With talk of the Switch successor being revealed this week I thought it'd be worth doing a review of all the major leaks and rumors that have appeared surrounding it. There's a comical amount of stuff to go over for a system that hasn't even been formally announced yet. Originally, I'd meant to do this at the end of last year but waiting might have paid off since it might only be a few days away from being unveiled to the world at long last. In fact, just today some new developments have happened that is even more smoke for a reveal being imminent. I'm not the most plugged into this topic or the technical aspects surrounding it so there may be some mistakes made but whatever. The broad strokes paint a pretty interesting picture.
The T239 Leak
Starting at the very beginning of the proper speculation for a Switch successor is the leak of the T239 chip which was acquired as collateral from Nvidia getting hacked. The T239 is a SoC that looks very much like it is designed to support a dedicated video gaming system and because of their existing partnership with the Switch people conclude that this is likely a new Nintendo console. Digital Foundry made this article covering details about the chip as well as some experimenting to try and approximate the power of the system as well as test it against some big gen eight games. I won't go too into the analysis as it's highly controversial amongst the spergs who argue with each other about Switch 2 leaks. The main takeaways are:
a custom built SoC made special for Nintendo; no more off the shelf Tegra X1
Ampere architecture
8x ARM A78C CPU (this wasn't in the original leak but was determined some time later)
1536 CUDA cores
128-bit/LPDDR5 memory interface
DLSS capable with tensor cores
ray tracing cores as well
maximum memory bandwidth of 102GB/s
a dedicated file decompression block
reportedly demo'd a build of Breath of the Wild with zero loading times and maybe also at 4K60FPS (the latter point is not as widely corroborated amongst journoids)
Overall, it's already looking like a substantial step-up from the og Switch in terms of performance and capability. As there really wasn't a generational leap between the Wii U and Switch this is an exciting prospect for Nintendo fans. It's also reason for optimism with regard to third party support as the test yielded decent performance for Death Stranding that is about equivalent to the PS4. At this point, however, there are still quite a few big question marks hanging over the system like RAM amount, the internal storage, the clock speeds, the process node it was fabbed on, the screen, the list goes on.
The Shipment Manifest Leaks
For a long while the T239 specs were all that there was to pore over for Switch 2 information. There were paltry nuggets found through datamining firmware and OS updates for the Switch that hinted at some form of compatibility with the successor but still nothing especially interesting. That is until the 
s and other assorted ResetEra rejects of Famiboards realized that they can look up customs data regarding international shipping through a number of different websites. I guess this is how people dig for info about upcoming smartphone models yet for whatever reason no one had thought to do the same for vidya consoles until now. Anyway, they set to work filtering results for data pertaining to Nintendo, Nvidia, and Nintendo's manufacturing partners in Vietnam and what do you know, they find some pretty interesting stuff. It looks like prototype models for a device with an identifying code that doesn't match any currently known Nintendo product.
Dimensions
These being shipping manifests some pretty comprehensive details about the physical dimensions of the system are discovered quickly.
Size of the Switch successor itself:
- 206mm x 115mm x 14mm (the original Switch is at 173mm x 102mm x 13.9mm)
The Joycons and dock are also a bit larger than their Switch 1 counterparts but I'm having trouble finding the exact measurements for them.
Peripherals
A built-in mic is discovered as well as details about magnets in the manifest that lead people to think that the new Joycons will attach magnetically to the Switch instead of via rails.
A giant fricking fan
Easily the most lolwtf find from the whole system.
Presumably this is going to go in the new dock for the system which raises... interesting possibilities for how hard they intend to push this thing in docked mode.
New Shipment Discoveries
Later on, after the investor's meeting where Nintendo first acknowledged the existence of a successor to the Nintendo Switch in May 2024 they would discover a new shipment with a few more details about the system.
RAM
Two 6 GB LPDDR5X chips are found in the data. This was a significant find for two reasons: it put to rest long held anxieties about 8 GB and hopes for 16 GB and it also puts the Switch 2 ahead of the Xbox Series S by 2 GB. It's also assumed that the Switch 2's OS will also be very lean and lightweight unlike what's used for Xbox. The Xbox Series S is seen as the best hope for the Switch 2 getting substantial third support since many big games will have already had to be scaled down in order to work on the S (and by extension be allowed to release on the Series X). As a result the Series S is what Nintendo fans are hopeful the Switch 2 will be comparable to.
Storage
A 256 GB UFS 3.1 storage chip is found. UFS 3.1 had long been speculated as the likely choice for storage based on what was popular in the smartphone space. It's a pretty fast storage medium and not too far off from what the PS5 and Xbox Series have.
