• Hey everyone, staff have documented a list of banned content and subject matter that we feel are not consistent with site values, and don't make sense to host discussion of on Famiboards. This list (and the relevant reasoning per item) is viewable here.

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (Read the staff posts before commenting!)

I’d like to gently remind folks (or inform you, if you were unaware) that Switch itself was supposed to launch in Holiday 2016 and was pushed to March 2017 at almost exactly this time.

We just didn’t know that’s what was happening at the time, because it didn’t leak. It wasn’t until the gigaleak that we got the original timeline.

The reason for the delay was (probably) neither games nor hardware, but the Switch OS being behind schedule. So regardless of the nature of this delay (or even if you can call it that, see LiC's comment), I wouldn't let it give you a taste of the ol' Wii U. These things happen, and we only perceive them because of a leak. Not everything leaks which mean these things happen more often than we will ever know.
Zero chance the Switch 2 sells as badly as the WiiU, but depending on the circumstances it could easily end up a 3DS-esque flop. Doesn't inspire confidence that nintendo won't overprice the Switch 2 when they insist on still not dropping the price of the Switch.
 
Wow, +20 pages in one day.


So while we haven't heard from the Horse's mouth, the possible reasons why Nintendo could be delaying a late 2024 release to (hopefully) early 2025 release are:

1. Shippment issues due to war
2. November elections (lol)
3. Delaying so they can have more games at launch from 1st and/or 3rd?
4. Supposedly they are confident they can sell 15 million switches his year for whatever reason (and no threat from Sony and MS) and really want to ride it out and with little to no cross gen games on switch.
5. PS5 Pro rumored to be releasing this fall, and they don't want to go head to head with that.

IMO, if it really comes out in 2025 everywhere (versus say a few continents as Super Metal Dave thinks), I think it's #4 and maybe #1 (if it's really that bad).
3 seens to be the more likely option, to why Nintendo delayed Switch sucessor to 2025.
 
0
I have doubts Metroid Prime 4 will be targeted mainly for Switch 1 based on previous reporting.
We all had doubts. We know it isn't coming the first half of this year. If it isn't shown by summer. I would bet your buns it would be a cross gen game.
Is safe to say that Wonder was fairly big budget.

Plus, it definitely was a BIG title considering it sold over 10 millions already and was a 90+ release at Metacritic.
I mean I am not doubting that. I was thinking more big adventure game, like metroid. But like you said it is await and see thing.
 
0
It just seems like he knew something happening.

Any and all alleged insider will be like "ah, yeah, I already knew about this, took a while but I'm glad you commoners are finally up to date now" whenever any news break from any source that aren't themselves.
 
Any and all alleged insider will be like "ah, yeah, I already knew about this, took a while but I'm glad you commoners are finally up to date now" whenever any news break from any source that aren't themselves.
Haha! I knew 10 months ago that they would have done delay!
 
Zero chance the Switch 2 sells as badly as the WiiU, but depending on the circumstances it could easily end up a 3DS-esque flop. Doesn't inspire confidence that nintendo won't overprice the Switch 2 when they insist on still not dropping the price of the Switch.
Well the 3DS didn’t flop, one could argue that it underperformed but given the circumstances it did better than most thought. And we can’t just look at the overprice like it wasn’t several issues combined to have it be that way:
  • 3D that was old hat & had articles scaring away potential customers
  • Relying on inconsistent 3rd parties at launch
  • 1st party drought due to the WiiU
Nor do I really see how not dropping the price of the Switch correlates in confidence that Nintendo may overprice Redacted.
 
T239 is not a pre-existing chip like the TX1 was when the Switch began development. A 1:1 comparison here doesn't really work.

Does it really matter though?

Don’t people usually consider the architecture inside the chip in determining age of hardware and not “tape out date” or whether it’s custom or not?

Any knowledge we have of t239 comes from that Nvidia hack from early 2022 which shows what the DLSS Nvidia API files were targeting for the t239 SoC spec wise. We know it’s an Orin SoC. Those t239 file targets dated back to 2019.

We have been basing our expectations of what Drake is, based on architecture going way back to the Orin reveal.

Tegra X1 and its architecture was revealed Jan 2015. Orin and its architecture was revealed Nov 2019.

