Intel Has Fixed its 7 nm Node, But Outsourcing is Still Going to Happen

[AleksandarK]

Intel has today reported its Q4 2020 earnings disclosing full-year revenue with the current CEO Bob Swan, upcoming new CEO Pat Gelsinger, and Omar Ishrak, Chairman of Intel's board. During the call, company officials have talked about Intel's earnings and most importantly, addressing the current problems about the company's manufacturing part - semiconductor foundries. Incoming Intel CEO, Pat Gelsinger, has talked about the state of the 7 nm node, giving shareholders reassurance and a will to remain in such a position. He has made an argument that he has personally reviewed the progress of the "health and recovery of the 7 nm program."

The 7 nm node has been originally delayed by a full year amid the expectations, and as with the 10 nm node, we have believed that it is going to experience similar issues. However, the incoming CEO has reassured everyone that it is very much improving. The new 7 nm node is on track for 2023 delivery, when Intel is expected to compete with the 3 nm node of TSMC. Firstly, Intel will make a debut of the 7 nm node with client processors scheduled for 1H 2023 arrival, with data center models following that. The company leads have confirmed that Intel will stay true to its internal manufacturing, but have stressed that there will still be a need for some outsourcing to happen.


View at TechPowerUp Main Site

 

[TheLostSwede]

Uhm, hasn't the 7nm node been delayed a lot more than just a year by now?

 

[chris.london]

Intel fixed its 7nm node yet the first 7nm processors won't be launched for at least 2 more years. This is the weirdest fix ever. (NB back in 2019 Intel's 7nm processors were also only 2 years away.)

 

[freeagent]

Progress!

 

[xtreemchaos]

havnt we been here before ?. smoke n mirrors me thinks . fingers crossed anyways.

 

[john_]

I don't see a "have fixed" somewhere in the article or in news. What we understand from Intel's comments, is that there are no delays, at least nothing bad happened those last 3 months.

 

[mtcn77]

You have to consider 7nm not only delivers a new node, but also the next processes Intel foundries will develop on. When you lose one, you lose the other. You cannot adopt another foundry without banishing your own.
This is good news. Production will continue. Intel has to give the right signals in order to continue as an independent manufacturer.

 

[bug]

I feel the lack of fabs goes far beyond Intel. Wherever you look, it seems there's only shortages and inflated prices
Of course, Intel just happens to have hit a low within an industry-wide low, which doesn't really help.

 

[TheLostSwede]

I feel the lack of fabs goes far beyond Intel. Wherever you look, it seems there's only shortages and inflated prices
Of course, Intel just happens to have hit a low within an industry-wide low, which doesn't really help.

There are plenty fabs, the issue is that most of those fabs are not producing at small enough nodes, nor do they use large enough wafers to produce the kinds of chips that there is a shortage of.

List of semiconductor fabrication plants - Wikipedia

en.wikipedia.org

Number of open fabs currently listed here: 529

 

[bug]

There are plenty fabs, the issue is that most of those fabs are not producing at small enough nodes, nor do they use large enough wafers to produce the kinds of chips that there is a shortage of.

List of semiconductor fabrication plants - Wikipedia

en.wikipedia.org

Yes, that's what I meant. I know "old" fabs stick around for a while because not everything needs to be built on cutting-edge nodes.
Tbh, I'm not even sure if there are too few cutting edge fabs available or demand has just soared. Either way, it still sucks for end-users.

Edit; Going over that list, I didn't realize there are so many fabs states-side. Also, where does Taiwan put all those fabs? Do they expand Dutch-style?

 

[Verpal]

There are plenty fabs, the issue is that most of those fabs are not producing at small enough nodes, nor do they use large enough wafers to produce the kinds of chips that there is a shortage of.

List of semiconductor fabrication plants - Wikipedia

en.wikipedia.org

If only there is a way to easily standardize and make design more portable between fabrication node, I am sure plenty of people and AIB willing to take on Ampere/Navi fabricated on Intel 14nm, Glofo 12nm or, god forbid, some ancient 28nm node only suitable for a 3050(ti), considering the persistent shortage and prices.

 

[TheLostSwede]

If only there is a way to easily standardize and make design more portable between fabrication node, I am sure plenty of people and AIB willing to take on Ampere/Navi fabricated on Intel 14nm, Glofo 12nm or, god forbid, some ancient 28nm node only suitable for a 3050(ti), considering the persistent shortage and prices.

Sorry, but that's impossible. You can't jump node sizes like that without re-doing the entire chip layout.

 

[mtcn77]

Sorry, but that's impossible. You can't jump node sizes like that without re-doing the entire chip layout.

That's what I tried deliberating, without a present node, there is no development for a future node. It is done, the costs are too high to start over.

 

[LemmingOverlord]

I think people should realise that there is a serious mis-reporting issue going on at Intel, where someone down the line is feeding top management some serious bullshit and then things derail. Gelsinger might've read a report about the 7nm health and recovery, but to what extent is this not a totally bogus report?

 

[Slizzo]

I think people should realise that there is a serious mis-reporting issue going on at Intel, where someone down the line is feeding top management some serious bullshit and then things derail. Gelsinger might've read a report about the 7nm health and recovery, but to what extent is this not a totally bogus report?

Well, Murthy was kicked to the curb, so they're probably getting reliable information again, fortunately.

 

[Flanker]

I totally believe you, Intel

 

[ThrashZone]

Hi,
Code name lavalake.

 

[TheLostSwede]

Hi,
Code name lavalake.

You sure it's not Sunlake?

 

[Arc1t3ct]

Intel Has Fixed its 7 nm Node​

Nope...

 

[Steevo]

Setting up the new blood for delayed failure, but at least giving them time to fail.

Also very misleading title, Intel has NOT fixed their 7nm nde as if there were a power switch somewhere that someone forgot to turn on, its more of, they are promising to have their node operational if not profitable by 2023

 

[mtcn77]

they are promising to have their node operational if not profitable by 2023.

They are essentially the same thing, a process develops as it is scaled.

 

[Luminescent]

Isn't it strange that almost everybody can make 7nm for some time but Intel, well, everybody meaning Samsung and TSMC.
Maybe Intel wanted to buy back some shares and they couldn't do that at a reasonable price so they keept stalling with the " we can't make 7/10nm " and eventually or already bought some shares back.

 

[bug]

Isn't it strange that almost everybody can make 7nm for some time but Intel, well, everybody meaning Samsung and TSMC.
Maybe Intel wanted to buy back some shares and they couldn't do that at a reasonable price so they keept stalling with the " we can't make 7/10nm " and eventually or already bought some shares back.

It would be strange if 7nm meant the same for everyone. As it is, it's not strange at all.

 

[Luminescent]

That's why i said "we can't make 7/10nm", i know about the density and all that.
But it's strange they can't make anything close to what Samsung and TSMC makes, even if they lowered the density.
This could be just bull...it, biggest semiconductor company in the world stuck at the same node for 6 years, no progress in 6 years!

 

[ZoneDymo]

"Everything is fine"