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

 

[DrMrLordX]

I have no doubts that they will get there. They may go a different route than ASML took, but they'll get there.

They're free to spend as much money as they like getting there. The more, the merrier!

 

[NostaSeronx]

Canon Nanoimprint tool starts tool mass production in September 2025 with full capacity of tool shipping expected by 2027.

Only known official pre-orders are for NAND and DRAM. Any Logic node potential will be stuck on mature/low-volume till at least 2027~2028. When 25-nm metal pitch should be validated.

 

[dangerman1337]

I don't think China will get anything like High NA EVU any time soon but even just being able to make equivalent to TSMC N3 nodes but way more affordably would be hugely consequential; imagine $200 Smartphones with the power of a A18 Pro SoC? Would be utterly crazy and transformative across the world.

 

[511]

I don't think China will get anything like High NA EVU any time soon but even just being able to make equivalent to TSMC N3 nodes but way more affordably would be hugely consequential; imagine $200 Smartphones with the power of a A18 Pro SoC? Would be utterly crazy and transformative across the world.

There is only so much you can chug for the cost it ain't possible to make it that cheap.

 

[fastandfurious6]

China 🤝 TSMC

not too far away.

Good read on the history of EUV research https://cset.georgetown.edu/wp-cont...rgence-of-Extreme-Ultraviolet-Lithography.pdf

 

[DZero]

Ehh. We haven't had any smoking gun moments like when Broadcom publicly declared 18A to be inadequate for their needs. 14A is still too far off for us to know anything about it.


SMIC can't sub in for TSMC yet, and possibly never will. There's only so much they can do.

Never say never. Also TSMC would leave chinese companies in the air if the wants and if US pressures enough.

I don't think China will get anything like High NA EVU any time soon but even just being able to make equivalent to TSMC N3 nodes but way more affordably would be hugely consequential; imagine $200 Smartphones with the power of a A18 Pro SoC? Would be utterly crazy and transformative across the world.

Of course, the only good bet is releasing their own DUV tech

[News] China’s SiCarrier’s Bold Roadmap: ASML-Compatible Tools Reportedly Coming in 2026

With SiCarrier stealing the show at SEMICON China, the company has disclosed more details on its product roadmap. The Chinese chip equipment firm clai...

www.trendforce.com

With that, they need to improve to the max, which is 14nm without multiple partnering and with multiple partnering can reach up to 4 nm

 

[Josh128]

I have no doubts that they will get there. They may go a different route than ASML took, but they'll get there.

Yep, people that dont understand this dont understand the mindset and focus of China and the CCP. They have tremendous resources, tremendous focus and discipline, and a tremendous amount of people to choose from to make this happen. This "not selling AI" hardware and "not selling lithography machines" to them to stifle their progress is a fools errand. Huang is right. Sell to them like anyone else and develop a more cordial trade relationship / partnership rather than becoming more isolationist and sowing animosity. Because when they do inevitably catch up, you now have a hostile rival vs merely a competitive friend / partner.

 

[johnsonwax]

Never say never. Also TSMC would leave chinese companies in the air if the wants and if US pressures enough.

America's ability to dictate these things is in pretty steep decline. We'll see if that holds.

 

[Win2012R2]

Ehh. We haven't had any smoking gun moments like when Broadcom publicly declared 18A to be inadequate for their needs. 14A is still too far off for us to know anything about it.

Here it is:

"If we are unable to secure a significant external customer and meet important customer milestones for Intel 14A, we face the prospect that it will not be economical to develop and manufacture Intel 14A and successor leading-edge nodes on a go-forward basis.

"In such event, we may pause or discontinue our pursuit of Intel 14A and successor nodes and various of our manufacturing expansion projects.""

Source: https://www.intc.com/filings-reports/all-sec-filings##document-5862-0000050863-25-000109-2

Nope and Nope to both Raichu leaded Parametric Yield already

And that's why they use N2 for main core IP - their CPU? Actions speak louder than words, clearly 18A isn't as good as N2, internally Intel would know this for sure because they get both sets of data.

 

[511]

Here it is:

"If we are unable to secure a significant external customer and meet important customer milestones for Intel 14A, we face the prospect that it will not be economical to develop and manufacture Intel 14A and successor leading-edge nodes on a go-forward basis.

"In such event, we may pause or discontinue our pursuit of Intel 14A and successor nodes and various of our manufacturing expansion projects.""

Source: https://www.intc.com/filings-reports/all-sec-filings##document-5862-0000050863-25-000109-2



And that's why they use N2 for main core IP - their CPU? Actions speak louder than words, clearly 18A isn't as good as N2, internally Intel would know this for sure because they get both sets of data.

This was in SEC Filings and it's entire purpose was to get money for Fabs and to scare the government as well.

 

[Win2012R2]

This was in SEC Filings and it's entire purpose was to get money for Fabs

And give 10% of the company in process or that bit wasn't part of the plan to get CHIPs money from the Act which was basically designed for Intel in the first place?

Either way lies on SEC disclosures like this is a sure way for execs to end up in jail.

 

[511]

And give 10% of the company in process or that bit wasn't part of the plan to get CHIPs money from the Act which was basically designed for Intel in the first place?

Which has String attached and it would take time they need immediate money you forget the scaring government part and believe it or not it worked they got $2B from Softbank as well.
Chips Act money was drop in bucket compared to INTC Investment they only got $2 Billion for investing $90-100 Billion.

 

[Win2012R2]

they got $2B from Softbank as well

How much a 14A fab costs?

Intel lost $2.9 bln last quarter alone, even $10 bln is spare change, which they planned on getting anyway - this does not change anything if 18A is not a good process.

