• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Discussion Leading Edge Foundry Node advances (TSMC, Samsung Foundry, Intel) - [2020 - 2025]

Page 261 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

DisEnchantment

Golden Member
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.

---------------------------------------------------------------------------------------------------------------------------------------------------


FEEL FREE TO CREATE A NEW THREAD FOR 2025+ OUTLOOK, I WILL LINK IT HERE
 
Last edited:
ifs
 

Attachments

  • Screenshot_20260325_113009_Adobe Acrobat.jpg
    Screenshot_20260325_113009_Adobe Acrobat.jpg
    236.3 KB · Views: 44
  • Screenshot_20260325_112949_Adobe Acrobat.jpg
    Screenshot_20260325_112949_Adobe Acrobat.jpg
    276.5 KB · Views: 44
  • Screenshot_20260325_112928_Adobe Acrobat.jpg
    Screenshot_20260325_112928_Adobe Acrobat.jpg
    236.9 KB · Views: 42
  • Screenshot_20260325_112905_Adobe Acrobat.jpg
    Screenshot_20260325_112905_Adobe Acrobat.jpg
    335.3 KB · Views: 40
  • Screenshot_20260325_112843_Adobe Acrobat.jpg
    Screenshot_20260325_112843_Adobe Acrobat.jpg
    265.3 KB · Views: 38
  • Screenshot_20260325_112843_Adobe Acrobat.jpg
    Screenshot_20260325_112843_Adobe Acrobat.jpg
    265.3 KB · Views: 42
Intel investor lmao it might be true though

It is no surprise. Demand for both logic and memory is through the roof.


"We are seeing that TSMC is hitting (production capacity) limits," Natarajan Ramachandran, director of product marketing in Broadcom’s Physical Layer Products division, told reporters on Tuesday, adding he would have described TSMC's capacity as "infinite" until a few years ago.
 
  • Like
Reactions: 511
@RTX would be interesting to see more specific information on GF's claims. What exactly did Tower steal from them, if anything?

The court filings would list the 11 patents. I'm sure there will be articles written that link them.

Edit: here is the relevant text from the lawsuit:

1. This is a civil action for infringement of United States Patent Nos. 11,476,244 (“the
’244 Patent”), 8,283,193 (“the ’193 Patent”), 11,658,177 (“the ’177 Patent”), 7,566,653 (“the ’653
Patent”), and 9,269,666 (“the ’666 Patent”) (collectively, “the Asserted Patents”) under the patent
laws of the United States, 35 U.S.C. § 1 et seq.
2. To address Tower’s widespread infringement of GlobalFoundries’ patents,
GlobalFoundries U.S. has also brought suit against Tower in both the Western District of Texas,
and the United States International Trade Commission, asserting U.S. Patent Nos. 10,062,748;
8,507,983; 9,093,425; 9,865,546; 10,707,167; and 8,330,235. Those patents are distinct from, and
unrelated to, the Asserted Patents here.
 
The court filings would list the 11 patents. I'm sure there will be articles written that link them.

Edit: here is the relevant text from the lawsuit:

Okay, that gives us the patents. But that doesn't show us where Tower actually used those patents without paying royalties.
 
Okay, that gives us the patents. But that doesn't show us where Tower actually used those patents without paying royalties.

Did you read the whole filing? I sure didn't, but I'm guessing it at least hints at it there though the meat of the argument would be shown in front of a judge (or shown to opposing counsel in pretrail negotiations hoping to get them to say "oh sh-t!" and settle)
 
Did you read the whole filing? I sure didn't, but I'm guessing it at least hints at it there though the meat of the argument would be shown in front of a judge (or shown to opposing counsel in pretrail negotiations hoping to get them to say "oh sh-t!" and settle)

Not yet. Kinda hoping for a product/process summary or possible injunctions.
 
But they're hungry! I tend to avoid AI's, but, for chewing through dense legal documents when I don't have a lot of time to waste, they are VERY useful to give you the general idea. If you ask the right questions, they'll even tell you where exactly to look in the document for specific language.
 
But they're hungry! I tend to avoid AI's, but, for chewing through dense legal documents when I don't have a lot of time to waste, they are VERY useful to give you the general idea. If you ask the right questions, they'll even tell you where exactly to look in the document for specific language.

Yeah this sort of thing is the perfect use for AI. Give it a large block of text you aren't interested in reading or uses some terms you aren't familiar with (like medical research) and ask it some questions about what it says or what conclusions it makes.

And as you say can ask it to point you to the relevant section so you can check up on it if you suspect it of hallucinating.
 
Intel investor lmao it might be true though

Exactly.

I have seen just about everything that happened (or did not happen) be twisted by Intel investors on Twitter to mean only one thing - that outside customers for Intel foundry are inevitable.

Everything imaginable always implies only one thing - inevitability of new external customers for Intel foundry.
 
No.
A14 is a sizeable chunk faster than A16.
I don't think I can call 4-5% sizable chunk faster according to TSMC themselves.
That's surprising, since A16 is just a derivative of N2. Does A16 than have significant power savings over N2?
8585d767-f371-469a-ad8f-3a6ee8ef1ba2_542x582.png
You can see in the slides clearly A14 is not much improvement over A16. A16 reference is N2P and A14 is base N2 and here is N2P slide
images(33).jpg
e4uUhFVTSisqcMUdhptuBk.png
 
If you set the baseline to N3E and use the midpoints of the ranges stated, it reads as A16 being 3% faster than A14. The power reduction is only slightly better, ~2%.

Seems.. unsatisfactory.
 
If you set the baseline to N3E and use the midpoints of the ranges stated, it reads as A16 being 3% faster than A14. The power reduction is only slightly better, ~2%.

Seems.. unsatisfactory.
It all comes to design next couple of yearS. Who ever has the best designs wins

AMDs zen7 that’s leaked so far looks promising
 
Back
Top