I think Microsoft's approach is to build a "series" of consoles with regular hardware updates.

The Playstation 5 and Xbox Series X released on TSMC's "7nm" process.
We are pushing 5nm in the GPU space now and TSMC's 3nm is starting to pick up traction.

Microsoft and Sony also pushed out clockrates a little harder which tends to have a corresponding increase to power consumption... A more balanced chip would likely remove much of that.

I think in a year or two, a PRO console is more than justifiable... Even with technology of today we can already expect a 50% performance improvement on most aspects.

As for the architecture/compatibility issue...

Microsoft went from Graphics Core Next 1.0 to Graphics Core Next 4.0 with the Xbox One > One X and never broke backwards compatibility... And they were chalk-and-cheese in terms of core design.
Sony and Microsoft also went from GCN to RDNA 2 which is a larger leap than RDNA 2 to RDNA 3 and never broke backwards compatability.

Microsoft and Sony are employing monolithic OS's, drivers and kernels that smooth over discrepancies in hardware which makes it all a non-issue... Their focus is to be able to move everything forwards with new hardware.

I don't think I'd have to wait that long to gather that Microsoft made a serious miscalculation. Actually the bolded pretty much shows why it's indeed misfired. MS thought this generation will play out like X1/PS4 vs X1X/PS4 Pro where the low powered/cheap SKU sees higher demand or sales. They kept the Series S a secret until not long before launch as if to checkmate Sony.

It seems that by design, the Series X was positioned to be their limited-quantity/enthusiast SKU that's just there to assist the Series S, their primary SKU. This would not pan out well in a long generation. Series S has very little room for effective long term pricedrops compared to PS5 and Series X (see Wii vs PS3), and it's underpowered right off the bat, meaning it will age rather poorly especially if a midgen upgrade is planned. When/if midgen upgrades are launched, a lot of AAA developers will push for fidelity high enough for base (by then cheap) PS5 and Series X to struggle in achieving 1080p+/40+ fps (You wanna go higher? Get the PS5 Pro, or the Xbox Series XL, or a capable PC!), Series S won't be able to adequately handle games with such high workload. We're not yet feeling the S limitations because we're still stuck in the crossgen period, the vast majority of console-grade games are souped up/upgraded Xbox 1 games.

If Series X meets demand and continues to sell less than the S, I'd blame it on Microsoft's marketing and people giving up and switching to PS5 or PC. As always, I may be wrong and Series S does manage to appeal to a new large dempgraphic, but so far I'm not feeling it and I don't think it'll happen.

The Series X and PS5 aren't priced as "hardcore gamer" consoles. With inflation in mind, they're priced in line with typical powerful consoles. I don't know why you're assuming demand will decrease over time in favor of the Series S. The only notably underpowered home consoles I can think of are the Wii and Wii U, one managed to find huge success by appealing to a new dempgraphic early in its lifecycle, but ultimately having weak legs because it aged poorly (I predicted that), and the other was a commercial disaster. It's less that PS5/Series X are "hardcore", and more that Series S is too cheap/underpowered. The existing console playerbase have no problem paying up to $500 ($400 for digital editions) for a powerful console, the Series S is trying to solve an issue that doesn't exist.

I agree that the Series S has little room for effective long-term price drops.  What I see as the flaw in your analysis is that Microsoft is predicting that the PS5 and Series X *also* have little room for effective long-term price drops.  Typically consoles are sold at a loss up-front, but with cost reductions so substantial that they can later be sold at a profit despite significant reductions in MSRP.  What Microsoft said in the Eurogamer interview that I've been referring to, is that their road map doesn't show substantial cost savings due to moving to smaller processes.  They will likely realize some cost savings over time, but not as much as in the PS4/X1 era, which itself was not as much as the PS360 era.  Therefore, Microsoft and Sony are likely going to have to choose between selling them at a profit, or significant MSRP reductions, they likely won't enjoy both.  They could maybe split the difference (sell them at cost at moderate MSRP reductions).  But that significantly changes the rest of your analysis, IMO.

