• 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.
  • Do you have audio editing experience and want to help out with the Famiboards Discussion Club Podcast? If so, we're looking for help and would love to have you on the team! Just let us know in the Podcast Thread if you are interested!

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

It's not about making a big announcement and show for consumers at gamescom it's about the opportunity it provides for nintendo to show off development tools for the new console to devs.

Would be extremely weird for nintendo to skip an important place to show off those development tools to devs for a console that's been announced, so it's safe to assume it won't be announced this summer.
If they're demonstrating development tools to devs at Gamescom 2024, rather than some form of devkits and documentation already being in developers hands at that point, then something has gone badly wrong. Not being present at Gamescom is completely meaningless to Switch 2 plans, because Directs have completely displaced games conferences for this kind of presentation.
Late 2025 onward and this system does start to look very outdated.

A cutting edge 2026 system would have three generations (!) of NVIDIA architectural improvements and LPDDR6 and mobile OLED screens seem to be getting cheaper as well.

Late 2025 would just come off as a massive fuckup compared to just designing a system that launched in 2026. Missing FP4 in particular I could see being a really big deal for the Switch 2 as DLSS is fairly expensive and ray reconstruction is extremely expensive for its core count and I wouldn't be shocked at all to see a faster DLSS being made using FP4.
You were never going to get this fantasy "cutting edge" system. They're mass manufacturing hardware that requires months/years of advance planning in combination with drawing a line under a target spec so that they have a schedule of 1st and 3rd party software to roll out in the first 12+ months, while also juggling the need to keep Switch 1 ticking over until that time.

There will never be an adequate cut-off point for creating a market-feasible future-proof system offering RT and ML-driven rendering techniques for 10 years, probably more. It is a new rendering paradigm that will form the next graphical arms race in place of rasterization, which is now largely "solved" in that fidelity is now largely driven by budgets, rather than raster performance. Every potential cut-off point will be "outdated" in some way because of this. That's what the PC market is there to address, not fixed spec console hardware.

Nintendo are already releasing into the most favourable conditions they could hope for: the baseline target for multiplatform development has already been set, and it's 1st-gen AMD RT hardware.
I am curious as to why we haven't seen some more pointed questions to DF. For example, how do you square 8nm with a 1536 core GPU in a portable system that will target 3+ hours of battery life? They also speak of performance being less than Steam Deck, but again, how do you square that with a 1536 core GPU? They speak of T239, seem to agree that its the SOC to power SNG, but then somehow gloss over the fact that even with low clock speeds, its going to surpass Steam Deck by a decent margin, at least in docked performance. It seems like they are failing to connect a lot of the dots here and are not calling into question certain "accepted truths" that they seem to share. 8nm is not just big, but very power hungry, something that Nvidia would have known from the beginning. From the power curves we have seen for Orin, it seems impossible for a 8nm T239 to reach low enough power draw for it to meet the requirements from Nintendo. As soon as you look at T239 with the possibility its actually a 4N chip, suddenly the size and power draw situation falls right into place and the ceiling for performance goes way up.

Im going to guess that they really do not have any inside information that gives them any clarity on these details. I suppose as long as you remain convinced its 8nm, some of those opinions seem obvious, but even then they dont really do a good job when talking about performance. Its important when comparing it to something like Steam Deck, are we referring to portable or docked performance for SNG? Could Steam Deck be more powerful than SNG in portable mode? Sure, but docked mode its basically impossible.
I'm sure if you laid it out to them as such DF would probably agree with you - but there is little incentive for them to wade further into the discussion ahead of the hardware details being confirmed. The only thanks they'll get, even if they're later proven right, is the most feral console warriors on earth dismissing any positive claims about Nintendo hardware performance.
 
microsoft also doesnt really need to do a redesign because their series x cooling solution is shit simple but super effective, chip directly to vaporchamber directly to large mass heatsink, with every other heat producing component is connected to a solid metal skeleton. its crazy how much metal is in a series x actually
A good rule of thumb for US bulk shipping from east asian assemblers is $3-$5 per kilogram of weight. Let's assume MS and Sony both get the cheapest possible version, $3. That means the base Xbox Series X, and the original PS5 hardware cost $13.50 to ship to the States.