And regardless of any node change between 2019 and now, the basic architecture shouldn’t change, nor would the perception of hardware age. People consider Mariko Switchs to be have the same aging hardware as 2017 Switch’s for all intents and purposes.
 
Btw, did Brazil say that everything would be clear in March? Saw someone mention that on Twitter. If so, it would be weird to just have Nintendo send out a tweet saying, "Oh yeah, a successor to the Switch is in the works". How likely are we to see an actual teaser for the hardware?
No. I mentioned in the podcast that I'd hoped things would be clear by March, but that was based on prior info. As I've mentioned a couple times today, I've actually heard that they might say the thing exists in April, not March. But I have very little insight into reveal plans.
 
Last edited:
I’d like to gently remind folks (or inform you, if you were unaware) that Switch itself was supposed to launch in Holiday 2016 and was pushed to March 2017 at almost exactly this time.

We just didn’t know that’s what was happening at the time, because it didn’t leak. It wasn’t until the gigaleak that we got the original timeline.

The reason for the delay was (probably) neither games nor hardware, but the Switch OS being behind schedule. So regardless of the nature of this delay (or even if you can call it that, see LiC's comment), I wouldn't let it give you a taste of the ol' Wii U. These things happen, and we only perceive them because of a leak. Not everything leaks which mean these things happen more often than we will ever know.

Well, I'd argue things are a bit different now because the Switch was fairly revolutionary at the time, whereas everything we know of the Switch 2 points to an iteration of the same concept.

Gives me DS to 3DS vibes more than anything.
 
No. I mentioned in the podcast that I'd hoped things would be clear my March, but that was based on prior info. As I've mentioned a couple times today, I've actually heard that they might say the thing exists in April, not March. But I have very little insight into reveal plans.
😄😄😄 that would be crazy. I wonder if there are people at Nintendo laughing at us.
 
Does it really matter though?

Don’t people usually consider the architecture inside the chip in determining age of hardware and not “tape out date” or whether it’s custom or not?

Any knowledge we have of t239 comes from that Nvidia hack from early 2022 which shows what the DLSS Nvidia API files were targeting for the t239 SoC spec wise. We know it’s an Orin SoC. Those t239 file targets dated back to 2019.

We have been basing our expectations of what Drake is, based on architecture going way back to the Orin reveal.

Tegra X1 and its architecture was revealed Jan 2015. Orin and its architecture was revealed Nov 2019.

And regardless of any node change between 2019 and now, the basic architecture shouldn’t change, nor would the perception of hardware age. People consider Mariko Switchs to be have the same aging hardware as 2017 Switch’s for all intents and purposes.
ampere and lovelace don't have many architectural differences to be relevant for gaming

it's not an Orin SoC

there won't be a node change. that would require a redesign
 
I think the discussion about what constitutes a "launch window" is pretty pointless. If Nintendo is actually telling companies they're not going to be able to publish their games until 2025, there's no way that's them treating a bunch of third parties like second-class citizens not allowed to publish their games in an earlier part of a launch window.
Not the same thing. The Switch library is currently available to every Switch owner. No one had a Wii U so the ports were basically like new releases.

Mario Kart 8 won’t sell 60 million copies on Switch 2 for example but it did on Switch due to the Wii U’s failure.
As far as I see, in Switch's first calendar year Nintendo's Wii U ports were all of three games: Mario Kart 8, Pokkén, and Bayonetta 2. No game but the next Mario Kart is going to replace Mario Kart, but on the whole I think more significant third party releases will matter more than what's missing by way of Wii U ports.
When would third party devs be made aware of the node?
I don't know that they'd ever need to be? If they've got the specs and the tools, the fine details of how those things came to be aren't necessary for what they do.
 
0
No. I mentioned in the podcast that I'd hoped things would be clear by March, but that was based on prior info. As I've mentioned a couple times today, I've actually heard that they might say the thing exists in April, not March. But I have very little insight into reveal plans.
have you heard anything about prime 4 for the holidays or is that still a mystery box?
 
Ultimately, if Switch 2 delivers on the promise of quality Nintendo games with noticeable graphical/technical improvement from Switch 1, then I’ll be satisfied. I personally don’t really care if the technology under the hood is potentially “outdated” as long as it’s still a major leap from Switch 1.
 