You also misquote me.  At no point did I assume that Series X demand will decrease over time in favour of the Series S.  I indicated that Microsoft suggested that, not that I was saying it was necessarily going to be so.

I agree with Microsoft's naming convention suggesting a series of consoles with updates.  The question is at what cadence, and whether they designed to be "pro" consoles of what came before, or a new generation each time (or perhaps whether they'll even alternate between those extremes).  The span in between, and what past console iterations that developers are required to support, will tell the

As for pro consoles being justified, it's a question of two things IMO: is there a market for it (as there was for 4K-focused mid-gen refreshes last-gen), and can it be done economically.  Those 5 and 3 nm processes you speak of, Microsoft has signalled that they'll reduce heat and size, but not reduce cost as those processes are expected to cost enough more that the increase in yields will at best compensate.  Time will tell if Microsoft is proven correct, but at this point they're not anticipating big cost reductions in Series X|S and PS5.  That makes a "pro" console that can be offered at a cost that drives market acceptance hard to come up with, if that prediction holds true. 

As for those upgrades not breaking backwards compatibility, that was done in part through emulation of elements of past platforms.  Even Xbox One on Xbox Series has an emulation layer, from what I understand.  So Microsoft either needs a fully backwards compatible architecture, or a big enough power bump each time to emulate as necessary.

I don't think I'd have to wait that long to gather that Microsoft made a serious miscalculation. Actually the bolded pretty much shows why it's indeed misfired. MS thought this generation will play out like X1/PS4 vs X1X/PS4 Pro where the low powered/cheap SKU sees higher demand or sales. They kept the Series S a secret until not long before launch as if to checkmate Sony.

It seems that by design, the Series X was positioned to be their limited-quantity/enthusiast SKU that's just there to assist the Series S, their primary SKU. This would not pan out well in a long generation. Series S has very little room for effective long term pricedrops compared to PS5 and Series X (see Wii vs PS3), and it's underpowered right off the bat, meaning it will age rather poorly especially if a midgen upgrade is planned. When/if midgen upgrades are launched, a lot of AAA developers will push for fidelity high enough for base (by then cheap) PS5 and Series X to struggle in achieving 1080p+/40+ fps (You wanna go higher? Get the PS5 Pro, or the Xbox Series XL, or a capable PC!), Series S won't be able to adequately handle games with such high workload. We're not yet feeling the S limitations because we're still stuck in the crossgen period, the vast majority of console-grade games are souped up/upgraded Xbox 1 games.

If Series X meets demand and continues to sell less than the S, I'd blame it on Microsoft's marketing and people giving up and switching to PS5 or PC. As always, I may be wrong and Series S does manage to appeal to a new large dempgraphic, but so far I'm not feeling it and I don't think it'll happen.

The Series X and PS5 aren't priced as "hardcore gamer" consoles. With inflation in mind, they're priced in line with typical powerful consoles. I don't know why you're assuming demand will decrease over time in favor of the Series S. The only notably underpowered home consoles I can think of are the Wii and Wii U, one managed to find huge success by appealing to a new dempgraphic early in its lifecycle, but ultimately having weak legs because it aged poorly (I predicted that), and the other was a commercial disaster. It's less that PS5/Series X are "hardcore", and more that Series S is too cheap/underpowered. The existing console playerbase have no problem paying up to $500 ($400 for digital editions) for a powerful console, the Series S is trying to solve an issue that doesn't exist.

I doubt Sony would go for 2 separate models, they want to keep things simple for developers.

I doubt Sony would go for 2 separate models, they want to keep things simple for developers. If you compare the installed base of current gen, 30 million PS5, 15 million Series S and 6 million Series X (using roughly 70/30 distribution from UK data), it is clear where to developer focus is going to be. In general, it is the optimization of software that yields best gaming performance.