Heat sink materials are usually a non-ferrous aluminum copper alloy. Copper is expensive, $4-5 bucks a pound at the moment, but aluminum is cheap, something like $0.80 a pound. Let's use that as a proxy.

The PS5 redesign got the weight down to 7lbs. Let's call half of that metal. That's a $1.20 in raw materials saved, $4.05 in shipping costs saved. $4.25 saved total. Let's assume that PS5 underperforms the PS4, and just barely breaks 100 million units.

This redesign made Sony nearly a quarter of a billion dollars over the course of the PS5 generation. If MS isn't going for a similar redesign it says to me (an admitted casual observer) that they simply aren't selling enough Series consoles to justify the investment.
 
Like, come on, this chip being on 2020 architecture and being finalized early 2022 caused a lot of people in this thread to think this was releasing mid to late 2023.

Will be a little disappointing if this launches late 2025
 
A good rule of thumb for US bulk shipping from east asian assemblers is $3-$5 per kilogram of weight. Let's assume MS and Sony both get the cheapest possible version, $3. That means the base Xbox Series X, and the original PS5 hardware cost $13.50 to ship to the States.

Heat sink materials are usually a non-ferrous aluminum copper alloy. Copper is expensive, $4-5 bucks a pound at the moment, but aluminum is cheap, something like $0.80 a pound. Let's use that as a proxy.

The PS5 redesign got the weight down to 7lbs. Let's call half of that metal. That's a $1.20 in raw materials saved, $4.05 in shipping costs saved. $4.25 saved total. Let's assume that PS5 underperforms the PS4, and just barely breaks 100 million units.

This redesign made Sony nearly a quarter of a billion dollars over the course of the PS5 generation. If MS isn't going for a similar redesign it says to me (an admitted casual observer) that they simply aren't selling enough Series consoles to justify the investment.
A few things on the matter of the Xbox family. Xbox Series X has a smaller retail box than even the slimmest PS5 (Digital Edition Slim) and has already had revisions to shrink the heat sink since launch. While they haven't taken advantage of a die shrink yet, they're definitely keeping the weight and materials question in mind. It also helps they have the smallest, lightest console on the market making up a majority of their growth. The Xbox Series X cooling system is also very small, especially next to PS5, but even next to the Slim model family, so in a way it's difficult for them to shave off more than they already have.

As for redesigns and more Xbox consoles, beyond the leaked roadmap from the court documents, they've officially announced a hardware roadmap, including the next generation system. They specified they'd have more to show this coming holiday. It seems likely to me that portions of the leaked roadmap, including the die shrunk revisions of Xbox Series X|S and the new controller (which, despite being not officially revealed, already has three different names, classic Microsoft branding.) are true, not only did they say they'd have something to sjare this coming holiday, but when Tom Warren off-handedly asked about new console and controller "options" coming this Holiday, Phil Spencer didn't deny anything.

While more anecdotal and not necessarily related, Xbox accessories, particularly controllers and Expansion Cards, have been steeply discounted in some markets, namely the US, which would line up with the leaked timeline for controller/accessory refreshes starting in late May. While I think this has probably been pushed back, they are clearing stock and it's almost certainly coming.

Assuming the roadmap indeed holds water, at least, some, Xbox has a LOT of hardware to be launching over the next few years, not including their next generation system, with refreshes to the Xbox Wireless Controller and Elite, a new Universal controller to slot in between them, a one-handed/new accessibility controller, a Joy-Con-style first party mobile controller, and the two console revisions. Notably, the handheld was listed as low priority in this leak, so I wouldn't expect it for another two years at minimum. I wouldn't be surprised to see the Keystone/Cloud device make it out between now and then.

Still, between the leaks, the interviews and the roadmaps, they do seem to a lot of refreshes to get through, and honestly I'm not sure all of it will pan out. At the very least we know they planned to go pretty hard on the Xbox ecosystem over the next few years.
 
