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

 

[511]

yes but they get the funding rn looks like 14A statement was for this

 

[coercitiv]

Idk... I feel like money wouldn't be a problem if they didn't have technical issues with their nodes.

Root cause and symptoms. Sometimes the symptoms need to be addressed first. One does not learn how to thread water while drowning.

You're right though int he sense that money just buys them time, not a solution.

 

[DrMrLordX]

Isn’t everything funded by the already awarded grants? Like the gov got it for basically for agreeing to follow up to give what it already promised...

Yes, but I think the government wasn’t going to be actually giving that money anymore unless Intel agreed to this.

It's not entirely clear. Intel pushed hard for CHIPS money, only to be met with a lot of um, how do you put it, interesting requirements from the previous administration which it appears that Intel never quite met, because the actual distribution of funds to Intel was delayed. By quite a bit. Then the current administration showed up and wiped out some DoE personnel who were (allegedly) responsible for administering the CHIPS grants which (allegedly) prevented anyone from receiving more than the small amount of CHIPS money that had dribbled out during the previous administration. So there was essentially a pot of $10 billion sitting there waiting to go to Intel without any processes in place to expedite the actual distribution of funds. Intel's agreement to hand the government shares somehow magically broke loose the multi-year, multi-administration logjam and finally got Federal money into Intel's pockets.

Or at least that's my take on it.

Hmm, money looks to be Intel's largest problem at the moment. The one time impairments have not been as one time as Intel claimed.

The real question is: where does this money go, and whom does it help? Intel is still one company (for now), but CHIPS was mostly directed at sustaining American leadership in silicon manufacturing. The money should be going to IFS 2.0, should it not? But what if Intel uses it to shore up their books to compensate for problems on the design side?

 

[511]

The real question is: where does this money go, and whom does it help? Intel is still one company (for now), but CHIPS was mostly directed at sustaining American leadership in silicon manufacturing. The money should be going to IFS 2.0, should it not? But what if Intel uses it to shore up their books to compensate for problems on the design side?

isn't it the design side that is keeping the fab alive and vice versa so both part of the company did what was required to sustain manufacturing so it's not wrong for both of them getting the funding. They just need to continue manufacturing.

anyway now from this and bunch of other stuff 14A is going to happen

 

[marees]

But what if Intel uses it to shore up their books to compensate for problems on the design side?

Imo, The lunar lake chip was perfectly fine. So I don't think design is the issue. I am afraid fab is a major issue. They desperately need anchor clients for 14A to justify huge capital investments needed

 

[DrMrLordX]

isn't it the design side that is keeping the fab alive and vice versa so both part of the company did what was required to sustain manufacturing so it's not wrong for both of them getting the funding.

The designers moved to N3B at considerable expense and still screwed it up, so not really sure on that one. Obviously the fabs have problems, but they aren't the only ones. Also the job of the designers in an IDM + foundry model is to promote the IFS nodes that actually have PDKs with internal products that are meant to make Intel's IP portfolio and nodes look good to prospective customers. Meteor Lake did a poor job of that for Intel 4 (which wasn't really a foundry node anyway, except for Ericcson), Arrow Lake-U ain't getting it done for Intel 3, and . . . where are the 18A chips? Oh that's right, delayed until later.

Imo, The lunar lake chip was perfectly fine. So I don't think design is the issue.

Mateor Lake wasn't, Arrow Lake wasn't. Panther Lake is looking iffy. It's about salesmanship and using designs and the IDM model to promote the nodes so that foundry customers will want access to the same thing for their designs. Which is something that is not going well for Intel. Plus there is a volume issue on everything EUV.

 

[511]

Mateor Lake wasn't, Arrow Lake wasn't. Panther Lake is looking iffy.

Right on ARL/MTL but PTL ain't out yet so it's hard to say about that one.

The designers moved to N3B at considerable expense and still screwed it up, so not really sure on that one.

yes but that was in 2020 though and 7nm and nodes beyond that had very bad outlook when they moved to N3B and all other stuff to TSMC to not loose market share but Intel design screwed up

Arrow Lake-U ain't getting it done for Intel 3, and . . . where are the 18A chips? Oh that's right, delayed until later.

ARL-U has more tiles from TSMC than IFS loosing volume hurts for a fab be it a small tile unless it's cheaper to outsource the only node that matches that description is N6 even than this means their fabs gets to sit empty also advanced packing doesn't help either as for 18A chips it is yet to enter HVM also PTL QS is yet to be finalized.

 

[DrMrLordX]

ARL-U has more tiles from TSMC than IFS

While true, the critical compute tiles are on Intel 3. Also the interconnect designs are holding everything else back, which is the crux of why Intel's design side is not being a good cheerleader for the fabs.

