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

[Deleted member 1324]

If it's out next May for Zelda I wouldn't expect it to go into full production until late Jan'23. They will delay the max possible to stop leaks before Xmas which might put people off buying a Switch or upgrading their Switch to an OLED if they hear there's a new model coming next Summer.*

*I'm talking about the casual audience, I know hardcore people collect Switch devices like stamps (I have a launch, Lite, OLED and Splatoon 3 OLED lol).

 

[John Omaha]

If it's out next May for Zelda I wouldn't expect it to go into full production until late Jan'23. They will delay the max possible to stop leaks before Xmas which might put people off buying a Switch or upgrading their Switch to an OLED if they hear there's a new model coming next Summer.*

*I'm talking about the casual audience, I know hardcore people collect Switch devices like stamps (I have a launch, Lite, OLED and Splatoon 3 OLED lol).

With the current global economic situation, I don't think they have the luxury of delaying production just to stop some potential leaks.

 

[arsen]

LG Display to start producing mid-size WOLED panels as demand for TVs declines | OLED Info

According to reports, LG Display will start producing mid-sized W-OLED panels, as it is concerned that demand for TV panels is on the decline.LGD will start producing 27-inch WOLED panels as early as by the end of October, targeting small TVs and gaming monitors. LGE will be LGD's first...

www.oled-info.com

Hopefully that means we're getting 27-32" OLED monitors with gaming features soon (hopefully before Drake), always wanted one but the gaming monitor industry seems to endlessly milk those LCD panels.

 

[Simba1]

The V2 and OLED are definitely on their last legs if next Switch is within $400.

I would be surprised if at least one of them (alongside Lite) not still selling in 2024.

Next Switch will most likely be heavily supply constrained hole 1st year, so Nintendo would keep selling some of current models alongside.

 

[Thraktor]

Semiconductor Fab Buildout Delays – 2023 to 2025 Wafer Fabrication Equipment Outlook – 2023 Below $70B WFE?

The semiconductor manufacturing world has seemingly flipped in a few months, with multiple fab buildouts and expansions slowed or delayed. Today we will address these slowdown rumors by company, sp…

semianalysis.com

Thanks for sharing this. While PS5 and XBSX/S were launched at pretty much exactly the wrong time from a supply chain point of view, it's possible that the new Switch model could be the opposite; launching at a time when new capacity has been coming online, while demand is dropping. Of course there's more to it than just the SoC, as there's always the possibility of shortages of smaller components like display driver ICs, PMICs, or whatever, but it looks like manufacturing will be much less supply-constrained in 2023 than it has been in the past couple of years.

The other thing I just noticed on SemiAnalysis is this article. Mainly regarding Nvidia's costs for Ada GPUs, it includes this comment, which I think is quite relevant for this thread:

SemiAnalysis sources indicate that the wafer cost of TSMC N5/N4 is more than 2.2x that of Samsung 8nm. With that wafer cost increase comes 2.7x higher transistor density.

I've been saying for a while that, despite the increased wafer cost of TSMC's 5nm processes, it could actually be cheaper per chip than Samsung 8nm, and this seems to confirm it.

If we assume Drake is 8 billion transistors (just a rough guess), on 8nm we would expect a 175.4mm2 die size, and 337 dies per wafer. Assuming 0.1 defects per cm2 on 8nm, that gives us an average of 280.1 yielded dies per wafer. On TSMC 5nm, we would get a 66.1mm2 die size, and 916 dies per wafer. With a 0.1 defect rate per cm2, that would leave 854.4 yielded dies on average per wafer. Combine these with the 2.2x price difference between 8nm and 5nm, and you get a Samsung 8nm Drake being 39% more expensive than a TSMC 5nm Drake. I would actually expect the real numbers to show a larger gap, as Samsung 8nm yields are unlikely to be as good as TSMC 5nm yields. If we bump the 8nm defect rate up to 0.2, it becomes 70% more expensive on a per-die basis.

 

[Skittzo]

Thanks for sharing this. While PS5 and XBSX/S were launched at pretty much exactly the wrong time from a supply chain point of view, it's possible that the new Switch model could be the opposite; launching at a time when new capacity has been coming online, while demand is dropping. Of course there's more to it than just the SoC, as there's always the possibility of shortages of smaller components like display driver ICs, PMICs, or whatever, but it looks like manufacturing will be much less supply-constrained in 2023 than it has been in the past couple of years.

The other thing I just noticed on SemiAnalysis is this article. Mainly regarding Nvidia's costs for Ada GPUs, it includes this comment, which I think is quite relevant for this thread:



I've been saying for a while that, despite the increased wafer cost of TSMC's 5nm processes, it could actually be cheaper per chip than Samsung 8nm, and this seems to confirm it.