Wait...why are suddenly people drawing the conclusion that the lack of a Gamescom pressence = delayed to late 2025? What?

And DF made more speculation and they're taking their thoughts as gospel?

There’s just a lot of nothing happening for a system that has to be announced very soon to realistically release before April 2025.

Getting a little anxious
 
Wait...why are suddenly people drawing the conclusion that the lack of a Gamescom pressence = delayed to late 2025? What?

And DF made more speculation and they're taking their thoughts as gospel?
I saw someone mention it on reddit, that it's most likely that Nintendo would prefer doing their own showcase of the switch 2, compare to showing it and renting booths in gamescom.

Since having a fullblown presentation seems more likely
 
I saw someone mention it on reddit, that it's most likely that Nintendo would prefer doing their own showcase of the switch 2, compare to showing it and renting booths in gamescom.

Since having a fullblown presentation seems more likely
Right now I am hoping for an early May "reveal" (at least the name or concept) and a summer presentation.
 
There’s just a lot of nothing happening for a system that has to be announced very soon to realistically release before April 2025.

Getting a little anxious
I mean...nothing happening has been a staple during console generation transitions...

But this amount of doomposting just because of their lack of presence in Gamescom is rather unreasonable.
 
And DF made more speculation and they're taking their thoughts as gospel?
I‘m the one who said that (I think?), as I searched for anyone else who phrased it that way, but that’s just the way I’ve seen some loose comments or answers being transformed into something that’s confirmed or definitive at times, hence the term; gospel :p.
I think the latest speculation brought forward by DF in their latest direct, just created a minor repeat of discussions about expectations. Not much more than that imo.

Getting a little anxious

There’s no need to be anxious, because in the end you will receive a new generation of Nintendo hardware eventually. Yes, I wish it was sooner too, but if speculations or the anticipation about what the hardware is going to give you anxiety. It really might be time to step back from the conversation for a while. Just my 2cents.
Because it’s really not that serious imo.

Nintendo was one of the many large Publishers that were responsible for the death of E3 by refusing to go, claiming that it was far too much trouble and expensive to set up a booth. I can possibly believe that Gamescom might be cheaper than E3, but if so, hardly by much.

Hmm ok, I uhh think it’s way too simple to boil it down to; not going or boycotting, but let’s just leave it at that, because this isn’t really the thread for it.
 
Last edited:
0
I mean...nothing happening has been a staple during console generation transitions...

But this amount of doomposting just because of their lack of presence in Gamescom is rather unreasonable.
Here's my thing. This is for the people who are too eager for the next gen. You guys really need to work on yourself if it is really bothering you that much that we get little to no information on the NGS. What happens in the future when games take longer to make? We are in territory of where games can be delayed up to 10 years. Also, Nintendo is making sure no one yaps. You can't blame them for that. This isn't a Nintendo problem it is a you problem. Fix it and this isn't toward you SiG.
 
Last edited:
Gamescom is and has been quite a good conference for gaming.
Yes it may not have the same legacy as E3, but Nintendo has been at Gamescom for many years. I have no idea where you’re coming from lol.

Edit; compare Nintendo presence E3 2017 v.s. Gamescom 2017.


Oh haha I wasn’t implying that you were saying NG > PS5, I was just speaking generally.
Moreover iirc, with respect to RT. AMD’s hardware also require more memory compared to NVIDIA for RT, so even if the switch has more memory available than the Series S, due to NVIDIA’s superior RT implementation it also requires less memory and bandwidth.

edit; for the latter statement I’ll try to find the exact analysis of where I got that from.

Nintendo was one of the many large Publishers that were responsible for the death of E3 by refusing to go, claiming that it was far too much trouble and expensive to set up a booth. I can possibly believe that Gamescom might be cheaper than E3, but if so, hardly by much.

So to see them boycott E3 but have no problem prior to this year going to Gamescom was a slap in the face to gamers everywhere.
 