Answer me this:

If Nintendo was so short stocked/panicked on software and had to delay everything to accommodate, then why didn't they hold onto Super Mario Wonder? Switch didn't need Wonder last year with TotK, Fire Emblem Engage, Pikmin 4, SMRPG, Pokemon, Xenoblade Future Redeemed, Mario Kart DLC, Advance Wars, etc? Could have used that for Switch 2 as a cross-platform game if they're so desperate.

Or better yet, why are they clumping/releasing most of their first party lineup in the first 3 months of 2024 when software is already anemic? Another Code, Mario VS Donkey Kong, Princess Peach Showtime, Splatoon 3 Side Order. Then to have to resort to a partner direct when there's only 2 remasters left on your plate, no original games for the rest of 2024 and with no release dates on said games, sounds ingenious.

It doesn't make sense, I don't buy these rumors.

The guess that the delay would be attributed to software was speculation made by one guy, not part of the coalition of sources suggesting knowledge of a delay.

I wouldn’t put any stock on that notion of delay reasoning.

If Nintendo did indeed notify 3rd party devs about a delay in the new hardware, they certainly wouldn’t tell them it’s because they are having 1st party development issues…
 
Does it really matter though?
Yes, using age as a sole determinant of obsolescence isn't helpful when T239 is a customized chip designed for a certain degree of longevity and support as a hybrid console. If you read my follow-up posts you'll see I'm well aware that the architecture is four years old. Please don't selectively quote one post from pages ago and make a lengthy response when I elaborated several times.

That and simply saying 'this architecture is four years old, therefore it's outdated' does not track, there should be an explanation as to what features or capabilities are missing that would render it obsolete. Considering the GPU is DX12 complaint and supports features like DLSS, it seems very current to me.
 
Last edited:
Ultimately, if Switch 2 delivers on the promise of quality Nintendo games with noticeable graphical/technical improvement from Switch 1, then I’ll be satisfied. I personally don’t really care if the technology under the hood is potentially “outdated” as long as it’s still a major leap from Switch 1.
I don't think the architecture is a big deal. Like what is out right now that can't be played on the Switch 2? I mean like feature wise? That technically have been the case since the Wii.

I think there will be a slow down in graphical advancement anyway... shit, we have a slowdown in gaming making.
 
0
You cut out this part of my reply:

And the Switch Pro report wasn't wrong. It existed and was cancelled. We know this from Nate and DF. Unless Nintendo can somehow undelay this thing, I don't see how so many sources can be wrong.
I thought the common consensus around here at this point was that it simply never existed. Though, I may be misremembering some details.
 
0
There will probably still be some level of confirmation between now and their FY briefing so they can speak about future direction and hardware with shareholders. Maybe that's still March. After that I imagine they'll keep it quiet until Fall and Winter, with the Summer direct being the last hurrah for Switch 1.
 
0
Yes, using age as a sole determinant of obsolescence isn't helpful when T239 is a customized chip designed for a certain degree of longevity and support as a hybrid console. If you read my follow-up posts you'll see I'm well aware that the architecture is four years old. Please don't selectively quote one post from pages ago and make a huge response when I elaborated several times.

That and simply saying 'this architecture is four years old, therefore it's outdated' does not track, there should be an explanation as to what features or capabilities are missing that would render it obsolete. Considering the GPU is DX12 complaint and supports features like DLSS, it seems very current to me.
I’m also interested to see how the machine learning aspect of the chipset will play out. Unlike the hardware, that should be nearly finalized outside of the clockspeeds, DLSS and it’s usage of tenor cores will continue to evolve.
 
0
I do feel like now there is this subtext of Nintendo DESERVES to have the Switch 2 fail as punishment which is kind silly
Can't blame them for feeling that way when this current situation is turning out to be a literal repeat of what happened with the Wii to Wii U transition- a situation that Furukawa repeatedly ensured they'd avoid.
 
0
Zero chance the Switch 2 sells as badly as the WiiU, but depending on the circumstances it could easily end up a 3DS-esque flop.
I mean it could absolutely crater. There is far more to a console's life than the contents of a leaked spec sheet. I don't think Nintendo is totally going to blow it, but we don't know the size, feature set, launch price, or launch game for the thing. Nintendo could launch this thing with Nobody 1-2 Switch and nothing else.

