Discussion Leading Edge Foundry Node advances (TSMC, Samsung Foundry, Intel)

[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.

TSMC Radically Boosts CapEx to Expand Production Capacities, To Reach $14B For 2019

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.

Samsung’s Aggressive EUV Plans: 6nm Production in H2, 5nm & 4nm On Track

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.

 

[DrMrLordX]

Isn't their biggest customer Samsung, anyway?

I'm not actually sure how to parse that statement, since it can be taken in so many different ways. Can you elabourate?

 

[NTMBK]

I'm not actually sure how to parse that statement, since it can be taken in so many different ways. Can you elabourate?

Samsung is the biggest customer of Qualcomm leading-edge SoCs.

 

[Rannar]

Samsung is the biggest customer of Qualcomm leading-edge SoCs.

That + Samsung own Exynos series SoCs produced in their own fabs.

 

[DrMrLordX]

Samsung is the biggest customer of Qualcomm leading-edge SoCs.

Oh, okay. That's possible. I haven't tracked how many different phones/other products use Snapdragons, but S-series Samsung phones certainly do have Snapdragon options. In that light, Samsung certainly is shooting themselves in the foot.

 

[JoeRambo]

Is this even big deal at all? Apple is sure happy, but Android users are used to SoCs that are like half performance of latest Apple stuff.

 

[moinmoin]

Is this even big deal at all?

Seems like MediaTek is turning into a serious contender at the high end due to this.

 

[Ajay]

Is this even big deal at all? Apple is sure happy, but Android users are used to SoCs that are like half performance of latest Apple stuff.

IDK. Phones are a weird market. Apple and Samsung are making upwards of 90% of the profits in that market. (Last I checked). Other than their CPUs, Apple makes bank on it’s premium style, and their ecosystem. Samsung has a weaker CPU, but makes bank on equally impressive specs in others areas (like cameras) and compatibility with the rest of the known universe ( ). So, both approaches work.

PS - Also, Apple runs allot of ads!

 

[Kenmitch]

PS - Also, Apple runs allot of ads!

Some pretty dumb looking one's to note. Not sure what's up with that?

 

[JoeRambo]

Seems like MediaTek is turning into a serious contender at the high end due to this.

Yeah, forgot about them. They stand to benefit the most from Samsung delays. And honestly in TSMC process "volume" pie charts their share is getting scary real fast.

 

[Doug S]

I don't get it. Where in the article or the quotes @DisEnchantment posted does it say anything about 3GAE being either gutted OR delayed? The addition of 4LPE doesn't mean 3GAE's status has changed, unless somewhere in the subscribers only part of the article they are separately claiming that.

I assume that the 4LPE/4LPP nodes serve the same purpose for Samsung as the N4 node for TSMC - a refinement of their 5nm not a completely separate process, and thus would not influence when 3GAE arrives.

We've been hearing "H2 2022" for N3 for TSMC pretty consistently over the past year and don't yet know what they mean, since they typically nail that down to a quarter than a half. Does it mean initial volume shipments of N3 wafers in July/August, just in time for iPhones to ship with N3 SoCs in fall 2022? Or is it toward the end of the year, thus forcing Apple to instead use N4 next year? If that happened it might free up some of the early N3 capacity that Apple would otherwise be gobbling up.

So it is possible TSMC may have a 'delay' of its own at 3nm, though IMHO given how tightly TSMC and Apple cooperate on process rollouts my money would be on "later than TSMC and Apple would prefer, but still in time for 'iPhone 14' September shipments". Referring to H2 rather than Q3 may be some element of hedging their bets - maybe they aren't quite as confident with N3's status as they were with N5 and N7 at this stage so they don't want to publicly commit to something they aren't 100% sure they will be able to deliver.

 

[DrMrLordX]

I don't get it. Where in the article or the quotes @DisEnchantment posted does it say anything about 3GAE being either gutted OR delayed?