If we assume Drake is 8 billion transistors (just a rough guess), on 8nm we would expect a 175.4mm2 die size, and 337 dies per wafer. Assuming 0.1 defects per cm2 on 8nm, that gives us an average of 280.1 yielded dies per wafer. On TSMC 5nm, we would get a 66.1mm2 die size, and 916 dies per wafer. With a 0.1 defect rate per cm2, that would leave 854.4 yielded dies on average per wafer. Combine these with the 2.2x price difference between 8nm and 5nm, and you get a Samsung 8nm Drake being 39% more expensive than a TSMC 5nm Drake. I would actually expect the real numbers to show a larger gap, as Samsung 8nm yields are unlikely to be as good as TSMC 5nm yields. If we bump the 8nm defect rate up to 0.2, it becomes 70% more expensive on a per-die basis.

I guess the more relevant question would be, would all of this have been known back when they finalized plans for Drake's design in 2019-2020?

 

[Thraktor]

I guess the more relevant question would be, would all of this have been known back when they finalized plans for Drake's design in 2019-2020?

Yeah, that's the pertinent question. My guess is yes. The price of 8nm wafers likely haven't changed much (if at all) since then, and yields on 8nm would have been a known quantity. TSMC 5nm prices have likely changed since projections that would have been made in late 2019/early 2020, but only upwards, meaning the TSMC process would if anything have seemed like a better deal back then. Yields on TSMC 5nm would have been a bit of an unknown, but given the extremely high yields of TSMC's 7nm process, it seems very unlikely that Nvidia would have predicted yields bad enough to cancel out the cost and density advantages.

The main reference point we have for this is Nvidia's Ada series of GPUs. They likely would have made the decision to fabricate them on TSMC 5nm at the same time (or perhaps earlier), and leaks indicate everything down to the entry-level AD107 is being manufactured on TSMC 5nm. If they expected 5nm to make economic sense for an entry-level GPU like AD107 releasing in 2023, then I would be surprised if they thought 8nm would be a cheaper option for Drake in 2023.

Of course, this doesn't say anything about TSMC 7nm/6nm, which may have been projected to be a cheaper option at the time. However, I assume again the same calculations would have applied to AD107/AD106, where Nvidia appear to have chosen 5nm instead.

 

[Hirogami]

Hi everyone,
I don't think Drake/T239 is in 8nm from samsung because I didn't find any chip in a78/a78c from samsung in 8nm but in 5-4nm.
It is very light to say that but I do not think that nintendo and nvidia remain on a 8nm doomed to disappear and not very profitable.
It would be better for nintendo and nvidia to change the engraving of drake/t239 knowing that tsmc has ampere in production line since 2019-2020 .

I imagine that drake had a modification of the base around 2020-2021 thanks to the advance made on atlan which was abandoned thereafter, seen the money put in the R&D of nintendo since 2019-2020 that it is nvidia or nintendo everyone was winning.

 

[ILikeFeet]

Hi everyone,
I don't think Drake/T239 is in 8nm from samsung because I didn't find any chip in a78/a78c from samsung in 8nm but in 5-4nm.
It is very light to say that but I do not think that nintendo and nvidia remain on a 8nm doomed to disappear and not very profitable.
It would be better for nintendo and nvidia to change the engraving of drake/t239 knowing that tsmc has ampere in production line since 2019-2020 .

I imagine that drake had a modification of the base around 2020-2021 thanks to the advance made on atlan which was abandoned thereafter, seen the money put in the R&D of nintendo since 2019-2020 that it is nvidia or nintendo everyone was winning.

a prior SoC with a78 made on 8nm wouldn't have any bearing on whether or not Drake would be made on the node. when Nvidia started designing the SoC, the intended node would have been made there. Atlan/Thor probably wasn't used as any sort of reference since Drake was already forked from Orin R&D

 

[ShaunSwitch]

To add to Thraktors point around Drake potentially being on 5nm. We have already spoke of the potential savings from not having to do a die shrink of Drake, Drake is likely a product with a target 6-7 year lifespan so putting it on an advanced node, that also works out to be cheaper per die and result in a better product makes sense up front.

The fact Orin on 8nm Samsung is also a moot point. Yes Drake is based on Orin to an extent however Orin isn't a product that will be the best in its class for the next 6/7 years, it is being surplanted by Atlan in 2025 so a move to an advanced node can happen then and provide another jump in performance from Nvidia automotive chips. Its fine having this regular release cycle for automotive products, as you don't have hundreds of dev studios needing dev kits and having to write games for the new architecture, Nintendo however have no need for new architecture every three years, it would in fact be a detriment due to how long it takes to release games.

For the switch, this chip will most likely be the future of the console for the next 6/7 years so do it right the first time and there's no need for a costly shrink. Plus, 66mm^2 is insane, battery life better than Mariko is a possibility.

 

[Mr Swine]

To add to Thraktors point around Drake potentially being on 5nm. We have already spoke of the potential savings from not having to do a die shrink of Drake, Drake is likely a product with a target 6-7 year lifespan so putting it on an advanced node, that also works out to be cheaper per die and result in a better product makes sense up front.