They won't, but if we're gonna be pessimistic doomers, let's at least explore all the unknowns. There are myriad ways for Nintendo to fuck this up.

Doesn't inspire confidence that nintendo won't overprice the Switch 2 when they insist on still not dropping the price of the Switch.
I'm not saying that Nintendo won't fuck up the pricing strategy, but I cannot understand a world where the Switch's pricing strategy is evidence of failure. We might not like the pricing strategy, but Nintendo was the only one of the big three to not raise their prices during this generation. They're about to be the best selling console of all time. And they're going to do it by selling well at the end of the generation.

It's not only the best selling console of all time, it's likely the most profitable. Sony and Microsoft are facing a situation where they can't succeed no matter how many units they sell, because they lose so much money per unit of hardware, that their software can't ever overcome the gap, so both are considering a multiplat/3rd party strategy to make up the difference.

The Lite is the cheapest console on the market, and adjusted for inflation it's cheaper than the DS at launch, or the even the 3DS's cut price. The base model is as cheap as the Series S, the premium model is the best selling. And in last year the console launched a full fat Zelda and Super Mario game - the same as the launch year of the console, so it's not like you can argue the software support hasn't been there keeping the value of the console high.

Maybe the pricing strategy for the next device pisses us all of. I remember a lot of folk being unhappy at the Switch's price point in the first place. But arguing that the Switch is evidence that Nintendo will price themselves to failure seems a little wild to me. The next thing isn't even announced, so not only do we not know what it's price strategy will look like, we still don't know Nintendo's pricing strategy for the Switch in its last year.

I'm not an investor and I could give three shits about Nintendo's objective business success, but the 3DS sold 75 million units and had a baller library. So I don't think a "3DS-esque flop" is anything to be upset about from my perspective
 
Well, I'd argue things are a bit different now because the Switch was fairly revolutionary at the time, whereas everything we know of the Switch 2 points to an iteration of the same concept.

Gives me DS to 3DS vibes more than anything.
I mean, shitloads of things are different. But a 4 month delay of the launch does not indicate "Nintendo about to fuck up" it just indicates a delay. The kind of thing that seems to happen with regularity, we just don't hear about it a year in advance.
 
So I'm reading the history of the rumors/leaks of the Switch 2 in the OP, and have to say it's extremely fascinating watching all of these small details come together over the years to form a larger picture.

But a few questions -

1. There were reports of dev kits being sent out for a "4K" capable Switch all the way back in 2021 and 2022. Was this for T239? We would know if there was some other chip that was developed for those dev kits right?

2. The illegal Nvidia hack mentioned Samsung as the semiconductor foundry company for T239 (which I wasn't aware of). So....that seems like pretty damning evidence that this is not being planned for a TSMC node. What changed recently to make people think this is now on TSMC 4N?
 
1. There were reports of dev kits being sent out for a "4K" capable Switch all the way back in 2021 and 2022. Was this for T239? We would know if there was some other chip that was developed for those dev kits right?
If this even happened in the first place, it was for T214/Mariko. The so-called "Switch Pro" - again if it existed - directly evolved (devolved?) into the Switch OLED.
 
What are the chances Nintendo gave up on the T239 and is using something else, either from Nvidia or another company (and thus had to delay the launch)?
 
So I'm reading the history of the rumors/leaks of the Switch 2 in the OP, and have to say it's extremely fascinating watching all of these small details come together over the years to form a larger picture.

But a few questions -

1. There were reports of dev kits being sent out for a "4K" capable Switch all the way back in 2021 and 2022. Was this for T239? We would know if there was some other chip that was developed for those dev kits right?

2. The illegal Nvidia hack mentioned Samsung as the semiconductor foundry company for T239 (which I wasn't aware of). So....that seems like pretty damning evidence that this is not being planned for a TSMC node. What changed recently to make people think this is now on TSMC 4N?
1. The Pro (if it did exist) was likely what we know as Aula, except clocked higher. Tegra239 is too distinct from the TegraX1 for the seemless backcourt expected from a Pro console (alongside of its hardware being finalized/taped out far too late).
2. Thraktor's prior analysis on power draw (alongside some other factors like the Front Level Clock Gating power saving tech backported from Lovelace, discovering engineers that worked both on Drake/Lovelace) lead most to entertain the possibility that Samsung being labeled as a manufacturer was nothing more than a holdover from Orin. As it stands it seems to be either all ot nothing with either Samsung 8N vs TSMC 4N, with how little murmurs we've heard of Nvidia maybe allocating Samsung's 5nm nodes on the cheap.
 