It's already been stated before, but Nintendo skipping conferences this year could not only be signal for a probably slow year software wise, but also because they want to arrange those expenses into something else, like a possible Switch 2 event.
 
Nintendo was one of the many large Publishers that were responsible for the death of E3 by refusing to go, claiming that it was far too much trouble and expensive to set up a booth. I can possibly believe that Gamescom might be cheaper than E3, but if so, hardly by much.

So to see them boycott E3 but have no problem prior to this year going to Gamescom was a slap in the face to gamers everywhere.
tenor.gif
 
Nintendo was one of the many large Publishers that were responsible for the death of E3 by refusing to go, claiming that it was far too much trouble and expensive to set up a booth. I can possibly believe that Gamescom might be cheaper than E3, but if so, hardly by much.

So to see them boycott E3 but have no problem prior to this year going to Gamescom was a slap in the face to gamers everywhere.
Wasn't it Sony who started the Domino affect of E3 Dying or was it Nintendo?
 
Wasn't it Sony who started the Domino affect of E3 Dying or was it Nintendo?
Nintendo stopped doing physical conferences in 2013, Sony pulled out altogether in 2019, and the pandemic was the real opening of the floodgates. E3 would have likely died off slowly regardless, but like many things in life, the pandemic made publishers question why they even needed a physical show in the first place.
 
Wasn't it Sony who started the Domino affect of E3 Dying or was it Nintendo?
Both, kinda? Sony was the one who said "Wait, we do not need this" and stopped attending the show, but Nintendo was the one who decided livestreaming pre-recorded conferences was the way to go long before then.
 
Both, kinda? Sony was the one who said "Wait, we do not need this" and stopped attending the show, but Nintendo was the one who decided livestreaming pre-recorded conferences was the way to go long before then.
probably because of this E3 Demo. Like the switch 2 presentation should be live though
 
I don’t know if spec finalization needs to be done in 2021 for a system releasing in late 2025 if that happens, lol.

The Xbox Series X uses architecture from 2020.
The difference with the PS5 and Xbox Series is that they're essentially midrange PCs. AMD just asks them for a performance target, and they can whip up whatever combination of CPU and GPU power is necessary based on their existing APU architectures. Because Nvidia only makes huge car and robot SoCs anymore, they had to completely Ship-of-Theseus Orin into Drake, which also requires cooperation with Arm because Nvidia doesn't make its own CPUs, and they have to hit Nintendo's very strict battery life and size requirements, and they also have to get vendors for screens, all of the stuff in Joy-Cons, etc. That takes a lot more in R&D than a PlayStation or Xbox.
 
All of Nintendo's studios have put so much work into their own internal engines that it would be odd to see any of them ditching what they currently have in favor of Unreal.
And Nintendo is too big of a video game company to not use internally developed game engines. No nead to pay fees for using externa engines.
 
Here's my thing. This is for the people who are too eager for the next gen. You guys really need to work on yourself if it is really bothering you that much that we get little to no information on the NGS. What happens in the future when games take longer to make?
I play other games. But preferably not with a 2017 frame rate and resolution.
So to see them boycott E3 but have no problem prior to this year going to Gamescom was a slap in the face to gamers everywhere.
Maybe that subset of gamers who ran the ESA, but I'm not sure why anyone else would take it as a slight.
 
Late 2025 onward and this system does start to look very outdated.

A cutting edge 2026 system would have three generations (!) of NVIDIA architectural improvements and LPDDR6 and mobile OLED screens seem to be getting cheaper as well.

Late 2025 would just come off as a massive fuckup compared to just designing a system that launched in 2026. Missing FP4 in particular I could see being a really big deal for the Switch 2 as DLSS is fairly expensive and ray reconstruction is extremely expensive for its core count and I wouldn't be shocked at all to see a faster DLSS being made using FP4.
I understand what others are saying and why the strategy you're asking for may be a fantasy... But I have to agree with you there. I already thought early 2025 was pushing it for a chip packing custom 2022 tech, but in the absolute nightmare scenario this thing is indeed a late 2025 launch... It'll be missing (nearly) three generations of improvements and overall future-proofness since Nvidia's machine learning is constantly evolving and leaving old architectures in the dust, Ampere already can't use frame gen and dual issue as it is... DLSS4 and beyond starts seeming like a huge ask real quick. In comparison, the TX1 was only one generation behind (Maxwell 2.0>Pascal).
 
