Discussion Leading Edge Foundry Node advances (TSMC, Samsung Foundry, Intel) - [2020 - 2025]

[DisEnchantment]

TSMC's N7 EUV is now in its second year of production and N5 is contributing to revenue for TSMC this quarter. N3 is scheduled for 2022 and I believe they have a good chance to reach that target.


N7 performance is more or less understood.


This year and next year TSMC is mainly increasing capacity to meet demands.

AnandTech Forums: Technology, Hardware, Software, and Deals

Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

www.anandtech.com

For Samsung the nodes are basically the same from 7LPP to 4 LPE, they just add incremental scaling boosters while the bulk of the tech is the same.

Samsung is already shipping 7LPP and will ship 6LPP in H2. Hopefully they fix any issues if at all.
They have two more intermediate nodes in between before going to 3GAE, most likely 5LPE will ship next year but for 4LPE it will probably be back to back with 3GAA since 3GAA is a parallel development with 7LPP enhancements.

AnandTech Forums: Technology, Hardware, Software, and Deals

Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

www.anandtech.com

Samsung's 3GAA will go for HVM in 2022 most likely, similar timeframe to TSMC's N3.
There are major differences in how the transistor will be fabricated due to the GAA but density for sure Samsung will be behind N3.
But there might be advantages for Samsung with regards to power and performance, so it may be better suited for some applications.
But for now we don't know how much of this is true and we can only rely on the marketing material.

This year there should be a lot more available wafers due to lack of demand from Smartphone vendors and increased capacity from TSMC and Samsung.
Lots of SoCs which dont need to be top end will be fabbed with N7 or 7LPP/6LPP instead of N5, so there will be lots of wafers around.

Most of the current 7nm designs are far from the advertized density from TSMC and Samsung. There is still potential for density increase compared to currently shipping products.
N5 is going to be the leading foundry node for the next couple of years.

For a lot of fabless companies out there, the processes and capacity available are quite good.

---------------------------------------------------------------------------------------------------------------------------------------------------


FEEL FREE TO CREATE A NEW THREAD FOR 2025+ OUTLOOK, I WILL LINK IT HERE

 

[Hans Gruber]

I think the better argument is that TSMC has good silicon but they keep raising the prices for diminished returns. They make good silicon but charge too much for what you get. If 18A completely sucks, that would be unfortunate. TSMC charges and arm and leg for N3P and will charge a kings ransom for N2. We kind of need Intel to stop failing in the semiconductor space.

 

[511]

I think the better argument is that TSMC has good silicon but they keep raising the prices for diminished returns. They make good silicon but charge too much for what you get. If 18A completely sucks, that would be unfortunate. TSMC charges and arm and leg for N3P and will charge a kings ransom for N2. We kind of need Intel to stop failing in the semiconductor space.

TSMC is kind of deserving of it as for IFS or Intel they have been having lot's of delays and poor decision past decade .....

 

[DZero]

Snapdragon 8 Elite Gen 6 & Dimensity 9600 To Adopt TSMC’s 2nm ‘N2P’ Process Instead Of N2 To Obtain A Higher CPU Frequency Advantage Over A20, A20 Pro

Qualcomm and MediaTek switching to TSMC’s 2nm N2P process for the Snapdragon 8 Elite Gen 6 and Dimensity 9600 will be all about reaching a CPU frequency

wccftech.com

🫠 the current android SoCs can’t even maintain 4.2GHz and 4.6GHz even for a bit

Even Apple phone SoCs can't maintain over 4 Ghz for long time, on tablets is another story, Android and Apple alike.

 

[511]

Even Apple phone SoCs can't maintain over 4 Ghz for long time, on tablets is another story, Android and Apple alike.

kind of a given cause it requires more watt than the Phones can dissipate

 

[jdubs03]

Is N2P even going to be available for Q4 2026?
If it is, it’s barely and I’d imagine a decent amount more expensive.

 

[eriksp92]

18A-P will be viable and competitive if they get the fmax close to or exceeding N2 and the yields are good; the Intel 4 to Intel 3 transition gives me hope for that. SRAM scaling is dead, and if Intel can offer up this node cheap enough per wafer that might be enough to sway customers if BSPD actually delivers on curbing leakage at low voltage and the PDK is usable. A slight density disadvantage is nothing if those things come together.

