with. I doubt they will repeat it exactly. But a "Mac Pro" with reduced modularity is definitely possible. Especially, if they are chopping the volume of the system.