Last edited:
I understand what others are saying and why the strategy you're asking for may be a fantasy... But I have to agree with you there. I already thought early 2025 was pushing it for a chip packing custom 2022 tech, but in the absolute nightmare scenario this thing is indeed a late 2025 launch... It'll be missing (nearly) three generations of improvements and overall future-proofness since Nvidia's machine learning is constantly evolving and leaving old architectures in the dust, Ampere already can't use frame gen and dual issue as it is... DLSS4 and beyond starts seeming like a huge ask real quick. In comparison, the TX1 was only one generation behind (Maxwell 2.0>Pascal).
A hypothetical AD10F Drake likely wouldn't have used Frame Generation anyways. Nvidia only recommends it above 60 or more fps native, since the input lag and frame smearing at 60 fps would be a much worse experience than just rendering it at 30. That's why the PS5 and Xbox Series X don't use Fluid Motion Frames despite being more than capable architecturally. It's a feature designed for discrete graphics cards, not integrated console GPUs.
 
A hypothetical AD10F Drake likely wouldn't have used Frame Generation anyways. Nvidia only recommends it above 60 or more fps native, since the input lag and frame smearing at 60 fps would be a much worse experience than just rendering it at 30. That's why the PS5 and Xbox Series X don't use Fluid Motion Frames despite being more than capable architecturally. It's a feature designed for discrete graphics cards, not integrated console GPUs.
I'm aware, just used it as an example of how much a single architecture can shake up things. Now imagine two and a half, full on three if the 8nm doomerism has some truth to it? That's no longer futureproof in the slightest, to be honest.
 
I'm aware, just used it as an example of how much a single architecture can shake up things. Now imagine two and a half? That's no longer futureproof in the slightest, to be honest.
"Future-proof" is relative. The PS4 and Xbox One have garbage CPUs even by 2013 standards, but games are still getting made for those systems. I think it's reasonable to expect the Switch 2 to function as a low target profile for the rest of the 9th generation and well into the 10th.
 
"Future-proof" is relative. The PS4 and Xbox One have garbage CPUs even by 2013 standards, but games are still getting made for those systems. I think it's reasonable to expect the Switch 2 to function as a low target profile for the rest of the 9th generation and well into the 10th.
The CPUs were indeed, but even then they weren't outdated, it's the best AMD could offer at the time. The GPUs were as reasonably modern as possible... Not directly comparable but again, it's somehow uncomfortable to entertain such a possibility at the moment. Early 2025 must be the time... Please lord let it be.
 
Last edited:
With the Switch 2 delay making the Switch even more longeval, does anyone think that Nintendo could try to do the same with the Switch 3? The Switch is doing A-OK for a console that's pretty much in its eight year in the market
 
Late 2025 onward and this system does start to look very outdated.

A cutting edge 2026 system would have three generations (!) of NVIDIA architectural improvements and LPDDR6 and mobile OLED screens seem to be getting cheaper as well.

Late 2025 would just come off as a massive fuckup compared to just designing a system that launched in 2026. Missing FP4 in particular I could see being a really big deal for the Switch 2 as DLSS is fairly expensive and ray reconstruction is extremely expensive for its core count and I wouldn't be shocked at all to see a faster DLSS being made using FP4.
So anytime after it releases it will still be seen as outdated past the first few months it's released? Idk
 
0
With the Switch 2 delay making the Switch even more longeval, does anyone think that Nintendo could try to do the same with the Switch 3? The Switch is doing A-OK for a console that's pretty much in its eight year in the market
Technology is slowing down, so I think this is a very reasonable and perhaps even likely outcome. The system after next is well into the 2030s. That said, "Nintendo as a platform" could be more of a thing, and we could see things like the Tegra X1+ revision equivalents actually having a little more power available, with the launch model's handheld mode serving as the "base" of compatibility for the generation.
 