0
If Nintendo pop out with MH Wilds, GTA 6, 3D Mario and PROFESSOR LAYTON as launch games it's over for every other console fr fr
 
I was hoping that Switch 2 would launch before or around the holidays since I would expect third party holiday/annual releases like Call of Duty to be released alongside it. But with the delay, I wonder how many will launch later on Switch 2 or not launch at all and wait for the next holiday season.
 
The fact that this point is from SMD64............. Are we this desperate for reassurance?
If you have to ask...

What are the chances Nintendo gave up on the T239 and is using something else, either from Nvidia or another company (and thus had to delay the launch)?
I asked something similar and oldpuck gave a pretty detailed response below how the chances of that are less than slim to nil:

Not 100% sure what you're asking, but chip development like this, based on what we've seen from AMD and Nvidia, takes 3 years. So if you wanted to replace T239 with something for a 2025 release, you would have needed to start in 2022.

But let's pretend for a second that development is instantaneous, that I can package up a SOC from any existing technologies right this very second. The differences between the chip you'd make today right this very second and the leaked T239 specs would be little more than a footnote. And even those differences might not be good as they'd represent potentially increased costs for very little (or even no) improvements. Let's write them out even, so we can see what we're talking about.

We could replace T239's modified Ampere with stock Ada. This is almost no change at all, the architectures are so similar. You could get the larger Optical Flow Accelerator, but because of limitations of being a handheld, Frame Generation is not likely to ever work. The larger OFA might be a literal waste of space, so this could effectively be a downgrade.

You could update from LPDDR5 to LPDDR5X. It's possible this has already happened. 5X would decrease battery life for the sake of extra bandwidth. Benchmarking suggests that extra bandwidth might be much of a performance win. This is a shrug.

ARM processors have actually been progressing nicely. You could replace the A78C cores with A715s. This would break Switch 1 backwards compat for what ARM advertises as a whopping 10% increase in performance. You could use A710s instead. This would keep BC, and ARM advertises the same 10% performance increase, but benchmarks suggest the opposite with an actual drop in performance.

You could put your chip on TSMC N3. Going by existing products on that process node, Nintendo could probably do that and launch this thing at $750. It's a non-starter.

The technology to make T239 significantly better does not yet exist and is unlikely to till late 2027, at which point it will still be prohibitively expensive. That's just the reality on the ground.



No one is making games with PS6 specs in mind while they're still desperate to figure out how to make games with PS5 specs profitable.

Specs matter. Features matter more. Switch wasn't much more powerful than the 360, but games didn't look like the 360, and they certainly didn't only come from the 360 era. Why? Because Maxwell (the GPU arch) and the rest of the console (solid state storage, the RAM pool, the CPU design) were all much more modern.

If AMD delivers a truly innovative GPU design that causes Xbox Series 2 to completely alter how DirectX works, the Microsoft will push that version of DirectX into Windows, which will force Intel and Nvidia to redesign their GPU pipelines to be DirectX14 (or whatever) compatible, which will alter how PC software works, which will start to leave T239 behind.

Tellingly, Microsoft seems to be driving the updates and forcing AMD to keep up. And what MS is pushing are things like RT and ML, forcing AMD to keep up with Nvidia's design. Right now it looks like T239's feature set will have a long life.


Truth

No clue. Custom SOCs for mass market devices that last more than a couple years are pretty rare, and we only have knowledge on T239's tapeout because the Lapsus hack gave us data we could cross-check elsewhere. That was an unprecedented event.
Looking back, I'm not entirely sure how I meant to phrase my question either, but I do feel this explanation satisfies my curiosity either way, so thank you and @ILikeFeet earlier for your responses
 
gonna assume the q1 2025 stuff has weight to it because of how doom and gloom the thread is now. i stop checking fami for like 2 days...
 
0
Please read this staff post before posting.

Furthermore, according to this follow-up post, all off-topic chat will be moderated.
Last edited:


Back
Top Bottom