TSMC Planning a 4nm Node that goes Live in 2023

[btarunr]

TSMC is reportedly planning a stopgap between its 5 nm-class silicon fabrication nodes, and the 3 nm-class, called N4. According to the foundry's CEO, Liu Deyin, speaking at a shareholders meeting, N4 will be a 4 nm node, and an enhancement of N5P, the company's most advanced 5 nm-class node. N4 is slated for mass-production of contracted products in 2023, and could help TSMC's customers execute their product roadmaps of the time. From the looks of it, N4 is a repeat of the N6 story: a nodelet that's an enhancement of N7+, the company's most advanced 7 nm-class node that leverages EUV lithography.



View at TechPowerUp Main Site

 

[Vayra86]

Its all 7nm let's stop fooling each other.

Its not like we're magically getting multiple shrinks when the current ones are still marketable. Come on

 

[john_]

It looks to me like an extra precaution in case there are delays with the 3nm process.

 

[InhaleOblivion]

Why waste a perfectly good profitable number?

 

[steve360]

Meanwhile over at Intel, they are hard at work trying to squeeze another plus to their 14nm++++++++++++++++++++++++++++++

 

[Ravenas]

Its all 7nm let's stop fooling each other.

Its not like we're magically getting multiple shrinks when the current ones are still marketable. Come on

Marketable to their Primary customer AAPL? AAPL is driving TSMC growth, everyone else is just gravy on top (those marketable customers you are referring too). iPhone, or other iOS/MAC device, will likely have that die size in some product in 2023/24

 

[Vayra86]

Marketable to their Primary customer APPL? APPL is driving TSMC growth, everyone else is just gravy on top (those marketable customers you are referring too). iPhone, or other iOS/MAC device, will likely have that die size in some product in 2023/24

What now?

AMD set to become TSMC's largest 2020 client

Citing supply chain sources, Taiwanese outlet Apple Daily predicts that AMD is set to become TSMC's largest 7nm customer by the end of 2020, displacing Apple in the process. The spectacular success of Ryzen and Apple's move to 5nm are likely reasons.

www.notebookcheck.net

Apple is just grabbing what they can and they also want to tell a cool story about shrinks make no mistake. They thrive on the impression of being 'ahead of the curve'. Never underestimate the power of perception...

Why waste a perfectly good profitable number?

Inb4 the decimal nanometers Oh no, shit, didn't Intel already show us a roadmap slide for '1.4nm' ? To be achieved by ominous sounding 'new developments'... maybe they'll cut up a 14nm chip in 10 pieces.

 

[Ravenas]

What now?

AMD set to become TSMC's largest 2020 client

Citing supply chain sources, Taiwanese outlet Apple Daily predicts that AMD is set to become TSMC's largest 7nm customer by the end of 2020, displacing Apple in the process. The spectacular success of Ryzen and Apple's move to 5nm are likely reasons.

www.notebookcheck.net

Are you not understanding what I said? 7nm is marketable to gravy customers, and not their Primary customer AAPL. Your article which you sourced states that AMD is set to become something they are not (somewhat dumb reporting), and the only reason for this is because AAPL is transitioning to 5nm, once again driving TSMC growth.

 

[Vayra86]

Are you not understanding what I said? 7nm is marketable to gravy customers, and not their Primary customer APPL. Your article which you sourced states that AMD is set to become something they are not (somewhat dumb reporting), and the only reason for this is because APPL is transitioning to 5nm, once again driving TSMC growth.

Right, let's just have a long look at those 5nm actual fin pitch measurements first though. Let's just say the definition is highly volatile. I'm not contesting the Apple story, I'm questioning the actual nanometers we see on paper.

 

[Ravenas]

Right, let's just have a long look at those 5nm actual fin pitch measurements first though.

Not sure what that has to do with marketshare of TSMC customers.

AAPL perception has nothing to do with the reality of the situation. AAPL is the most valuable company in the US, and world over. That is not because they have pulled the wool over investor's eyes.

I will agree with you in regards to contesting die sizes, as each company has their own methodology for measuring distance (there is no world standard that I am aware of). However, AAPL along with all of the other large TSMC customer shares makes an insurmountable growth and investment strategy in comparison to Intel. Intel will go semi fabless soon in order to compete.