Nintendo stopped doing physical conferences in 2013, Sony pulled out altogether in 2019, and the pandemic was the real opening of the floodgates. E3 would have likely died off slowly regardless, but like many things in life, the pandemic made publishers question why they even needed a physical show in the first place.
Nintendo were fully present at the show floor for all the years that was relevant. Contrary to popular belief, the part that they "pulled out of" back in the Wii U/3DS era isn't even properly part of the show, and they replaced it with other live events that better took advantage of the format like Treehouse Live and tournaments.
 
Nvidia could definitely offer something better by then (late 2025). It might need an expensive new tapeout however which Nintendo might not want to do even with an 18-24 months headstart. You're all right hence "I understand", it's what it is.
what could nvidia possibly offer thats better than what nintendos going with atm?
 
Nvidia could definitely offer something better by then (late 2025). It might need an expensive new tapeout however which Nintendo might not want to do even with an 18-24 months headstart. You're all right hence "I understand", it's what it is.
what is that better supposed to be? we don't know if they're even moving to 3nm for Blackwell or if it'll still be a 5nm design. ARM hasn't improved their CPUs by a significant amount on the same node (though @Dakhil shown they made some improvements in power per mm2). there's LPDDR6, I guess. but we need to wait and see how it actually performs first
 
Nvidia could definitely offer something better by then (late 2025). It might need an expensive new tapeout however which Nintendo might not want to do even with an 18-24 months headstart. You're all right hence "I understand", it's what it is.
An expensive new tapeout being necessary kinda goes to show that "technically, they could whip up something better" isn't all that meaningful. Nintendo isn't clairvoyant, there will always be a risk of unforeseen circumstances. (And that's assuming that late 2025 is set in stone in the first place, when we're not at the stage where we can determine that.)
 
what is that better supposed to be? we don't know if they're even moving to 3nm for Blackwell or if it'll still be a 5nm design. ARM hasn't improved their CPUs by a significant amount on the same node (though @Dakhil shown they made some improvements in power per mm2). there's LPDDR6, I guess. but we need to wait and see how it actually performs first
How much of a performance jump is it from 4 nm to 3? Just a random question
 
what is that better supposed to be? we don't know if they're even moving to 3nm for Blackwell or if it'll still be a 5nm design. ARM hasn't improved their CPUs by a significant amount on the same node (though @Dakhil shown they made some improvements in power per mm2). there's LPDDR6, I guess. but we need to wait and see how it actually performs first
LPDDR6, FP4 (dual issue) and better future-proofing for DLSS features to boot. We still don't know ARM's cpu schedule for 2025, so that can't be verified at the moment. A proper 2025-ish handheld would be an improvement by architecture alone, that's just how far Ampere is expected to fall below the curve at the current pace... But again, late 2025 isn't exactly the most reassuring date to speculate about.
 
what could nvidia possibly offer thats better than what nintendos going with atm?
There's the Blackwell architecture, which Nvidia's using for Thor.

(Of course, Nvidia hasn't revealed much about the Blackwell architecture. And there's a possibility Nvidia makes datacentre Blackwell architecturally distinct from consumer Blackwell, similar to the Ampere architecture.)
 
LPDDR6, FP4 (dual issue) and better future-proofing for DLSS features to boot.
Not everything is about checking off boxes on a spec sheet. You still need to make hardware that costs money and plays games. You haven't mentioned yet a technology which either cost less money, play same games or cost same money, play game better.

LPDDR6 might be out next year, but I don't see what it has to offer. LPDDR5 offers almost nothing - it has one minor power saving feature which doesn't really work in game devices. Most everything else that LPDDR5 offered over 4 came from the node shrink of the memory and the controller. Drake isn't a memory starved design, it doesn't need more bandwidth, and it is unlikely to save power from a protocol upgrade. The benefits come from the node, and there isn't a new node that will be available to Nintendo next year that is going to offer a more cost effective product than whatever choices were available to them in 2022, or even 2020, frankly.