The fact Orin on 8nm Samsung is also a moot point. Yes Drake is based on Orin to an extent however Orin isn't a product that will be the best in its class for the next 6/7 years, it is being surplanted by Atlan in 2025 so a move to an advanced node can happen then and provide another jump in performance from Nvidia automotive chips. Its fine having this regular release cycle for automotive products, as you don't have hundreds of dev studios needing dev kits and having to write games for the new architecture, Nintendo however have no need for new architecture every three years, it would in fact be a detriment due to how long it takes to release games.

For the switch, this chip will most likely be the future of the console for the next 6/7 years so do it right the first time and there's no need for a costly shrink. Plus, 66mm^2 is insane, battery life better than Mariko is a possibility.

That depends entirely on what Nintendo wants and what they have come up with since the development and birth of the SoC.

If they go with 8nm then it will be a really big chip compared to 6/7nm. But I hope that since Nintendo have stopped chasing after old and cheap tech they actually invest in something really good.

Something that will last them 7 years at least

 

[ILikeFeet]

That depends entirely on what Nintendo wants and what they have come up with since the development and birth of the SoC.

If they go with 8nm then it will be a really big chip compared to 6/7nm. But I hope that since Nintendo have stopped chasing after old and cheap tech they actually invest in something really good.

Something that will last them 7 years at least

They were forced to stop with old cheap tech long ago. They even stopped with the 3ds as they were looking at the relatively modern Tegra 2.

 

[ReddDreadtheLead]

It’s on the Nintendo nanometer, mystery solved yall.

Everyone go hunt for any job listing from nvidia or Nintendo! (Or ARM)


Let’s see any relevant info of background things.

 

[Mr Swine]

They were forced to stop with old cheap tech long ago. They even stopped with the 3ds as they were looking at the relatively modern Tegra 2.

Yet they made the Frankenstein Wii U which married old and new.

 

[ILikeFeet]

Yet they made the Frankenstein Wii U which married old and new.

The gpu was modern-ish at least. If there was another PowerPC option at the time, they might have gone with that, but shit was dead and Nintendo demanded BC

 

[Hermii]

They were forced to stop with old cheap tech long ago. They even stopped with the 3ds as they were looking at the relatively modern Tegra 2.

The Wii U was just old, expensive tech.

 

[Hermii]

Yeah, that's the pertinent question. My guess is yes. The price of 8nm wafers likely haven't changed much (if at all) since then, and yields on 8nm would have been a known quantity. TSMC 5nm prices have likely changed since projections that would have been made in late 2019/early 2020, but only upwards, meaning the TSMC process would if anything have seemed like a better deal back then. Yields on TSMC 5nm would have been a bit of an unknown, but given the extremely high yields of TSMC's 7nm process, it seems very unlikely that Nvidia would have predicted yields bad enough to cancel out the cost and density advantages.

The main reference point we have for this is Nvidia's Ada series of GPUs. They likely would have made the decision to fabricate them on TSMC 5nm at the same time (or perhaps earlier), and leaks indicate everything down to the entry-level AD107 is being manufactured on TSMC 5nm. If they expected 5nm to make economic sense for an entry-level GPU like AD107 releasing in 2023, then I would be surprised if they thought 8nm would be a cheaper option for Drake in 2023.

Of course, this doesn't say anything about TSMC 7nm/6nm, which may have been projected to be a cheaper option at the time. However, I assume again the same calculations would have applied to AD107/AD106, where Nvidia appear to have chosen 5nm instead.

My question is then, why did they use 8nm for Orin? If they thought it would end up as expensive/ more expensive.

Orin and Drake are very much part of the same family.

 

[Hirogami]

a prior SoC with a78 made on 8nm wouldn't have any bearing on whether or not Drake would be made on the node. when Nvidia started designing the SoC, the intended node would have been made there. Atlan/Thor probably wasn't used as any sort of reference since Drake was already forked from Orin R&D

It seems to me that the a78 is not at all advised with an engraving higher than 5nm otherwise we might as well stay on a76-a77.
We have an orin in 8nm and a drake with a new cpu which is much smaller, the DLA and PVA absent as the chip orin nano, more orin and drake only share the same gpu we do not know yet for the lpddr5 8-12Gb 68-102GB/s
I think that seen the size of orin and its consumption, nintendo and nvidia have switched drake to tsmc with a smaller node for the sustainability of the chip, nintendo will not make the mistake committed with the switch tx1>20nm then tx+1>16nm.
I imagine that given that ampere exists in 7nm at tsmc drake will be in 7-6nm which will be a better compromise than the 8nm of samsung.

 

[ILikeFeet]

My question is then, why did they use 8nm for Orin? If they thought it would end up as expensive/ more expensive.

Orin and Drake are very much part of the same family.

Orin isn't intended for space and thermally constrained applications so they could get away with a larger and more power consuming due.