 

[511]

How much a 14A fab costs?

Ohio ones are $28 Billions each

 

[Win2012R2]

Ohio ones are $28 Billions each

TSMC's A14 is quoted at $49 bln, not sure how it compares size wise though, but either way $2 bln and even $10 bln is nothing now.

 

[511]

TSMC's A14 is quoted at $49 bln, not sure how it compares size wise though, but either way $2 bln and even $10 bln is nothing now.

If we assume that building fab is US is about 10% expensive than 2 Fabs in Ohio combined($56 Billion) are in similar ballpark to $53.9B ($49 Billion*1.1) .

 

[DavidC1]

The money won't matter if they can't get EUV up and running.

You are thinking the progress is limited by technology.

The word "technology" is created by us humans. It's the other way around. Smart people create the technology required to progress. China has TEN times the startups and graduates in computer science compared to US. It's not even close. Not to mention the general work ethics are better because of complacency in US. Like that woman that climbed mountains everyday to go to school so she can get to university.

Also, this difference will accelerate if they become adversaries to each other.

 

[DrMrLordX]

Yep, people that dont understand this dont understand the mindset and focus of China and the CCP. They have tremendous resources, tremendous focus and discipline, and a tremendous amount of people to choose from to make this happen.

Sorry, but China/the CCP are looking more and more like a tofu dreg tiger as the years go by. Things aren't all well there.

"If we are unable to secure a significant external customer and meet important customer milestones for Intel 14A, we face the prospect that it will not be economical to develop and manufacture Intel 14A and successor leading-edge nodes on a go-forward basis.

"In such event, we may pause or discontinue our pursuit of Intel 14A and successor nodes and various of our manufacturing expansion projects.""

That says "if", not "when". What it means is, IF external customers turn up their noses at 14A, then Intel's goose is cooked, that's it, say gnight Gracie. They can't have another delayed/ignored node like 18A. It will kill the company.

What I'm talking about is more like this article:

https://www.reuters.com/technology/intel-manufacturing-business-suffers-setback-broadcom-tests-disappoint-sources-2024-09-04/

(as others have observed, it's often Reuters carrying these stories, but still it was a bad look for Intel)

You are thinking the progress is limited by technology.

The word "technology" is created by us humans. It's the other way around. Smart people create the technology required to progress. China has TEN times the startups and graduates in computer science compared to US. It's not even close.

And? You can throw infinite manpower and money at a problem and fail. China has done a LOT of that lately.

Not to mention the general work ethics are better because of complacency in US.

Oh please. It isn't American fabs they've got to catch up to! It's TSMC!

Also, this difference will accelerate if they become adversaries to each other.

Isolation doesn't lead to innovation.

 

[DavidC1]

And? You can throw infinite manpower and money at a problem and fail. China has done a LOT of that lately.

It's not "throwing" manpower. Out of greater number of people advances happen. One could even say modern advances were enabled by record number of people.

Intel was also an undisputed leader until they weren't. If you are talking about the worst case scenario where they become adversaries, TSMC will be irrelevant because they'll be annexed/destroyed.

 

[DavidC1]

Oh please. It isn't American fabs they've got to catch up to! It's TSMC!

Intel was the undisputed leader until they weren't. And in a scenario where countries become adversaries TSMC will be gone.

Isolation doesn't lead to innovation.

The world will be split in half in that scenario. US and Taiwan will be isolated just as China is.

 

[johnsonwax]

You are thinking the progress is limited by technology.

The word "technology" is created by us humans. It's the other way around. Smart people create the technology required to progress. China has TEN times the startups and graduates in computer science compared to US. It's not even close. Not to mention the general work ethics are better because of complacency in US. Like that woman that climbed mountains everyday to go to school so she can get to university.

Also, this difference will accelerate if they become adversaries to each other.

China has a lot of startups but they are all starved of cash. It's not an easy place to build capital intensive businesses from startup. The US is the best in the world in that regard. And China's graduate education programs are not comparable to the US yet. They're getting there but they have a ways to go. It takes a lot of infrastructure and coordination to build these kinds of cutting-edge technology R&D programs. China may get there, but their inability to bring in talent from around the world is a big problem. That was the main strength of the US until this year. We'll see if we can get it back later.

 

[511]

What I'm talking about is more like this article:

https://www.reuters.com/technology/intel-manufacturing-business-suffers-setback-broadcom-tests-disappoint-sources-2024-09-04/
(as others have observed, it's often Reuters carrying these stories, but still it was a bad look for Intel)

It's 2024-09 Months away from HVM and it's reuturs I will not believe Reuturs they have been dishing out fake news.

 

[fastandfurious6]

China may get there, but their inability to bring in talent from around the world is a big problem. That was the main strength of the US until this year.

Yeah that's a big point

>99% of population in China are Chinese
>99% of PHD engineers in China are Chinese

>60% of PHD engineers in USA are Foreigners (or ethnically)

It's extremely impressive for a single country with Tiny immigration to achieve so much so fast

 

[DrMrLordX]

The world will be split in half in that scenario. US and Taiwan will be isolated just as China is.

Given how well TSMC is doing thus far, that doesn't seem to be a problem.

It's 2024-09 Months away from HVM and it's reuturs I will not believe Reuturs they have been dishing out fake news.

Why are you disputing this old article? By now we know that 18A isn't getting any major external customers. Certainly not Broadcom. The article was essentially correct.

 

[511]

Why are you disputing this old article? By now we know that 18A isn't getting any major external customers. Certainly not Broadcom. The article was essentially correct.

18A ain't getting major customer but they were saying 18A is bad which ain't just cause someone doesn't get customer doesn't mean the tech is bad look at Optane and the death it died.