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.

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.

 

[darkswordsman17]

Raja was AMD's unwilling Troyan Horse, he set back Intel GPU Division at least 5 years

Ah, so Intel is going to cancel GPUs 3 years from now then? (Assuming the first GPU was due out more like 2021)

I think Intel is just wasting time and resources (and will continue to do so by keeping with it) at consumer dGPU, which in turn wasted time and resources in enterprise (because they should have been focused solely there with much more focused designs). Heck, I think Intel would've been better off just focused on developing a pure ray-tracing accelerator.

Something I wish someone had the clout to do is call out the tech journalists, who I think are the reason that Raja was enabled. We've now seen he simply is just not very good at overseeing GPU divisions. His rep seemed built on him being friendly to tech journalists who fluffed him up and gave him celebrity status in the industry. Raja seems like a decent guy that just isn't suited for the positions he was being put in.

 

[Exist50]

But Raja was just "Retired" Right?

Yeah, that's new. But tbh, it was completely expected after they removed him from any management role. No one survives that kind of fall down the corporate ladder.

 

[Panino Manino]

Something I wish someone had the clout to do is call out the tech journalists, who I think are the reason that Raja was enabled. We've now seen he simply is just not very good at overseeing GPU divisions. His rep seemed built on him being friendly to tech journalists who fluffed him up and gave him celebrity status in the industry. Raja seems like a decent guy that just isn't suited for the positions he was being put in.

What? Are you saying that Intel only hired Raja because they read too much yellow journalism? Nonsense.
He may not be the best at what he does, but he is still someone that knows what he does.

 

[jpiniero]

Guess I missed, or just forgot about it.

My expectation is that Tower will run IFS basically if the deal goes through. Maybe through the spinoff if it comes to that.

 

[DrMrLordX]

Raja nooooooooooo

 

[lightmanek]

Yeah, that's new. But tbh, it was completely expected after they removed him from any management role. No one survives that kind of fall down the corporate ladder.

But the parties he thrown on the way down ... they must've been EPIC!

 

[Exist50]

Guess I missed, or just forgot about it.

Just linking an article for reference. Was announced in November.

Intel Foundry Services Chief Steps Down

Before the business unit's prime time.

www.tomshardware.com

Though I'm a bit surprised they didn't appoint someone external or from Tower to fill the role. Gelsinger seems to be heavily leaning on "old Intel" for upper management.

 

[jpiniero]

Just linking an article for reference. Was announced in November.

Intel Foundry Services Chief Steps Down

Before the business unit's prime time.

www.tomshardware.com

Though I'm a bit surprised they didn't appoint someone external or from Tower to fill the role. Gelsinger seems to be heavily leaning on "old Intel" for upper management.

Tower deal hasn't gone through yet IIRC.

 

[Exist50]

Tower deal hasn't gone through yet IIRC.

True, but they could have put Stu Pann (or someone else) in as an interim leader. Instead, he seems to be the permanent choice.

 

[naukkis]

What? Are you saying that Intel only hired Raja because they read too much yellow journalism? Nonsense.
He may not be the best at what he does, but he is still someone that knows what he does.

Didn't he left AMD because poor AMD didn't have resources to implement his visions? He then did go to Intel to design Ponte Vecchio...... It actually surprises that he didn't get booted sooner.

 

[igor_kavinski]

Didn't he left AMD because poor AMD didn't have resources to implement his visions?

AMD got rid of him because his visions were gonna bankrupt them!

 

[Panino Manino]

Intel is the Google of semiconductor.

 

[moinmoin]

Intel is the Google of semiconductor.

Does an Intel counterpart to the Google Graveyard site already exist? Would be nice to have indeed.

 

[Ajay]

Samsung is good at announcing new nodes and making it sound like they are even with TSMC, but they are terrible at delivering working processes ever since EUV arrived.

My off the cuff assessment is that using Samsung’s fabs, chipmakers can get good electrostatics, good density or good yields - pick two. It’s not the kind of trade off most chipmakers want. There must be more to it, otherwise I would think they’d have more customers on older nodes whose needs aren’t as demanding.