 

[Hermii]

Orin isn't intended for space and thermally constrained applications so they could get away with a larger and more power consuming due.

They can get away with yes, but that doesn’t explain why they would go for a worse and possibly more expensive solution.

 

[ReddDreadtheLead]

They can get away with yes, but that doesn’t explain why they would go for a worse and possibly more expensive solution.

They aren’t going to really be made at the same volume as the switch and are considerably more expensive than the switch.

So it’s already a different segment.

 

[Mr Swine]

They can get away with yes, but that doesn’t explain why they would go for a worse and possibly more expensive solution.

Kinda like why they went with 8nm for the RTX 30xx cards. I’m guessing they got a good deal but in the end it was a mistake performance, size and wattage wise

 

[Hermii]

They aren’t going to really be made at the same volume as the switch and are considerably more expensive than the switch.

True and true, but if Orin at tsmc would be cheaper and more energy efficient that answer is still not satisfactory imo.

 

[JoshuaJSlone]

I'm out of my areas of expertise here, but: keeping basically the same hardware going for 5+ years is pretty unusual outside of game machines. Anyone making machines using Orin today probably has less concern with viability of continued production of it in 2028 and beyond.

 

[Kevin]

If new Switch hardware is coming in early 2023, when do we expect rumors and leaks to start picking up? This has been a very quiet year (opinion) for rumors and leaks regarding new hardware and it leaves me more and more skeptical that we are getting a 4K DLSS Switch in 2023. It will be such a huge bummer if 2023 comes and goes and still no new hardware or even a hint of new hardware coming. Switch turns 6 in March. I want to move onto something better. Don't want to be in year 7 of Switch. Ugh.

So are you guys still confident in new hardware in 2023? When do you guys expect the leaks and quality rumors to start picking up from the drought we have been in?

 

[NineTailSage]

Orin isn't intended for space and thermally constrained applications so they could get away with a larger and more power consuming due.

They can get away with yes, but that doesn’t explain why they would go for a worse and possibly more expensive solution.

They aren’t going to really be made at the same volume as the switch and are considerably more expensive than the switch.

So it’s already a different segment.

In comparison as well the mark-up for Orin is way higher than Drake will be In the Switch...
From $200 (Orin Nano) up to $1600 (Jetson AGX Orin 64GB), so the profit margin should easily make up for Samsung's 8nm shortcomings.

 

[Hermii]

If new Switch hardware is coming in early 2023, when do we expect rumors and leaks to start picking up? This has been a very quiet year (opinion) for rumors and leaks regarding new hardware and it leaves me more and more skeptical that we are getting a 4K DLSS Switch in 2023. It will be such a huge bummer if 2023 comes and goes and still no new hardware or even a hint of new hardware coming. Switch turns 6 in March. I want to move onto something better. Don't want to be in year 7 of Switch. Ugh.

So are you guys still confident in new hardware in 2023? When do you guys expect the leaks and quality rumors to start picking up from the drought we have been in?

These things are hard to predict. I expect things to really pick up next year.

Edit: But you never know, there might be a huge eurogamer article tomorrow.

 

[ReddDreadtheLead]

True and true, but if Orin at tsmc would be cheaper and more energy efficient that answer is still not satisfactory imo.

Nvidia already knows that, and energy efficiency for an battery isn’t what they were really aiming for clearly, or it wouldn’t be on that node but the same node as their Datacenter GPUs (GA100).

They aimed for a very performant piece of kit for lower power, not the lowest power.

The price of these things can exceed $1500 making them enough back in profit due to really good margins.

The lower bracket ones are those that most likely had defects and they were binned to a lower price range.

Nvidia is really making the most out of the silicon and how it’ll be used/allocated to make the maximum profit.



There was a rumor about nvidia trying to get a good deal with TSMC 7nm by using SEC 7, but TSMC gave them the finger and NV were stuck with SEC. 7 wasn’t producing good results, yields being terrible, so they had to pull back to 10nm and improve it to get the 8N that the ampere line uses. And funnily enough, nothing exists for “7nm” at SEC.

 

[Dakhil]

Yeah, that's the pertinent question. My guess is yes. The price of 8nm wafers likely haven't changed much (if at all) since then, and yields on 8nm would have been a known quantity. TSMC 5nm prices have likely changed since projections that would have been made in late 2019/early 2020, but only upwards, meaning the TSMC process would if anything have seemed like a better deal back then. Yields on TSMC 5nm would have been a bit of an unknown, but given the extremely high yields of TSMC's 7nm process, it seems very unlikely that Nvidia would have predicted yields bad enough to cancel out the cost and density advantages.

The main reference point we have for this is Nvidia's Ada series of GPUs. They likely would have made the decision to fabricate them on TSMC 5nm at the same time (or perhaps earlier), and leaks indicate everything down to the entry-level AD107 is being manufactured on TSMC 5nm. If they expected 5nm to make economic sense for an entry-level GPU like AD107 releasing in 2023, then I would be surprised if they thought 8nm would be a cheaper option for Drake in 2023.

