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

 

[adroc_thurston]

blatantly false claim outside of couple of nodes .....

Where are the customers broski

 

[Hans Gruber]

Do the new desktop CPU's on 18A start with Panther Lake or will the circus start with Nova Lake?

 

[adroc_thurston]

Do the new desktop CPU's on 18A start with Panther Lake or will the circus start with Nova Lake?

There are no desktop CPUs on 18A.

 

[511]

Where are the customers broski

Intel is big enough customer lmao and even if customer will sign for 18AP/14A we won't see their products before 28-29 at least

 

[adroc_thurston]

Intel is big enough customer lmao

Would've been true a decade ago.
Now? Lmao.

and even if customer will sign for 18AP/14A we won't see their products before 28-29 at least

If 18A didn't suck you would've had customer sign ons a while ago with products ramping H2'26.

 

[511]

If 18A didn't suck you would've had customer sign ons a while ago with products ramping H2'26.

It's the PDK that sucks which is getting fixed with 18AP alongside process improvement

Would've been true a decade ago.
Now? Lmao.

Even Now they consume large amount of wafers but that's still nowhere enough to fill fabs

 

[adroc_thurston]

It's the PDK that sucks which is getting fixed with 18AP alongside process improvement

PDK is like, not a separate thing from the process node. If your PDK sucks, your node sucks also.

Even Now they consume large amount of wafers but that's still nowhere enough to fill fabs

Intel as a business can no longer sustain bleeding edge fab capex.

 

[DZero]

They are still going to release the Pro and Pro Max.

Apparently there will be a delay, we will return to yearly launch, so expect being it launched by november instead of September, similar to the rest of phones to get delayed between 2 or 3 months.

 

[511]

PDK is like, not a separate thing from the process node. If your PDK sucks, your node sucks also.

Yes but that's true for all Intel nodes for external you mean to tell me that me From 90 to 45 to 32 to 22 to 14nm cause i can gurantee no one can use those nodes beside Intel

Intel as a business can no longer sustain bleeding edge fab capex.

I agree they don't have the money to sustain Capex but they won't go out of nodes cause they have US government to fund it and US will fund it cause "national security".

 

[Geddagod]

IIRC, he is one of the large contingent of INTC fan club on Twitter (and SemiWiki), who twist every news into "this means that Intel Foundry is definitely getting huge orders".

Or, "there is no other possible outcome as a result of XYZ news than companies flocking to Intel Foundry"

Too many of those nowadays due to Intel stock owners needing hopium.

 

[jdubs03]

Apparently there will be a delay, we will return to yearly launch, so expect being it launched by november instead of September, similar to the rest of phones to get delayed between 2 or 3 months.

According to whom? I haven’t seen this reported anywhere. This sounds illegitimate.

 

[jdubs03]

Intel as a business can no longer sustain bleeding edge fab capex.

Let’s hope that changes.

Purchases of high-NA are conflicting in the narrative that they can’t invest in bleeding edge. In addition, the new D1X Mod 4 complex starting construction soon.

 

[dangerman1337]

Let’s hope that changes.

Purchases of high-NA are conflicting in the narrative that they can’t invest in bleeding edge. In addition, the new D1X Mod 4 complex starting construction soon.

AFAIK they're installing them so I think they've gotten strong interest with 14A from external? I don't know why Intel would be installing them when LBT said if there was no interest/big orders they'd can 14A which depends on High NA EUV?

 

[Doug S]

They are still going to release the Pro and Pro Max.

Yes, there have been rumors brewing for almost a year that Apple would release iPhone 18 Pro / Pro Max and "fold" next fall at the usual time (maybe "fold" is delayed a bit) but wait until the next spring to release the base iPhone 18, iPhone 18e, and Air. The claim was that Apple was doing it to smooth out manufacturing over the year rather than have a surge in summer/fall then a comparative lull in winter/spring. i.e. that this would not be a one time thing but a change to their release schedule going forward. That change in production would make things easier for Foxconn to manage, especially with Apple expanding more manufacturing outside of China where getting that surge in temporary employees could be more difficult.

I've seen it reported a lot but it still isn't clear if this is one rumor in an echo chamber or multiple independent confirmations.

 

[DZero]

Yes, there have been rumors brewing for almost a year that Apple would release iPhone 18 Pro / Pro Max and "fold" next fall at the usual time (maybe "fold" is delayed a bit) but wait until the next spring to release the base iPhone 18, iPhone 18e, and Air. The claim was that Apple was doing it to smooth out manufacturing over the year rather than have a surge in summer/fall then a comparative lull in winter/spring. i.e. that this would not be a one time thing but a change to their release schedule going forward. That change in production would make things easier for Foxconn to manage, especially with Apple expanding more manufacturing outside of China where getting that surge in temporary employees could be more difficult.

I've seen it reported a lot but it still isn't clear if this is one rumor in an echo chamber or multiple independent confirmations.

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.

 

[Joe NYC]

AFAIK they're installing them so I think they've gotten strong interest with 14A from external? I don't know why Intel would be installing them when LBT said if there was no interest/big orders they'd can 14A which depends on High NA EUV?

If they are installing them (High NA machines) it is because Gelsinger ordered them some 3 years ago, not because LBT got a phone call from a potential customer 3 weeks ago.

 

[poke01]

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

 

[eriksp92]

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, just as Intel 3 actually compares very well to N3.

 

[511]

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, just as Intel 3 actually compares very well to N3.

Intel 3 is decentish behind PPA vs N3B maybe slightly behind N3B in perf/watt and i would expect 18AP to be slightly behind N2 in perf/watt or maybe match it

 

[eriksp92]

I would be very surprised if it's behind N3B in PPW. The density is AFAIK slightly higher if comparing Intel 3 HP libraries to N3B HP libraries (and further revisions of N3 relaxed the density even further, no?), but N3B wins handily in HD library density. N3B is also lower leakage at low power.

 

[511]

I would be very surprised if it's behind N3B in PPW. The density is AFAIK slightly higher if comparing Intel 3 HP libraries to N3B HP libraries (and further revisions of N3 relaxed the density even further, no), but N3B wins handily in HD library density.

That's the problem outside Intel HP lib is used sparingly usually a mix of HP/HD

 

[eriksp92]

Yeah, Finflex is a definite advantage for TSMC, as is their lower leakage at low power. But to match Intel 3's lack of vdroop at high voltage and fmax, they had to relax their densities, and one of 18A's BSPD's goals was to remedy this if I remember correctly. TSMC will still be ahead because of their experience in dealing with external customers, and I'm unsure what Intel has to combat TSMC's Finflex, but if 18A-P sees the same boost as Intel 4->Intel 3, I think 18A-P will compare very well with N2.

 

[adroc_thurston]

I think 18A-P will compare very well with N2.

dawg it's slower than N3p

 

[eriksp92]

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.

 

[adroc_thurston]

Intel 3 is already around N3P in terms of fmax and outright performance.

lmao dawg it's slower than N4.