 

[Doug S]

With the current situation... I don't be surprised if the iPhone line (all of them) ends being delayed by 2 months due RAM, storage and even other chips supply.

One of the reasons Apple prepays billions for guaranteed delivery of DRAM and NAND is to insure there are no supply disruptions. Their product release schedule won't be affected by all the chaos in the memory market.

 

[Doug S]

Is N2P even going to be available for Q4 2026?
If it is, it’s barely and I’d imagine a decent amount more expensive.

Supposedly TSMC is pulling in N2P and it will follow N2 by only six months instead of the usual year. Still, given that N2 only entered mass production in recent weeks, if N2P begins mass production in June that doesn't leave much time for wafers exiting the line, testing, packaging, etc. to deliver finished products to customer hands before the end of the year.

 

[jdubs03]

Supposedly TSMC is pulling in N2P and it will follow N2 by only six months instead of the usual year. Still, given that N2 only entered mass production in recent weeks, if N2P begins mass production in June that doesn't leave much time for wafers exiting the line, testing, packaging, etc. to deliver finished products to customer hands before the end of the year.

Hmm. No it doesn’t especially when some of these devices typically release only a month or two after Apple releases theirs. Unless they delay them…

 

[Geddagod]

Hmm. No it doesn’t especially when some of these devices typically release only a month or two after Apple releases theirs. Unless they delay them…

Mediatek explicitly said they would have those parts would be available in late 2026

 

[Doug S]

Hmm. No it doesn’t especially when some of these devices typically release only a month or two after Apple releases theirs. Unless they delay them…

If Apple is delayed why would others be able to ship on time? If Apple is delayed due to N2 delays there's no chance N2P isn't also delayed. If Apple is somehow delayed due to RAM/NAND supply issues, how would anyone else not be delayed for the same reasons? And likely much worse because if Apple is delayed despite their ironclad prepaid contracts, no one else has a hope of getting enough supply to release product (other than maybe Samsung)

 

[jdubs03]

Mediatek explicitly said they would have those parts would be available in late 2026
View attachment 136104

Hmm well that “late” wording is doing some heavy lifting there; and doesn’t really contradict a potential delay. Also, does that mean next year when Apple is on N2P, the Snapdragon Elite Gen 7 and Dimensity 9700 are on A16? It seems weird that Apple would consistently not be on the leashing edge node.

If Apple is delayed why would others be able to ship on time? If Apple is delayed due to N2 delays there's no chance N2P isn't also delayed. If Apple is somehow delayed due to RAM/NAND supply issues, how would anyone else not be delayed for the same reasons? And likely much worse because if Apple is delayed despite their ironclad prepaid contracts, no one else has a hope of getting enough supply to release product (other than maybe Samsung)

Sorry there was some ambiguity there. I meant other manufacturers delaying their typical release window to get N2P in them. I don’t think Apple will delay anything.

 

[Geddagod]

I'm personally very excited for 18A-P; much like Intel 4 was a janky version of Intel 3, once they get the fmax to where it's intended (and needs) to be, I imagine it will compete very well against N2, j

Following this logic, Intel 18A-P should be used in desktop for client for at least the gen after NVL, if not NVL itself, right?
There's really nothing else stopping it from being used this way.

I would be truly shocked if it's behind N3P in performance - Intel 3 is already around N3P in terms of fmax and outright performance.

The highest frequency Intel 3 part matches "ARL-U" matches ARL-H and is worse than ARL-S. Surely it can't be this shocking, especially looking at GNR ppw vs Turin iso core count too...

Hmm well that “late” wording is doing some heavy lifting there; and doesn’t really contradict a potential delay.

If the original schedule was that N2P would begin HVM in 2H 2026, it would make it pretty hard for those companies to be confident that they would be able to launch and have those chips in market also by late 2026. I don't think they make that claim if they think there will be a delay.

Also, does that mean next year when Apple is on N2P, the Snapdragon Elite Gen 7 and Dimensity 9700 are on A16?

