As I see it there are effectively four categories of games when it comes to patches/support for the new hardware:
1. Games releasing at the same time or after the new device. Obviously we should expect these to all support it, typically with higher resolution output, and possibly also leveraging the new hardware more heavily, such as higher quality assets, RT/DLSS, etc.
2. Games that are still receiving patches/updates/DLC, such as Mario Kart 8, Splatoon 3, Animal Crossing, etc. If they're already shipping patches for these games, then it's relatively easy for them to add support for increased resolution on the new hardware. I don't expect much more than resolution bumps for these titles, though. Not even DLSS, just straight higher rendering resolution. Perhaps 60fps on one or two titles at best.
3. A handful of older titles which will be used to show off the new hardware. The above two categories won't cover very many games at launch, so to sell the hardware, I expect Nintendo to patch a small number of older titles with higher rendering resolutions. Basically just a few "prestige" titles that they can show off in 4K to sell the new hardware. I would expect games like Mario Odyssey, Luigi's Mansion 3, Link's Awakening, etc. to be in this group. Again, nothing more than a rendering resolution bump here, they're not going to go back and add something like ray tracing to a codebase nobody's touched in three or four years. There is one possible exception for DLSS that I'll mention below, but I think these will be the simplest updates possible. I also expect them to all arrive at the launch of the new hardware. They'll be there to pad out the number of games that leverage the new device during the launch period, then as new games are released there's little reason for Nintendo to go back and patch any more of them.
4. Older games which won't receive patches. This is most games released prior to 2023, as it's not worth digging up old code to add patches for anything but the small number of prestige games in category 3. This would include everything from older Pokemon games to the likes of ARMS or Captain Toad, or anything developed by a third party, like Astral Chain. Basically the majority of Nintendo games on Switch. I would expect in many of these cases, where a game either has an unstable framerate or a dynamic resolution, running on the new hardware should give us more stable framerates and higher resolutions, but still within the confines of the original code.
The question really is how many games Nintendo decides to include in category 3. I don't think it's zero, because (a) Nintendo needs to flesh out the line-up for the new device and (b) they have a lot of evergreen titles on their roster, so it's not like they're patching games which won't continue to sell. My guess is single digits, probably five or six games.
Three of these potential games that I'm quite interested in (as I'm playing through the trilogy right now) are the Xenoblade games. XBC3 will probably be in category 2, still receiving some content by the time the new hardware arrives, but XBC: DE and XBC2 could count as prestige games in terms of showing off the new hardware. XBC2, in particular, would probably benefit more than any other game from a patch to run on Drake. Not just in straightforward resolution increase, but because the Xenoblade games use temporal upscaling/AA solutions, it's
theoretically possible for them to be adapted to use DLSS without major changes to the code. The word
theoretically is important, as there could be many technical reasons it wouldn't work, but it's not like a game which doesn't use TAA, where the game engine just doesn't generate the necessary inputs to DLSS.
The reason I'm interested in DLSS for XBC2 isn't just resolution (although it would certainly help), but image quality. The particular implementation of temporal upscaling Monolith used for XBC2 seems to have a very aggressive sharpening filter being applied to it, which means that when the internal resolution drops, the upscaling heavily emphasises the aliasing, rather than removing it. This makes the resolution drops far more noticeable than they would be otherwise. The benefit of DLSS is that it would replace this upscaling and do completely the opposite; provide a clean, anti-aliased image regardless of the internal resolution. DLSS is effectively both an upscaling and anti-aliasing solution, which is why you often hear people saying it looks better than native, because it's trained on super-sampled images.
So, combining both an increased rendering resolution due to Drake, with a big jump in output resolution due to DLSS, and the infinitely better image quality that DLSS would provide over the current solution, I think XBC2 would probably be not only the biggest possible upgrade from base Switch to Drake, but also likely the biggest increase in image quality even from one generation to the next, let alone from something like the PS4 Pro or Xbox One X. In purely IQ terms, it would be like jumping from the Xbox 360 to the Series X. From a point of view of getting people interested in the new hardware, just showing some before/after screens of XBC2 using DLSS on Drake would be a hell of a sales pitch.
That said, XBC2 is an old game by this point. They apparently
added Korean language support in late 2020, but the main content updates stopped in 2018. It's quite possible that whoever worked on the upscaling solution back then isn't at the company any more, and even if they are, they may not remember all the pertinent details, like data formats for motion vectors, etc. Diving back into old code can be a pain even when you're familiar with it, so adding DLSS to XBC2 would almost certainly take a lot more time than doing the same thing with XBC3. I feel that it's only likely to happen if it's something that's pushed for within Monolith, ie if a few employees who worked on XBC2 asked management to allow them to work on updating it for the new hardware, or perhaps doing so for all three XBC games on Switch.
Personally, if someone asked me to go back and work on code I wrote 5 years ago that's barely been touched since then, I'd try to avoid it if at all possible, but if the code I wrote was for Xenoblade Chronicles 2 I'd probably have a very different attitude. Fingers crossed that some folks in Monolith were playing around with Drake dev kits when XBC3 development was finishing up, and thought that going back to XBC2 (and XBC: DE) was something they wanted to do with their time.