This post is a much denser breakdown of the shipping finds if anyone is interested.
Possible Factory Leaks Begin
The September Leak
It's in September of 2024 that photos allegedly of the Switch 2 in various states start to appear on the Internet.
These photos end up being quite contentious. The perspective of the case sitting on the desk looks off to me and feels like it was a poor shop. For what it's worth people closely examining the motherboard find product codes that match that of one of the prototypes found in the shipping manifests.
The sole revelation from this batch of photos is the extra USB-C port on the top of the system as seen in this CAD render.
Lastly here's the front facing view of the complete system.
The December Leak
Guy on Xitter pops up claiming to have photos of a completed Switch 2 model. I believe that he goes on to torch his own credibility by claiming that he used AI manipulation on the photos to better protect his identity.
(The left photo is allegedly the new dock.)
The January Leak
Another leak from Chinese social media. This time of the Joycons. One aspect in particular makes it (in my opinion) the most fun photo leak by a country mile.
Notice the little square sensor near the middle of the Joycon? Looks a bit like a mouse sensor, doesn't it? My God, it looks like it even has glide pads...
The debate between the mouse sensor camp and IR sensor camp has been on-going. The OG Switch has IR sensors in its Joycons (for the like 12 games that use it lmao) so it's not impossible that is what these are. However, recent findings like mouse support being found in firmware updates for the Switch and shipping data indicating the presence of mouse soles would seem to support the mouse theory. Zany if true.
The Clockspeeds Leak
Finally, just today another big mystery about the specs seems to have at long last been solved. An established dataminer revealed on Famiboards what he believes to be the various clockspeeds that will be used in the Switch 2.
The user in question will come to Famiboards every once and a while to post about new discoveries made in the Switch's firmware and OS so it's assumed these numbers also came from that. The more interesting theory is that the Switch 2's SDK has leaked somewhere and that it's where he's actually getting this.
Some rough estimations for the TFLOPS of each power profile have been made and come out to 1.72 TFLOPS for portable and 3.09 TFLOPS for docked.
Fun visualization of the power difference.
The difference in CPU speeds has weirded a lot of people out. The Switch 1 keeps the same clockspeed for docked and portable and it's doubly weird that the Switch 2's docked mode has the lower clockspeed. It's lead some people to cast doubt on these being accurate numbers. We'll have to wait and see.
Lingering Questions
The Screen
Despite being listed in the shipment data nothing concrete has been found about the Switch 2's screen. People have made claims that they say is sourced from insider information but all of it has sounded like safe, educated guesses. They'll probably go with LCD screens again because it's cheaper (and so they can make tendies double dip with OLED models later down the line 
). A 7.91 inch screen is something specific that has been claimed. This would be a huge increase over the base Switch's 6.2-inch screen and even the OLED model's 7-inch screen. It is assumed by everyone that it will be a 1080p screen, up from the Switch's 720p screen.
The Process Node
The question of what process node the Switch 2's hardware will be made on has probably been the single most hotly debated question about the thing. Heated arguments over whether it will be 4nm (le good) or 8nm (LE BAD) have been raging since the T239 was discovered and it gets very neurodivergent. What the frick do these numbers mean, exactly? I dunno, I guess before 28nm nodes were a thing it was some kind of specific measurement in a circuit board but now it's just branding basically. What matters is that smaller = better and better process nodes allow for smaller form factor and better power efficiency. Both are very important in a handheld. It's been a near constant back and forth of people citing a somewhat reliable but also spotty guy on Xitter who claims it's for sure 8nm, followed by people saying that 8nm is obviously too old for something that finished in 2022 and that Nintendo would NEVER cheap out on hardware. Every time an alleged photo of the motherboard drops people rush to estimate the process node based on analysis of the photo and napkin math which then prompts other people to write essays explaining why trying to estimate the node size that way is fricking r-slurred and completely unreliable. The leaks of the clockspeeds has tempered this somewhat and it looks like 7nm might be what people gravitate towards as general consensus. Either way, we won't know until tech reviewers get their hands on the final product and do teardowns. Not like it matters at this point.
The Games
Not a fricking thing has been leaked about games for the Switch 2. In the wake of basically everything about the hardware being laid bare almost a full year in advance of Nintendo saying anything about the system this is a bit surprising. Just goes to show how airtight Nintendo are
Conclusion
That's basically everything worth mentioning. It's a little wild to finally be on the verge of getting something official about this system after all this time feeling like I already know quite a bit about it.
Jump in the discussion.
No email address required.
At this rate switch 3 is just going to be a gaming laptop
Jump in the discussion.
No email address required.
More options
Context