Of course, this doesn't say anything about TSMC 7nm/6nm, which may have been projected to be a cheaper option at the time. However, I assume again the same calculations would have applied to AD107/AD106, where Nvidia appear to have chosen 5nm instead.

Considering Jensen Huang mentioned that Nvidia needs to secure process node capacity 1.5 years in advance, I think there's a possibility Nvidia was at least considering having mid range to entry level Ada GPUs fabricated using TSMC's N6 process node in the similar vein to how AMD's having Navi 33 fabricated using TSMC's N6 process node. And there's a rumour about how Nvidia was considering using TSMC's 7 nm** process node in 2021 (here and here). (Of course, the reliability of that rumour is a different story altogether.)

But when Nvidia saw how competitive RDNA 2 GPUs were compared to Ampere GPUs, Nvidia could have thought that having all Ada GPUs fabricated using TSMC's 4N process node was necessary to ensure all Ada GPUs were at the very least very competitive against AMD's RDNA 3 GPUs. And therefore, Nvidia could have made that decision in early 2021.

** → a marketing nomenclature used by all foundry companies

---

Anyone making machines using Orin today probably has less concern with viability of continued production of it in 2028 and beyond.

Yeah, especially with Nvidia mentioning that all the Jetson Orin modules (Jetson AGX Orin to Jetson Orin Nano) will be available until January 2028.

 

[KMStwo]

They can get away with yes, but that doesn’t explain why they would go for a worse and possibly more expensive solution.

I think it may be because 8nm is a more mature and kind of more reliable node for automotive products(which iirc from someone here or the old place has been said to neccessitate more reliability) since it is much older than the newer nodes.

 

[fwd-bwd]

BTW, the following was posted on Baidu, a Chinese forum, on July 7. What do you think?

今年不会有新机的【ns吧】_百度贴吧

今年不会有新机的..先说一下我知道的情报（信息来源烟台xx康）没有oled版lite，Lite和续航版作为低端线被保留，估计直到NS退役。新机是加强版，非换代。直接沿用Oled的面板和产品线，只做了芯片的升级。oled的

tieba.baidu.com

Thanks for posting this. I saw the rumor in July, but didn't want to share it here due to the OP's lack of track record as an insider. But now that part of the rumor has been validated, perhaps it's worth a revisit. A quick breakdown of their claims (translations mine):

• Source of information was the Fo***nn in Yantai.
• There won't be an OLED Lite. The Lite and v2 models will be kept as the low-end products.
• The "enhanced" model will inherit the OLED model's display panel and production lines.
• "Only the SoC will be upgraded." [not sure if "only" should be taken literally; could mean "mainly"]
• The assembly lines of OLED model began shifting to the new model in May. [unclear whether the shift was also completed in May]
• The new model won't be released this year, due to the time needed for production and distribution. [unclear if the timing was a learned fact or merely inferred]
• Two special editions of OLED model were produced in the first half of this year, in order to clear out the inventory of prior SoC. [more about this below]

With the introductions of Splatoon 3 SE and Scarlet & Violet SE, the last bullet point has been partially substantiated. Granted, it could be just a lucky guess.

It's not clear why they'd want to be clearing stock of "old chips" if the other two models are still being produced.

The OP addressed this further down the thread. They are clearing out the OLED model's motherboard, not just the SoC by itself; the v2 and OLED motherboards differ, despite sharing the same SoC. Because of the manufacturing and QA process, a certain amount of inventory needs to be prebuilt.

 

[Dakhil]

Hi everyone,
I don't think Drake/T239 is in 8nm from samsung because I didn't find any chip in a78/a78c from samsung in 8nm but in 5-4nm.

Well, all of the Jetson Orin modules (Jetson AGX Orin to Jetson Orin Nano) are fabricated using Samsung's 8N process node and use the Cortex-A78AE, which I believe ReddDreadtheLead mentioned is practically the Cortex-A78C, but with automotive features added.

 

[ILikeFeet]

In comparison as well the mark-up for Orin is way higher than Drake will be In the Switch...
From $200 (Orin Nano) up to $1600 (Jetson AGX Orin 64GB), so the profit margin should easily make up for Samsung's 8nm shortcomings.

let's not forget the orin board markup so high, GM said "no thanks, we'll make our own"

 

[karmitt]

Thanks for posting this. I saw the rumor in July, but didn't want to share it here due to the OP's lack of track record as an insider. But now that part of the rumor has been validated, perhaps it's worth a revisit. A quick breakdown of their claims (translations mine):

• Source of information was the Fo***nn in Yantai.
• There won't be an OLED Lite. The Lite and v2 models will be kept as the low-end products.
• The "enhanced" model will inherit the OLED model's display panel and production lines.
• "Only the SoC will be upgraded." [not sure if "only" should be taken literally; could mean "mainly"]
• The assembly lines of OLED model began shifting to the new model in May. [unclear whether the shift was also completed in May]
• The new model won't be released this year, due to the time needed for production and distribution. [unclear if the timing was a learned fact or merely inferred]
• Two special editions of OLED model were produced in the first half of this year, in order to clear out the inventory of prior SoC. [more about this below]

With the introductions of Splatoon 3 SE and Scarlet & Violet SE, the last bullet point has been partially substantiated. Granted, it could be just a lucky guess.


