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

 

[Doug S]

Prices for TSMC’s latest manufacturing process, N3P of the 3nm family, are about 20% higher than the prior generation, and 2nm will be over-50% higher than N3P as price inflation continues to hit semiconductors, media report, citing unnamed supply chain sources.

TSMC is not allowing any 2nm (N2) discounts due to its huge production line equipment spend.

I also saw stuff claiming that TSMC was able to keep the number of N2 EUV layers similar to N3, so the price increase wouldn't be nearly as bad as the 50% that had been claimed.

The truth is that anyone who knows at this stage will be under NDA, so most of the numbers out there are guesswork or disinformation. I remember seeing claims a couple years ago that N3 was $25K per wafer. The prices that have been reported more recently are under $20K.

 

[adroc_thurston]

The prices that have been reported more recently are under $20K.

19k-ish iirc.
N2 is like 20.5/21k a wafer.
But TSM is hiking prices though.

 

[jpiniero]

19k-ish iirc.
N2 is like 20.5/21k a wafer.
But TSM is hiking prices though.

I'll take the over on both, Bob.

 

[511]

19k-ish iirc.
N2 is like 20.5/21k a wafer.
But TSM is hiking prices though.

Last I really heard was 22K ish for N2 but with 5-10% increase 24K seems possible.

 

[adroc_thurston]

Last I really heard was 22K ish for N2

I've heard 21k for the top 5 customers but yeah it's very WIP and TSMC wants moar moneyz.

 

[511]

I've heard 21k for the top 5 customers but yeah it's very WIP and TSMC wants moar moneyz.

They will simply say and customer will give like fabless has any options

 

[Win2012R2]

I've heard 21k for the top 5 customers but yeah it's very WIP and TSMC wants moar moneyz.

And I've heard it's 21.5k actually. I guess they use different numbers just to trace leakers...

 

[511]

And I've heard it's 21.5k actually. I guess they use different numbers just to trace leakers...

It's per customer and depending on your relationship and volume you get different number what is well known that Apple gets the best price.

 

[Win2012R2]

And has some of the smaller chips for the best yield. Imagine a 700mm2 GPU on N2X.

Me wants!

 

[511]

And has some of the smaller chips for the best yield. Imagine a 700mm2 GPU on N2X.

Apple pro max are relatively large dies in 500-600mm2 sq range

 

[Josh128]

Apple pro max are relatively large dies in 500-600mm2 sq range

lol what?? M3 die are estimated to be around the 146mm2 range, hard data is not available for M4, but its estimated to be around 200-250mm2. Both based on N3. Where on earth did you come up with 500-600mm2? Source please.

Regardless, iphones dont use pro max die, they use A-series SOCs. A18 Pro has a 105mm² die size, while the A18 measures 90mm². Those SOCs ship AT LEAST 1000x (possibly MUCH more) the volume than M series SOCs do.

 

[Josh128]

So this is essentially confirmation that nothing N2P based is coming before July of next year, and nothing N2X based is coming before 2027. The 256C Zen 6 Venice that Lisa showed as the first N2 product a couple months back is almost certainly N2 only due to efficiency requirements and the below timeline.

Interesting that Fmax increase is only mentioned for N2X and not at all for N2P. Wouldnt it be something if all the 6.5GHz hoopla caused by MLID is for a product coming in 2027 and Zen 6 doesnt move the bar much at all and ends up at ~5.8GHz.

 

[511]

lol what?? M3 die are estimated to be around the 146mm2 range, hard data is not available for M4, but its estimated to be around 200-250mm2. Both based on N3. Where on earth did you come up with 500-600mm2? Source please.

Here is the image for M3 and here is the die size for M3 https://www.bilibili.com/opus/1003234408233172993 just use this and the image and you can calculate it from it

 

[511]

Here is a better method just take the pixel for them divide by area for M3 = 146mm2 for 359*301 and for M3 Max it's 609*626 using snipping tool and maths

 

[Josh128]

Here is a better method just take the pixel for them divide by area for M3 = 146mm2 for 359*301 and for M3 Max it's 609*626 using snipping tool and maths
View attachment 130814

Regardless, my point is that the vast majority of the silicon Apple sells is sub 120mm2. Likely >1000:1 vs M series, and also more than any other silicon from any other manufacturer.