 

[Vayra86]

Not sure what that has to do with marketshare of TSMC customers.

AAPL perception has nothing to do with the reality of the situation. AAPL is the most valuable company in the US, and world over. That is not because they have pulled the wool over investor's eyes.

I will agree with you in regards to contesting die sizes, as each company has their own methodology for measuring distance (there is no world standard that I am aware of). However, AAPL along with all of the other large TSMC customer shares makes an insurmountable growth and investment strategy in comparison to Intel. Intel will go semi fabless soon in order to compete.

I think we have proven, the world over and irrespective of ideology that humans are easy to fool into the next economical bubble. We're also keen on cognitive dissonance. What really happens is this: somewhere in some board room people think up a roadmap and it contains smaller nodes. Roadmaps become promises. Promises can't always be kept. That is how we got to that disparity between node measurements in the first place, as we did with all other places in the economy where we demand constant growth and improvement.

Something's gotta give, and usually a big part of that is hidden in marketing. For nanometers we already know that is the case. Why would 5nm be different? Even 7nm isn't true 7nm, and it certainly wasn't when they used DUV.

In the context of Apple and suppliers towards customers they're all playing the same ball game and want to come out winning. Why do you think wealth accumulates at some few % of the world? Its a big scratch your back and you scratch mine circus. The deals between TSMC and Apple are exactly that. Its not some weird conspiracy... its how the world works.

We regard node shrinks as progress towards some undisclosed higher goal. We view it as a pinnacle of our abilities and technology. It has immense marketing value and the vast majority has no idea what they're listening to. All they see is profit numbers and a shrink representing a bigger number.

 

[Ravenas]

In the context of Apple and suppliers towards customers they're all playing the same ball game and want to come out winning. Why do you think wealth accumulates at some few % of the world? Its a big scratch your back and you scratch mine circus. The deals between TSMC and Apple are exactly that. Its not some weird conspiracy... its how the world works.

I'm not sure where this is going or why it was brought up; it's somewhat wild to bring this up in my opinion...

A supply contract in a supply chain is just that. Whatever clause that is in place on each unit produced due to overall minimum volume is speculative, but there isn't a conspiracy involved and this really shouldn't need any explaining. The world is really getting out of hand with the conspiracy generation, and "in place" explanation when lack of knowledge or understanding is present to the person or entity (shall we say) deriving the conspiracy.

 

[cucker tarlson]

Why waste a perfectly good profitable number?

hell,nvidia will be using 8nm that's really 10nm and probably performs same as tsmc's excellent 12nm which is their 16nm only better

 

[ppn]

TSMC 7nm performance is 10nm.

if 7nm really was 7nm as they claim Big Navi 505mm2 would be as big as small Navi10 - 251mm2.

But instead provides only 60% better density than 14nm. NAVI10 on 14nm would be 404mm2.

 

[ARF]

Its all 7nm let's stop fooling each other.

Its not like we're magically getting multiple shrinks when the current ones are still marketable. Come on

Why waste a perfectly good profitable number?

Meanwhile over at Intel, they are hard at work trying to squeeze another plus to their 14nm++++++++++++++++++++++++++++++

What is worse - multiple pluses after 14 or going 7 -> 6 -> 5 -> 4 with the same result or like a quarter node shrink or even less ?


If you really want to make a proper and serious discussion - start discussing the ITRS Logic Device Ground Rules for each process node.

Here: page is 12 from 36 https://irds.ieee.org/images/files/pdf/2017/2017IRDS_MM.pdf

 

[RoutedScripter]

Guys .... it's the same story every time.

14 nm ... 5 nm ... oh there will be 7 nm stopgap, .... 5 nm .... 3 nm .... oh there will be a 4 nm stopgap, I think they have all this planed ahead and it's just so confusingly messaged to create drama, for the industry to play with prices between themselfs and the media enjoys it due to the fact there's more articles to write, but yeah, it's getting harder, but I also think they secretly have much more advanced processes they're not telling anyone, they want to suck this tech dry before they get anywhere near something completely new, that's the definition of the industry practice of cow milking, an incredibly predatory and disrespectful practice.