The OP addressed this further down the thread. They are clearing out the OLED model's motherboard, not just the SoC by itself; the v2 and OLED motherboards differ, despite sharing the same SoC. Because of the manufacturing and QA process, a certain amount of inventory needs to be prebuilt.

Thanks for the update. So the implication of the post is that the OLED model is going to be replaced after 2022? Or just that this particular factory is shifting gears?

I was under the impression that the OLED model had quite some effort put into the design, hence Nintendo having a write up on the differences around it (can’t find it but I know one was released). Throwing all that away after only 1-1.5 years sounds unlikely… If a system was going to get axed I’d have thought it would be the V2, just because it feels pretty dated now

 

[MP!]

I may be alone... but I think the Lite will go quietly... and they'll never bring it back

 

[Pokemaniac]

The OP addressed this further down the thread. They are clearing out the OLED model's motherboard, not just the SoC by itself; the v2 and OLED motherboards differ, despite sharing the same SoC. Because of the manufacturing and QA process, a certain amount of inventory needs to be prebuilt.

That makes a bit more sense if they were actually referring to the whole motherboard. All three systems would have unique motherboards. Would make the Switch OLED strangely short-lived, but it could make sense if they don't want to price cut anything.

 

[karmitt]

That makes a bit more sense if they were actually referring to the whole motherboard. All three systems would have unique motherboards. Would make the Switch OLED strangely short-lived, but it could make sense if they don't want to price cut anything.

Yeah this is the biggest question I have around that post. We have the other vague post about the size of the device possibly changing / a possible new backplate. If that is true, it makes the excellent kick-stand work they did for the OLED model feel kind of throwaway.

I may be alone... but I think the Lite will go quietly... and they'll never bring it back

I personally can’t see that ever happening. It’s more essential than having a V2 and OLED in the same market. It’s the closest thing to the affordability and size of yesteryear’s GameBoy line. The full size device clearly has the edge for most consumers, but I still see plenty of Lites in the wild. Just recently read something from IGNs Kat Bailey about seeing mostly Lites while traveling in Japan.

I’m sure Nintendo would rather have everybody buy the regular Switch, but it’s not gonna get any smaller than it currently is; That size barrier for travel and convenience is always going to be there

 

[Skittzo]

I may be alone... but I think the Lite will go quietly... and they'll never bring it back

Until they can get another model down to $199 I don't see that happening.

 

[JoshuaJSlone]

I was under the impression that the OLED model had quite some effort put into the design, hence Nintendo having a write up on the differences around it (can’t find it but I know one was released). Throwing all that away after only 1-1.5 years sounds unlikely… If a system was going to get axed I’d have thought it would be the V2, just because it feels pretty dated now

But this rumor would mean the opposite of that. All the effort put into the design would now live on in an even more appealing version because it's not stuck with the same 2017/2019 hardware.

But it seems pretty implausible to me they'd WANT the next big thing to look practically identical to something older.

 

[Thraktor]

My question is then, why did they use 8nm for Orin? If they thought it would end up as expensive/ more expensive.

Orin and Drake are very much part of the same family.

I'd say there are a few factors with Orin. For one, the chip was announced way back in March 2018, so was potentially in development as much as two years before Drake. It's also been made available before any of Nvidia's 5nm chips. It's been made widely available in Jetson products this year, but it had likely been shipping at the very least small volumes chips to automotive customers for quite a while before that, possibly going back to early last year. Unlike GPUs, where you can pretty much just dump them on the market, automotive customers will need a long time to basically build an entire autonomous driving system, infotainment system, etc., on top of these chips, and with the first Orin-powered cars apparently due to be delivered to customers in August of this year, they must have been shipping Orin units to at least some automotive customers a long time ago.

The other factor with automotive chips is that they generally use pretty old manufacturing processes. I assume that the ASIL-B/ASIL-D certification is pretty slow, and I would imagine using the latest manufacturing process adds potential delays and risks to that certification. Orin is probably on a pretty advanced process as far as ASIL-D certified chips go.

Edit: Also, in terms of costs, Orin and Drake are on opposite ends of the spectrum for Nvidia. Orin is probably one of Nvidia's highest-margin products, whereas Drake is possibly their lowest-margin. Drake is also much more power-constrained in its intended use-case, where it's all about how much they can squeeze out of 5W or so in portable mode, where Orin is a 50W product that's often attached to a 200W GPU, so wouldn't benefit as much from the improved efficiency of the more advanced node.