 

[NostaSeronx]

2021-2022 GlobalFoundries Fab8.2 plan was for FDX/HV:

(Slide timestamped for June 2022, but apparently Fab8.2 being FDSOI has been roaming around since the IPO.)

However, October 2022 to present, I have been whispered that Fab8.1 will be doing FDSOI sooner rather than later.

First billing: June/July 2023, Production Range: 2H2023 -> 1H2024.
Products for 22FDX(+):
- Automotive(1st tier/majority of production), AI/Blockchain, and Industrial APs.
Products for 12FDX:
- Datacenter MPU, APUs, SoCs(1st tier/majority of production), plus the above for premium customers.

with CV's pointed towards:
Integration/Yield-up/NPI for 12FDX at Malta => April-June 2022
Integration/Yield-up/NPI for 22FDX at Malta => October-December 2022

There is also suppose to be announcement late 2H2023 to early 1H2024 iterating on the 7nm FDSOI node for Crolles/Malta 8.2.

12FDX_FEOL is derived from 10FD and has its own BEOL. While 7nm FDSOI should be more closely a direct rip; 7nm FDSOI dev. alliance is currently: CEA, Soitec, GlobalFoundries and STMicroelectronics.

With the 7nm FDSOI node being slated for Fab8.2 and Crolles(later phases), both are ETA'd to 2025+ for 7nm FDSOI.

 

[DrMrLordX]

Does an Intel counterpart to the Google Graveyard site already exist? Would be nice to have indeed.

Wish we had pics of the disco ball.

 

[jpiniero]

True, but they could have put Stu Pann (or someone else) in as an interim leader. Instead, he seems to be the permanent choice.

Or maybe they think China won't approve of the deal so it might be dead.

Intel's Acquisition of Tower Delayed, Still Expected to Close by June 2023

Intel still hopes to close the deal with Tower by end of June.

www.tomshardware.com

 

[mderbarimdiger]

Intel Israeli division is still talking acquisition plans.

 

[A///]

Intel Israel was Intel's clutch division that saved their asses many times. They also happen to be the ones who dropped the ball when it came to security woes. Raja is such an amateur. A good cigar dancer can keep the cigar in their mouth while juggling a drink or two in their hands and not splashing it on their lady friend.

 

[dark zero]

Or maybe they think China won't approve of the deal so it might be dead.

Intel's Acquisition of Tower Delayed, Still Expected to Close by June 2023

Intel still hopes to close the deal with Tower by end of June.

www.tomshardware.com

With the current political situation I don't be surprised if they starts to NOT to care on China's desicions, even if it affects the bussiness in a future

 

[Saylick]

Wasn't sure which thread made the most sense to share this, but I thought this Wired magazine article about TSMC was really intriguing, especially as someone who considers themselves an atheist.

https://www.wired.com/story/i-saw-the-face-of-god-in-a-tsmc-factory/

 

[jpiniero]

With the current political situation I don't be surprised if they starts to NOT to care on China's desicions, even if it affects the bussiness in a future

Things may get to that point but we are nowhere near that kind of situation.

 

[maddie]

Things may get to that point but we are nowhere near that kind of situation.

Things will never get to that point in a sane world, only fools will do that. The global trends are irreversible and clear, however distasteful it is to some.

Having said that, we do have a lot of fools loose, so who knows.

 

[DisEnchantment]

Having said that, we do have a lot of fools loose, so who knows.

Its ~35% of Intel revenue at stake which AMD will happily pick up, they wont risk getting kicked out of that market for non compliance.

 

[Ajay]

Or maybe they think China won't approve of the deal so it might be dead.

Intel's Acquisition of Tower Delayed, Still Expected to Close by June 2023

Intel still hopes to close the deal with Tower by end of June.

www.tomshardware.com

Hmm, why does China have to approve? I didn’t think Tower had any fabs in China.