Regardless

If the Feds want to bolster American silicon manufacturing, the money should go to IFS to help it compete with TSMC rather than going towards helping Intel's design team compete against other designers like AMD, Qualcomm, and Apple.

 

[511]

While true, the critical compute tiles are on Intel 3

40mm2 out of like 160-170mm2.
For second part force Ohio to be production with 14A.

 

[511]

https://twitter.com/x/status/1959004733336682795

Is this true has anyone seen this confirmed somewhere cause this means even if Intel imports chips from TSMC no tariffs 😮

 

[DrMrLordX]

40mm2 out of like 160-170mm2.

That is missing the point. No other part of the package hints at the potential performance and power characteristics of an Intel node. Arrow Lake-U, Granite Rapids, and Clearwater Forest are all cheerleaders for Intel 3, and if there were a 20a variant of Arrow Lake then it would be a cheerleader for 20a (sadly that never happened).

 

[511]

That is missing the point. No other part of the package hints at the potential performance and power characteristics of an Intel node. Arrow Lake-U, Granite Rapids, and Clearwater Forest are all cheerleaders for Intel 3, and if there were a 20a variant of Arrow Lake then it would be a cheerleader for 20a (sadly that never happened).

Yeah i am just looking at PTL how it turns out to be

 

[Hitman928]

https://twitter.com/x/status/1959004733336682795

Is this true has anyone seen this confirmed somewhere cause this means even if Intel imports chips from TSMC no tariffs 😮

I haven't seen that mentioned anywhere. He's probably just assuming either that this will give Intel the money they need to get the fab side good enough that they don't need to use TSMC anymore or that the government won't want to tariff Intel while owning a large piece of the company. I don't think either is a safe assumption though.

 

[511]

I haven't seen that mentioned anywhere. He's probably just assuming either that this will give Intel the money they need to get the fab side good enough that they don't need to use TSMC anymore or that the government won't want to tariff Intel while owning a large piece of the company. I don't think either is a safe assumption though.

Government is the largest shareholder now though

 

[Josh128]

https://twitter.com/x/status/1959004733336682795

Is this true has anyone seen this confirmed somewhere cause this means even if Intel imports chips from TSMC no tariffs 😮

100% lawsuit from Nvidia and AMD incoming if thats true....and I hope they win 'em.

 

[gdansk]

100% lawsuit from Nvidia and AMD incoming if thats true....and I hope they win 'em.

On what basis? The government should give us money too even though we aren't manufacturers?

In a more logical world Intel would be split to "solve" that problem but I don't think any suit stands a chance given the record on past government investments.

Intel might have the best chance to sue because the government isn't upholding the terms of the CHIPS Act but they're not going to bite the hand that feeds.

 

[DZero]

https://twitter.com/x/status/1959004733336682795

Is this true has anyone seen this confirmed somewhere cause this means even if Intel imports chips from TSMC no tariffs 😮

I don't think so... maybe it will be an one off, but then it must be 100% US made.

On what basis? The government should give us money too even though we aren't manufacturers?

In a more logical world Intel would be split to "solve" that problem but I don't think any suit stands a chance given the record on past government investments.

Intel might have the best chance to sue because the government isn't upholding the terms of the CHIPS Act but they're not going to bite the hand that feeds.

Dumping.
Literally what China does. And Europe or even China itself might react even harder now.

 

[gdansk]

Dumping.
Literally what China does. And Europe or even China itself might react even harder now.

Haha. Hah... No. An $11 billion investment is nowhere near comparable to the good and logical investments conducted elsewhere.

Intel was getting this money one way or another and it was known since the CHIPS Act was discussed. Nvidia has done nothing to oppose it. And AMD is far less litigious than forum goers here suggest.

 

[511]

Intel was getting this money one way or another and it was known since the CHIPS Act was discussed. Nvidia has done nothing to oppose it. And AMD is far less litigious than forum goers here suggest.

How can fabless company sue the only US Manufacturer of leading edge chips receiving Chips grants even if they tried to sue I am 100% sure they will loose the case.

100% lawsuit from Nvidia and AMD incoming if thats true....and I hope they win 'em.

They won't if they want to win they better start a leading edge foundry .
Intel already invested $100 Billion in actual manufacturing what did AMD/Nvidia did they outsourced more to TSMC how do you think they will make the lawsuit.

 

[Hitman928]

Government is the largest shareholder now though

And?

 

[511]

And?

I don't how they are going to act they say non voting but how random this admin is I can't fathom what they would do to boost equity 🤣.