 

[karmitt]

But this rumor would mean the opposite of that. All the effort put into the design would now live on in an even more appealing version because it's not stuck with the same 2017/2019 hardware.

But it seems pretty implausible to me they'd WANT the next big thing to look practically identical to something older.

I was I guess trying to reconcile this rumor with the very unconfirmed talk of a new backplate.

Perhaps that backplate aspect could still be true but the rest of the changes are intact - screen, speaker constructions, material changes etc.

 

[Skittzo]

But it seems pretty implausible to me they'd WANT the next big thing to look practically identical to something older.

Yeah this has always been a bit of a sticking point for me too. It's pretty unlike them but then again the Switch is kinda a new frontier in general.

Either way it can reuse most of the same components and form factor and have some major differentiators.

 

[fwd-bwd]

Thanks for the update. So the implication of the post is that the OLED model is going to be replaced after 2022? Or just that this particular factory is shifting gears?

I was under the impression that the OLED model had quite some effort put into the design, hence Nintendo having a write up on the differences around it (can’t find it but I know one was released). Throwing all that away after only 1-1.5 years sounds unlikely… If a system was going to get axed I’d have thought it would be the V2, just because it feels pretty dated now

That is also my understanding. Yes, the OLED Switch allegedly will be replaced by the "enhanced" model. This isn't a popular or safe prediction for a fake insider to make, therefore the rumor caught my attention despite the unproven source. On the surface phasing OLED out this quickly seems counterintuitive, but it may have some merits from the product management standpoint.

• As pointed out by another poster earlier, the OLED model is targeting the enthusiasts and repurchasers. The Drake model will be targeting the same segment, rendering the former redundant.
• By keeping the dated v2 and limited Lite instead of the more attractive OLED, it might accelerate the adoption curve of the new model, which would be good for Nintendo and other developers.
• If the rumor is correct that the main differentiator from the OLED model is the SoC, discontinuing the OLED would create better product tiers ($200/$300/$400+) and lessen consumer confusion (not having two models with the same display).
• For the same reason, the R&D and production setup spent on the OLED essentially are shared by both models. (The mod uncle also mentioned that both models were prototyped together last year.) Even if the OLED is indeed short-lived, the efforts Nintendo put in would not be wasted.
• Nintendo stated a few times that the profit margin of OLED model is lower. By discontinuing it, the company could make some design changes on the new model to increase the profit margin, not unlike what Sony did with the new PS5 (CFI-1200).

 

[Deleted member 887]

Yeah this has always been a bit of a sticking point for me too. It's pretty unlike them but then again the Switch is kinda a new frontier in general.

Either way it can reuse most of the same components and form factor and have some major differentiators.

The 3DS looked so much like a DS I didn’t realize it wasn’t just a DS with a little joystick till years after it came out.

Whatever happens, it seems like Drake will be at least in some way a new strategy for Nintendo. But the more I ignore the TV console line and just focus on the handhelds, the less it seems out of place to me.

 

[My Tulpa]

So, this is a good place to do it.


People believe that the system will launch in May 2023 right? Because of ZELDA correct?


Let’s assume this scenario only for curiosity sake: May 12 happens and there is no new system. Thus, Zelda did not launch with Switch 2/Pro.

June13-16 is E3, Nintendo announces a few games, a few are set for holiday but not given a date.

June after E3/July Nintendo announces new system set in stores for holiday 2023.


Here’s what I want to know, I want a list of games (read: IPs) that can realistically be launched title software for this new platform if it isn’t Tears of the Kingdom. Because it could launch with a new 2D or 3D Mario, right? Those are huge selling titles.


Would Metroid Prime HD set for holiday 2023 make sense?

Tomodachi Life?

A brand new IP?

New Donkey Kong?

Splatoon and Zelda are out of the question, Pokemon may also be out of the question too, Fire Emblem is out of the question too. Mario Kart I don’t see as that is still getting software support by then…


Like if it were to release in H2 2023, what would be the title to release with?



I’m asking it here because the community here is filled with very passionate Nintendo fans who know more about their intellectual properties in a sense of a launch system software than I do, so I’m more curious on what your thoughts are if it does release on the second half of 2023, as in what on earth would even release it with if it was to set for that time. And it doesn’t release with Zelda in May. Because I know some people do believe that the second half 2023 or even 2024 makes more sense but I’m also drawing a blank on what software it would release with, you feel me?


So to try to actually analyze it from a perspective that is important for the discussion we have to actually pinpoint and narrow down what are the most possible release titles for this system.


And I’d like to remind people that the previous time a Zelda was cross generation, the following 3D Zelda was not cross-generation, and a successor released a year after, but it bombed. Also, in that previous time that they released a Zelda as a cross generation title, the original platform bombed, but the following platform was a massive success.