FP4 dual issue is useless. It's not a gaming feature, it's an AI research feature. There are no gaming related tensor core updates from Nvidia since Ampere in 2020. The ada lovelace whitepaper, which is 40 pages long, dedicates only three sentences to the tensor cores. It's a non-issue.

You have no evidence that Nvidia's plan is to lock DLSS features behind some RTX 50 architectural change. I don't know what's in Blackwell and neither do you. Or do you mean frame gen? We know DLSS gets demoed to Nintendo well before it gets released, we saw that from the leak. If Nintendo wanted the OFA that makes frame gen possible, they could have gotten it in Drake.

Presumably they chose not to have it because it would be a terrible idea. Frame generation isn't free, it takes 3-5 TFLOPS of raster performance. When your whole device only offers 1.5-3 TFLOPS of performance, frame gen costs more than native rendering. Adding the OFA that makes frame gen possible to Drake would have eaten up physical die space that could be used for something else. It was a waste.

You keep saying that it will be out of date by 2025, and that's just demonstrably not true. The architecture is still close to the top of the line. The performance level might be well below what other handheld PCs are doing in that space, but no one is going to offer significantly more power in a handheld at the cost that Nintendo does by next year. And launching 6 months earlier would not dramatically change that,

We still don't know ARM's cpu schedule for 2025, so that can't be verified at the moment.
LPDDR6 and Blackwell's release schedules aren't out either, but that didn't stop you there ;) But we do know that the 2025 ARM processor won't have 32-bit support. So if Nintendo waited for that, they'd have to give up on backwards compatibility with Switch 1 software.

ARM, Nvidia, AMD, Apple, Samsung, Mediatek, Qualcomm - their whole business model depends on yearly product releases. Branding those products as "next gen" helps sell them. That doesn't mean there is actually some generational leap happening every year, and Nintendo misses the Boat to the Undying Lands if they don't have everything at the moment of launch.




A proper 2025-ish handheld would be an improvement by architecture alone, that's just how far Ampere is expected to fall below the curve at the current pace... But again, late 2025 isn't exactly the most reassuring date to speculate about.
 
FP4 dual issue is useless. It's not a gaming feature, it's an AI research feature. There are no gaming related tensor core updates from Nvidia since Ampere in 2020. The ada lovelace whitepaper, which is 40 pages long, dedicates only three sentences to the tensor cores. It's a non-issue.
I already reiterated that we don't know enough about late 2025 to make any real hardware speculation out of it, we don't even know what Blackwell will offer to begin with... I think calling it a non-jump over the three years old Ampere is extremely premature, perhaps too much. The point is that three years of improvements are what they are, they're very much tangible and T239 will be missing them entirely, it already cannot be considered anything remotely "top of the line" by then, by definition. Excellent performer as everyone knows? Indeed, but that's that.

Frame gen was always a non-factor, but it's not even close to being the endgame for Nvidia's ML revolution... I only brought it up to exemplify how much a single architectural improvement can move the needle in significant ways. I do not believe Ampere will receive anything other than DLSS upscaler improvements in the future, and that is an issue for the "futureproofing" narrative some would like to be true. Just used this as a vehicle to say this aspect won't really be there.

I mostly disagree about dual issue being useless also, which is why I only quoted that... It has only been useless for gaming because of the open-ended nature of PCs and the shader compiler not being able to leverage those instructions often enough to impact performance, with a closed platform though? It's a completely different story, and Sony is actively betting on it for their Pro model as we speak. Videogames are already doing the "AI applications" you mention, and they're only going to get more specialized especially by the time of PS6 (ML inference mostly), which NG will actively compete with for a real while. LPDDR6 is self explanatory and yeah more of an schedule is required, but I seriously cannot call something like FP4 "useless" for what's to come... Just no.
 
Last edited:
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