If you compare the 2019 projections for 3GAE and the 2021 reality, they don't exactly line up. By a long shot.

 

[JasonLD]

If you compare the 2019 projections for 3GAE and the 2021 reality, they don't exactly line up. By a long shot.

Hard to say that is confirmed though. Samsung Foundry site still goes by the original 2019 projections on 3GAE. Being nerfed or not, I really doubt it will be worse than 4LPP.

 

[DrMrLordX]

Hard to say that is confirmed though. Samsung Foundry site still goes by the original 2019 projections on 3GAE. Being nerfed or not, I really doubt it will be worse than 4LPP.

Hmm. Do they contradict themselves? Strange.

 

[DisEnchantment]

Samsung 3GAE gutted and delayed, 3GAE could be behind 4LPE it looks like.

Seems official now for foundry partners, 3GAE gone, 3GAP is here,


From here

三星最新消息。3nm级暂定一个节点:3GAP【高通吧】_百度贴吧

三星最新消息。3nm..今天下午新鲜出炉。来源张国斌的微博。另外三星改了原来的PPT，性能/面积/功耗的实例都是用的A57，不再显示A75了。而我们都知道A57最多只能跑2.2~2.3的样子(现实中最高频是三星7420的2.

tieba.baidu.com

 

[DrMrLordX]

So what's the difference between 3GAP and 3GAE?

 

[DisEnchantment]

So what's the difference between 3GAP and 3GAE?

The difference is losing face.
I bet it is the same thing tweaked to fix the issues and renamed accordingly and then marketed as something new.

 

[maddie]

So what's the difference between 3GAP and 3GAE?

One is the 5th attempt and the other is the 16th. A silicon WD40.

 

[DisEnchantment]

TSMC 2021 Foundry Update: Foundry Roadmap

TSMC 2021 foundry update

fuse.wikichip.org

New updates from TSMC.
I won't copy the slides or detailed info, please subscribe and read there.

In 2021 there will only be a minor additional capacity increase in N7.
In 2022 the capacity for N5 will be more than 3.5x than that from last year. So lots to go around. N5P (HP) scaling only 1.38x big letdown. N5 HD 1.8x for logic.
In 2H22 N3 will come online for the likes of Apple probably. But not so interesting for HPC. Analog and SRAM barely any scaling.



4LPE/5LPP should get closer to N5. I think Samsung will get to fab a bunch of mobile SoCs.

 

[DrMrLordX]

The difference is losing face.
I bet it is the same thing tweaked to fix the issues and renamed accordingly and then marketed as something new.

So 3GAP may get closer to the expected density/performance targets than Samsung projected for 3GAE two years ago?

 

[JoeRambo]

In 2022 the capacity for N5 will be more than 3.5x than that from last year. So lots to go around. N5P (HP) scaling only 1.38x big letdown. N5 HD 1.8x for logic.
In 2H22 N3 will come online for the likes of Apple probably. But not so interesting for HPC. Analog and SRAM barely any scaling.

End of road for FinFET? Just like planar died on 20nm, it seems that GAA is needed to push forward.

 

[Thibsie]

FinFET has been relatively short lived or is it my memory failing ?

 

[jpiniero]

FinFET has been relatively short lived or is it my memory failing ?

Broadwell Core M was Q3 2014. So 7 years?

 

[Bigos]

Broadwell Core M was Q3 2014. So 7 years?

FinFET started with 22nm and Ivy Bridge, so Q2 2012 and 9 years.

Ivy Bridge (microarchitecture) - Wikipedia

en.wikipedia.org

 

[Thibsie]

Alright, older than I thought, then.
Thanks.

 

[Ajay]

End of road for FinFET? Just like planar died on 20nm, it seems that GAA is needed to push forward.

Not yet, IIRC, TSMC 3N is still FinFET. And plenty of devices; SoCs, microcontrollers, FPGA and ASICs will still be built in FinFET node for some time to come.