Of course, I don’t see TOTK being exclusive, but just that it isn’t always indicative of when a hardware will launch.

But if we narrow it down we can get a better idea I think of what can actually launch with it.


I know people say they have loads of options but when you put it into perspective they don’t really have that many….

and not releasing with Zelda would be sorta missed opportunity, even if they have a massive install base, there are perks the Pro/Succ/2/Dogwalker can get over the base




TL;DR: if it doesn’t release with Zelda what other IP do you think is actually realistic or plausible for an H2 release?

Late on this but I see no good argument why Nintendo would hold off releasing this new model that late (second half of 2023).

Everything points to them wanting to get this upgrade option out there as soon as possible.

As for software, it seems like Nintendo would release software around it to promote the software, not the hardware. ToTK does not need such promotion, so it will be something else. The only connection between Drake and ToTK would be Nintendo wanting gamers to have the ability to play the game with better graphics/performance if they want.

So I see it much like Metroid Dread launching with OLED. Some type of “relatively more niche” game that could benefit from having a spotlight on it.

 

[LiC]

But this rumor would mean the opposite of that. All the effort put into the design would now live on in an even more appealing version because it's not stuck with the same 2017/2019 hardware.

But it seems pretty implausible to me they'd WANT the next big thing to look practically identical to something older.

Since we're talking about the tablet part of the console here, I don't really see what they could change to make it not look practically identical. The backplate may be different, but that's about it. The Joycons and dock are probably the places where the model would receive its design flair -- if it even has any to speak of -- but that doesn't concern the rumor about OLED production lines.

I don't 100% agree with the justifications for the rumor, but it's not rendered implausible by any one thing, either.

 

[Raccoon]

Since we're talking about the tablet part of the console here, I don't really see what they could change to make it not look practically identical. The backplate may be different, but that's about it. The Joycons and dock are probably the places where the model would receive its design flair -- if it even has any to speak of -- but that doesn't concern the rumor about OLED production lines.

I don't 100% agree with the justifications for the rumor, but it's not rendered implausible by any one thing, either.

yeah, this sort of argument is kind of predicated on the idea that the switch isn't just a screen with two controllers lol

best they could do is give the joy-cons' buttons a snes circle

 

[karmitt]

That is also my understanding. Yes, the OLED Switch allegedly will be replaced by the "enhanced" model. This isn't a popular or safe prediction for a fake insider to make, therefore the rumor caught my attention despite the unproven source. On the surface phasing OLED out this quickly seems counterintuitive, but it may have some merits from the product management standpoint.

• As pointed out by another poster earlier, the OLED model is targeting the enthusiasts and repurchasers. The Drake model will be targeting the same segment, rendering the former redundant.
• By keeping the dated v2 and limited Lite instead of the more attractive OLED, it might accelerate the adoption curve of the new model, which would be good for Nintendo and other developers.
• If the rumor is correct that the main differentiator from the OLED model is the SoC, discontinuing the OLED would create better product tiers ($200/$300/$400+) and lessen consumer confusion (not having two models with the same display).
• For the same reason, the R&D and production setup spent on the OLED essentially are shared by both models. (The mod uncle also mentioned that both models were prototyped together last year.) Even if the OLED is indeed short-lived, the efforts Nintendo put in would not be wasted.
• Nintendo stated a few times that the profit margin of OLED model is lower. By discontinuing it, the company could make some design changes on the new model to increase the profit margin, not unlike what Sony did with the new PS5 (CFI-1200).

Other poster might have been me? Or maybe not, it's a very apparent position :]

Regardless, this feels like a reasonable call. This thread has spent countless posts trying to reconcile price points, naming, positioning etc. after the 'OLED Model' was thrown into the mix. If it's taken out and replaced with the new hardware at $400, some of the worst sticking points resolve. Repeating some of what you've already said but:

• The $450 or $500 could have worked for a strong launch but it felt a little too high for a product Nintendo would want everybody to own down the line
• The oddity of the name 'OLED Model' when the new product is also a 'Switch' with an OLED screen is also no longer a consideration.
• Edit*: The design of the OLED model manages to be distinct enough from the current V2 due to materials, kickstand and screen size, but there was always a question of how the new model might differentiate itself from the OLED

I just wonder how do they transition off of the OLED and onto the new model? There's heaps of marketing out there for the OLED model currently. Do they suddenly stop marketing it and wait for the remaining product to be sold off (probably discounted at retail)?

Also, is the OLED Model dock sufficient to support 4K/60/HDR? Or would that require additional changes?

 

[chocolate_supra]

best they could do is give the joy-cons' buttons a snes circle

yesyes

I wants

 

[cw_sasuke]

Man of this actually replaces the SWOLED and can be sold for around 400 bucks this would be a mayor win.

This would mean that Nintendo is probably be more agreesive in what they are able to ship initially than I thought.

Switch Lite 149€
Switch V2 249€
Drake 399€

Would be a really good setup for the future.