 

[511]

Regardless, Apple sells >1000X A series SOCs than they do M MAX SOCs, that was my point.

doesn't change the fact that M3 masx is ~500mm2 also M4 is 168mm2

 

[Joe NYC]

Here is a better method just take the pixel for them divide by area for M3 = 146mm2 for 359*301 and for M3 Max it's 609*626 using snipping tool and maths
View attachment 130814

Hmm, I did not know you could do that.

I use Excel for nearly everything or sometimes I just put something like that into Google Search

 

[511]

Hmm, I did not know you could do that.

I use Excel for nearly everything or sometimes I just put something like that into Google Search

If the ratio is same for the image than you can totally do that

 

[Josh128]

If the ratio is same for the image than you can totally do that

Problem is these images are clearly not at the same scale. Not drastically different, but different nonetheless.

 

[511]

Problem is these images are clearly not at the same scale. Not drastically different, but different nonetheless.

View attachment 130816

These are derived from this

Apple unveils M3, M3 Pro, and M3 Max, the most advanced chips for a personal computer

M3, M3 Pro, and M3 Max feature groundbreaking technologies that deliver dramatically increased performance and unleash new capabilities for Mac.

www.apple.com

 

[Doug S]

lol what?? M3 die are estimated to be around the 146mm2 range, hard data is not available for M4, but its estimated to be around 200-250mm2. Both based on N3. Where on earth did you come up with 500-600mm2? Source please.

Regardless, iphones dont use pro max die, they use A-series SOCs. A18 Pro has a 105mm² die size, while the A18 measures 90mm². Those SOCs ship AT LEAST 1000x (possibly MUCH more) the volume than M series SOCs do.

They are nowhere near "1000x" more. Apple ships about 10-12x more iPhones than Macs per year, so you're only off by two orders of magnitude!

 

[Josh128]

They are nowhere near "1000x" more. Apple ships about 10-12x more iPhones than Macs per year, so you're only off by two orders of magnitude!

When I say 1000x, Im talking their small SOCs vs their large ones. My estimate includes ipads and Apple watches, most of which include A series and smaller SOCs, and Im talking about vs the M series Pro ( I specifically called out the Pro) and Max SOCs that are larger than the standard M Series 150mm2, which is still a small die. Theres no way you can make me believe that for every 10 people on the planet that has ever purchased an iphone, ipad, or Apple Watch, 1 has purchased one of these $2000-$4000 M3 Pro or Ultra devices, but if you can back that up with sales numbers, I'll stand corrected.

 

[DavidC1]

When I say 1000x, Im talking their small SOCs vs their large ones. My estimate includes ipads and Apple watches, most of which include A series and smaller SOCs, and Im talking about vs the M series Pro ( I specifically called out the Pro) and Max SOCs that are larger than the standard M Series 150mm2, which is still a small die. Theres no way you can make me believe that for every 10 people on the planet that has ever purchased an iphone, ipad, or Apple Watch, 1 has purchased one of these $2000-$4000 M3 Pro or Ultra devices, but if you can back that up with sales numbers, I'll stand corrected.

Apple watches are fraction of iPhones.

iPhone is about 220 million per year. Apple has sold 22-25 million mac units. Now the iPhone data is 2022 and mac units are 2025 data. But it is roughly 1 out of 10. Even if you only include the M3 Max, it's not going to be anywhere near 100x, nevermind 1000x and that's a very contrived way of calculating things.

 

[Win2012R2]

iPhone is about 220 million per year. Apple has sold 22-25 million mac units.

iPhone profitability is much higher - plus big fat cut from all the apps sold.

 

[Doug S]

iPhone profitability is much higher - plus big fat cut from all the apps sold.

Overall? Obviously. Per unit? I bet the Mac wins there, at least at up front at purchase time.

The iPhone display is smaller but much more complex so it'll cost more per unit, there are up to four cameras (five if you include Face ID's IR camera) and it includes cellular which even when using Apple's modems still incurs patent licensing expense. The Mac has bigger chips and more RAM/NAND but also higher ASP. The Mac's gross margin is almost certainly higher than iPhone's, though AFAIK Apple hasn't broken out the gross margins of individual product lines for a long time so I have no way to prove that.

But yeah in terms of total profit its not even a contest thanks to selling 10x more iPhones, and the endless billions in post sale services income for as long as the phone remains operational even with second or third owners. Mac would be a megaprofitable business for almost any other company, but for Apple it is practically an afterthought.