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

Page 119 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

DisEnchantment

Golden Member
Mar 3, 2017
1,659
6,100
136
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.

1587737990547.png
N7 performance is more or less understood.
1587739093721.png

This year and next year TSMC is mainly increasing capacity to meet demands.

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.


1587739615344.png

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.
 

dr1337

Senior member
May 25, 2020
379
635
136
What a strange take. TSMC will make everything that AMD needs, as long as they order it and are willing to pay the price.

TSMC and AMD collaborate very closely, like with the X3D-chips, so it's harder for AMD to switch than Nvidia.
Its not a strange take at all, AMD has gone on record saying they undership. Its obvious with how they position themselves in the market right now most of their production is going towards MI300 and other market segments. I mean, did you already forget the shortages of 2021? And that had a third curveball of console silicon as well.

They have something like 20% comsumer marketshare, maybe 30% on the higher estimates. If they want to be at a healthy 50% then they need to literally or very nearly double their output. It seems rough that they constantly have to flipflop silicon between segments due to availability. I assume they would want to sell more chips than just change customers every year.

I guess they do already balance this by using multiple nodes at one time, but still even as 3nm comes online how much does that really expand their ability to sell new chips? Even Nvidia seems capacity limited with regard to their AI demand.
 

gdansk

Platinum Member
Feb 8, 2011
2,454
3,316
136
Intel and TSMC have talked about their 14å nodes so it is time for Samsung. The article itself calls it SF1.4. Don't get too sarcastic at a bad headline.
 

FlameTail

Platinum Member
Dec 15, 2021
2,888
1,627
106

Big news

AMD is in talks with Samsung Electronics to place orders for 3nm GAA (gate all around, new transistor type) chip production, media report, citing an unnamed industry source. AMD CEO Lisa Su reportedly attended the ITF Summit 2024 in Belgium, saying AMD will use the 3nm GAA process “to improve power efficiency and performance…” Samsung will also provide HBM memory chips and packaging services to AMD
Previously it was 4nm. Now they are taking about 3nm GAAFET.
 

Doug S

Platinum Member
Feb 8, 2020
2,420
3,912
136

adamge

Member
Aug 15, 2022
56
135
66

SiliconFly

Golden Member
Mar 10, 2023
1,118
573
96

Panther Lake power on already.
18A volume ramp is 25H1 confirmed.

And pat kept saying during the keynote about process leadership next year without a doubt!

Looks like Intel's execution engine is finally back online...
 
Last edited:
  • Like
Reactions: adamge

Hitman928

Diamond Member
Apr 15, 2012
5,523
8,566
136

Panther Lake power on already.
18A volume ramp is 25H1 confirmed.

And pat kept saying during the keynote about process leadership next year without a doubt!

Looks like Intel's execution engine is finally back online...

Intel announced MTL power on just over 2 years before it was released. It's an important milestone, but there is still a long way to go before you get a final product. Hopefully Intel can follow a quicker schedule this time.
 
  • Like
Reactions: SiliconFly
Jun 4, 2024
116
144
71
Intel announced MTL power on just over 2 years before it was released. It's an important milestone, but there is still a long way to go before you get a final product. Hopefully Intel can follow a quicker schedule this time.
One would hope that with design team not being able to hide margins behind foundry, and the greater emphasis on pre-silicon validation, Panther Lake won’t suffer the same delays. Meteor Lake needed too many steppings.
 

SiliconFly

Golden Member
Mar 10, 2023
1,118
573
96
One would hope that with design team not being able to hide margins behind foundry, and the greater emphasis on pre-silicon validation, Panther Lake won’t suffer the same delays. Meteor Lake needed too many steppings.
Meteor Lake was more of a test vehicle. And it appears they're already ditching it in favor of Lunar Lake.
 

lightisgood

Member
May 27, 2022
182
76
71
One would hope that with design team not being able to hide margins behind foundry, and the greater emphasis on pre-silicon validation, Panther Lake won’t suffer the same delays. Meteor Lake needed too many steppings.

Pat said that Intel 18A product, CWF & PTL, will be released in mid'25 in 1Q24 earnings report.
You saw Xeon 6 got the process leadership back from EPYC, didn't you?
Today, Intel is forcefully fulfilling the 5N4Y promise.
There is nothing to worry about 18A.

Meteor Lake was more of a test vehicle. And it appears they're already ditching it in favor of Lunar Lake.

the number of MTL shipment
1Q24 : 5M
2Q24 : 10M (forecast)

the number of AI processors (MTL&LNL&ARL) shipment
2024 : 40M

MTL shipment will decrease rapidly in 2H24 and LNL&ARL shipment will exceed 20M?
I don't think so.
 

SiliconFly

Golden Member
Mar 10, 2023
1,118
573
96
the number of MTL shipment
1Q24 : 5M
2Q24 : 10M (forecast)

the number of AI processors (MTL&LNL&ARL) shipment
2024 : 40M

MTL shipment will decrease rapidly in 2H24 and LNL&ARL shipment will exceed 20M?
I don't think so.
I don't have the sales numbers, but I do think MTL won't even hit half the numbers compared to previous generations.
 
  • Like
Reactions: lightisgood

Doug S

Platinum Member
Feb 8, 2020
2,420
3,912
136
Intel announced MTL power on just over 2 years before it was released. It's an important milestone, but there is still a long way to go before you get a final product. Hopefully Intel can follow a quicker schedule this time.

The power on indicates that their design is solid enough to boot, but it tells us nothing about 18A. Yield could be 0.01% and they had to dig through dozens of wafers until they found one that tests good. Or maybe every wafer has a bunch of working ones, but their architects will struggle so long getting from "boot" to "ready to sell" that 18A lines are ready to go with nothing to make.

I've mentioned at times over the last few years that I've been watching Intel and considering throwing some money their way but haven't remain convinced. I'm still not fully convinced, but the news of the last few months caused me to pull the trigger on 10K shares last week. I don't feel nearly as confident about it as I did when I belatedly bought Nvidia around this time last year, but I figure the worry about AMD and ARM eating their lunch has become overblown, and by taking PE money to help with their fab buildout they seem serious about building sufficient capacity to eventually compete for whales like Nvidia and Apple.
 

SiliconFly

Golden Member
Mar 10, 2023
1,118
573
96
The power on indicates that their design is solid enough to boot, but it tells us nothing about 18A. Yield could be 0.01% and they had to dig through dozens of wafers until they found one that tests good. Or maybe every wafer has a bunch of working ones, but their architects will struggle so long getting from "boot" to "ready to sell" that 18A lines are ready to go with nothing to make.
18A is 2nd gen GAAFET & 2nd gen BSPD. They've had enough time to work out the issues. 18A shouldn't face any major problems (like 20A's probably limited supply; something I expect).

...but the news of the last few months caused me to pull the trigger on 10K shares last week. ...
Too soon if you ask me.

...I don't feel nearly as confident about it as I did...
Thats because Intel's trying to turn the ship around. But they're not there yet. And the future can be either: Moonshot or Titanic. I don't think we should discuss money matters here, but my advice is, be cautious. Even though I like him a lot, Pat's words shouldn't be taken literally. Who knows what all skeletons Intel still has in its closet.