I don't think A16 is going to be used by the smartphone crowd at all.

It seems weird that Apple would consistently not be on the leashing edge node.

I think Apple wanted TSMC to have N2 out by end of year 2025, but they were not able to enter HVM early enough in 2025 for that to happen. I think a similar thing happened with N3 as well.

 

[jdubs03]

If the original schedule was that N2P would begin HVM in 2H 2026, it would make it pretty hard for those companies to be confident that they would be able to launch and have those chips in market also by late 2026. I don't think they make that claim if they think there will be a delay.

I don't think A16 is going to be used by the smartphone crowd at all.

I think Apple wanted TSMC to have N2 out by end of year 2025, but they were not able to enter HVM early enough in 2025 for that to happen. I think a similar thing happened with N3 as well.

1. Fair enough, it’s just a bit bewildering that N2P follows so soon after N2 in its availability. I mean to have devices with N2P out only 1-3 months after the first N2 devices is a tad unconventional.

2. So that would imply that the S8EG7 and Dimensity 9700 are on N2P as well. Abrnormal. Because I’m quite sure that A14 is a 2028 part.

3. Fair.

 

[Fjodor2001]

dawg it's slower than N3p

And you think this because…?

 

[Doug S]

1. Fair enough, it’s just a bit bewildering that N2P follows so soon after N2 in its availability. I mean to have devices with N2P out only 1-3 months after the first N2 devices is a tad unconventional.

Apple's release of iPhone with N2 in September isn't because that's the earliest they could get them out, its because that's when they release new iPhones. They may well do what they did with N3B when they release M4 iPad Pro in May and release an M6 iPad Pro next spring. If the gap between N2 and N2P was only two months, but N2P was too late for the iPhone production cycle they'd use N2 rather than delay its launch to gain a mere 5% boost.

But that M6 iPad Pro scenario may not happen this time. Apple may "ship" many N2 devices before the next iPhone launch but all of them end up going to Apple datacenters rather than customer hands.

 

[Win2012R2]

They may well do what they did with N3B when they release M4 iPad Pro in May

That was N3E (lovely node)

But that M6 iPad Pro scenario may not happen this time.

It won't happen this year, they've just updated it in a minor way, they will wait till 2027

1. Fair enough, it’s just a bit bewildering that N2P follows so soon after N2 in its availability. I mean to have devices with N2P out only 1-3 months after the first N2 devices is a tad unconventional.

AI HPC, big money talks

 

[jdubs03]

It won't happen this year, they've just updated it in a minor way, they will wait till 2027

AI HPC, big money talks

1. I could see them releasing the M6 iPad Pro in fall. The rumor for the MacBook Pro redesign with M6 is late this year (though I always suspected early 2027)

2. That doesn’t apply to phones with Qualcomm and Mediatek flagships though.

 

[marees]

Qualcomm CEO Cristiano Amon met with reporters at CES 2026, which opened in Las Vegas on the 5th, and said, "Among many foundries, we have begun discussions with Samsung Electronics for outsourcing production using the latest 2-nanometer (nm, 1nm = 1 billionth of a meter) process. We have also completed the design work with the goal of commercialization soon." This is the first time Qualcomm has announced plans to outsource AP production using the 2nm process to Samsung Foundry.

If the contract is finalized, Samsung will resume production of Qualcomm's cutting-edge products after a five-year hiatus since 2022. Qualcomm contracted Samsung Foundry to produce its cutting-edge APs until 2021, but then shifted its focus to Taiwan's TSMC, citing Samsung's failure to meet Qualcomm's demanding requirements.

[ë¨ë] ì¼ì±ì ì íì´ë리 ìµì í컴칩 ë§ë ë¤

[ë¨ë] ì¼ì±ì ì íì´ë리 ìµì í컴칩 ë§ë ë¤, ë°ëì²´ INSIGHT íì»´ CEO "2ëë¸ì¹© 맡길 ê²" ì°¨ì¸ëAP ìì¡°ì ê³ì½ ì ë ¥

www-hankyung-com.translate.goog

https://twitter.com/x/status/2008834242516787562