Trying to please AAA devs that are never happy is fruitless, I agree, but while turn based combat and also some new concepts can be implemented with little HW power, others require more, and even many frugal 1st and 3rd party devs will eventually start needing some more CPU power and main RAM, I'm not talking about graphics, but, for example, larger gaming worlds with higher level of interactivity, better AI, more items and NPCs. And I'm not even talking about now, but late 2023 or early 2024.

The flaw here is the actual reality of things is BOTW was a cross gen title built for the WiiU hardware level it still has more interactions than all other open world games even the likes of RDR2 which was built for more powerful hardware, we recently had CB2077 released and it's interactivity is nowhere near BOTW either and is smaller with worse AI despite having way more power than the WiiU to work with and taking 3 extra years to develop. Power is not what determines that at this point it's developer execution, polish and efficiency I'll bet you BOTW's sequel will also have more interactivity in it's world while being larger than many games that get built for PS5 and X1.

Late 2023/2024 would be 7-8 years for the Switch and the third-fourth year for PS5 and XSS that kind of makes the power point a bit well pointless because obviously at that point a successor would be ready for Switch as it would have done it's active years at that point.

RDR 2 the main quest is boring as hell. The problem is RDR heir the fail state mission of GTA V. The game needs more dynamic fail state missions. The world is great but the execution when you are on the main mission is scripted and boring. The live world and dynamic fall a cliff when you play the main path of history. 

I think  that at  this point, Nintendo is in a unique situation. the offset due to the WiiU makes this time around a bit different. I believe that nintendo can see that the next few years could be profitable with the switch, therefore it can take this opportunity to polish the next console instead of having to rush a new release. For this reason I think it could be possible the there will not be a new console from nintendo in about 3-4 years like @Wyrdness mentioned.

I think that the first quarter of next year could really tell us about the demand for the switch. If it somehow manages to keep pace with what it did in 2020, then I see that as a great sign that it will be able to be profitable for at least 4 years. These are the first three months of 2020:

1,331,008
1,314,715
2,508,862

I expected the BOTW example and I totally expect Ninty 1st party devs to use available power more efficiently than anybody else, but:

1) After Ninty HW designers themselves, they are the ones that know their HW best
2) Ninty 3rd party exclusive devs most probably come second in efficiency, but they'll run out of power before 1st parties
3) Further down 3rd party multiplats devs will have to sacrifice some efficiency for portability, even the most frugal and efficient ones
4) Deep optimisation is time consuming, small teams will need to accept compromises for lack of manpower
5) Eventually, 1st parties too will want more power

All the aforementioned stuff can influence, with different weights, Ninty's decisions on next gen schedule, even totally ignoring 3rd party AAA devs overblown needs.

BTW, late2023/2024 would be 6.5-7 years, not a short lifecycle as main console anyway, and after that NS will easily have at least another 3 years as last gen entry level offer, unless Ninty decide for an earlier death. But I don't expect a record long lifecycle, this gives competitors a chance of outselling it lifetme despite never reaching its peak. A chance, not a certainty, as HW power can extend POTENTIAL lifetime, but it's not by any means a guarantee of success.

That's kind of the point Nintendo designed their hardware to best suit themselves so they can support it better and that's the first of only two factors that's going to matter to them when deciding their schedule the other factor is how much of their library old and new they've released. To put it simply Nintendo is dancing to its own theme and no one else's and anyone who wants to join in has already accepted that's the theme that will play for the platform's run which is why power to them is not a problem because they already set out an ecosystem where that was never the focus so it mainly attracts and retains those not reliant on the power approach.

Nintendo doesn't need AAA titles because they have the best alternatives which is their first party the NS didn't get the likes of GTA or RDR but got BOTW and its sequel, it didn't get FFXV but got Xenoblade 2 and DE all the while it won't get FFXVI but Monolith already have another project on the way, it didn't get DMCV but got Astral Chain and Bayo 3 etc... All the while the AA titles have found either a preferred position on NS or treat it as a must for their projects. 

Switch really doesn't need replacement until 2023 at the soonest. Killing the goose that lays the golden eggs just to play graphics catchup with companies they can't outspend in that department would be about the dumbest thing Nintendo could do.

People don't buy the Switch for graphics or new AAA multiplats, so long as it gets a steady supply of quality first party content, it will continue to sell. 

That certainly worked well for Switch, so I can see it working well for its successor. March 2023 really is the absolute earliest though, any sooner would be a miscalculation.

I agree with @curl-6 and @trunkswd March is looking like a great strategy, I am actually kind of surprised that sony or Microsoft didn't try it out. The switch seems to be the console that sold the most units in its first 12 months.

The gameboy advanced sold 10 million and it was also released in march.

the 2023 seems even a bit too soon for me, if the switch manages to sell really well in 2021 (maybe more than 25 million) then it could be insane to stop it's momentum.