Question Intel Q2: 7 nm in bad shape

jpiniero

Lifer
Oct 1, 2010
14,487
5,154
136

Here's the money shot:

The company's 7nm-based CPU product timing is shifting approximately six months relative to prior expectations.
The primary driver is the yield of Intel's 7nm process, which based on recent data, is now trending approximately
twelve months behind the company's internal target.

Data Center volume up 29% (plus 5% ASP)
Notebook up 9%
Desktop down 14% - not as bad as I had thought

Q3 guidance bad - revenue of 18.2b vs 19.2b in 2019
 
Last edited:

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,439
14,409
136

Still reading, but here's the money shot:
Coming from Intel itself, I certainly believe this. What I am sick of, is the Intel advocates that keep telling us "oh, its coming sooner than expected" and "its got a fine yield, doing great".

Positive thinking is one thing, but right out lying or greatly exaggerating is another.

While I don't want Intel to fail as a company (as I am sure they will not) I am ready for competition to get some market share for a long term battle that will continue. Its healthy and good for the consumer.
 

LightningZ71

Golden Member
Mar 10, 2017
1,627
1,898
136
Who would make up the difference for desktop volume? AMD still has nowhere near enough capacity to make a bigger impact, unless they develop competing products for older nodes.
 

JasonLD

Senior member
Aug 22, 2017
485
445
136
Intel is accelerating its transition to 10nm products this year with increasing volumes and strong demand for an expanding line up. This includes a growing portfolio of 10nm-based Intel Core processors with “Tiger Lake” launching soon, and the first 10nm-based server CPU “Ice Lake,” which remains planned for the end of this year. In the second half of 2021, Intel expects to deliver a new line of client CPU’s (code-named “Alder Lake”), which will include its first 10nm-based desktop CPU, and a new 10nm-based server CPU (code-named “Sapphire Rapids”). The company's 7nm-based CPU product timing is shifting approximately six months relative to prior expectations. The primary driver is the yield of Intel's 7nm process, which based on recent data, is now trending approximately twelve months behind the company's internal target.

I have no idea what Intel's internal target was, but it seems like Alder Lake is 2021 product. It may mean 7nm Client is late 2022 bound.
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,439
14,409
136
Who would make up the difference for desktop volume? AMD still has nowhere near enough capacity to make a bigger impact, unless they develop competing products for older nodes.
How do we KNOW that AMD can't get out considerably greater volume ? Not saying they could make up all 80-90% of all CPU's, but even if they could provide 40% and then later 50%, That could change the market share.
 

LightningZ71

Golden Member
Mar 10, 2017
1,627
1,898
136
I believe that I’ve read in multiple locations that TSMC is still very capacity constrained on their leading edge node products. For AMD to get more capacity, it will take a bidding war with Apple and Nvidia for it among others.
 

turtile

Senior member
Aug 19, 2014
614
294
136
I believe that I’ve read in multiple locations that TSMC is still very capacity constrained on their leading edge node products. For AMD to get more capacity, it will take a bidding war with Apple and Nvidia for it among others.

AMD doesn't use the leading edge at first because of this. GPUs have been first and then CPUs. TSMC will be shipping 5nm this year while AMD will have an improved version in 2022.
 
  • Like
Reactions: Tlh97

NostaSeronx

Diamond Member
Sep 18, 2011
3,683
1,218
136
Couple of questions to answer though;
Is it with the qualification fleet(NXE:3300/NXE:3350/NXE:3400B @ Oregon) or a production fleet(NXE:3400C)?
Is the yield numbers with pellicle or without pellicle?

If it is qualification and without pellicle those expectations are a little early for doom and gloom.

TSMC 5nm/Samsung 3nm don't even have that as doom and gloom. But, we've seen everyone with sense avoid 7nm+/7LPP onwards though for 8LPP/N7P.
 

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
Couple of questions to answer though;
Is it with the qualification fleet(NXE:3300/NXE:3350/NXE:3400B @ Oregon) or a production fleet(NXE:3400C)?
Is the yield numbers with pellicle or without pellicle?

If it is qualification and without pellicle those expectations are a little early for doom and gloom.

TSMC 5nm/Samsung 3nm don't even have that as doom and gloom. But, we've seen everyone with sense avoid 7nm+/7LPP onwards though for 8LPP/N7P.

Oh be quiet for once. It didn't even take a week for you to be proven wrong beyond a shadow of a doubt. This, of course, isn't news to anyone experienced on this forum, but at least have the decency to spare us the rambling for a few weeks.
 

Hitman928

Diamond Member
Apr 15, 2012
5,160
7,595
136
Couple of questions to answer though;
Is it with the qualification fleet(NXE:3300/NXE:3350/NXE:3400B @ Oregon) or a production fleet(NXE:3400C)?
Is the yield numbers with pellicle or without pellicle?

If it is qualification and without pellicle those expectations are a little early for doom and gloom.

TSMC 5nm/Samsung 3nm don't even have that as doom and gloom. But, we've seen everyone with sense avoid 7nm+/7LPP onwards though for 8LPP/N7P.

If Intel still hasn't figured out how to make an EUV pellicle, they are in even worse shape than I thought.
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,683
1,218
136
It didn't even take a week for you to be proven wrong beyond a shadow of a doubt.
Like they aren't producing 7nm at all.
With internal pellicle => constant bad yields
Without internal pellicle => only for client heavy swings of good and bad yields.

Server products need a working pellicle, client products can do without but it gets into total yields/costs. (Client EOL extended support requires pellicle at Intel)
2nd generation internal pellicle or 3rd+ generation foreign pellicle. The foreign pellicle is going to be used for TSMC's 5nm HVM instead of their internal one. The only good 1st generation pellicle is Samsung's one.

Mass production is with NXE:3400C tools w/ 2nd gen pellicles at Intel unless the foreign pellicles are really good as believed. Of which the external pellicles only go in HVM service in Q1-Q2 2021. (Edit: https://jp.mitsuichemicals.com/en/release/2019/2019_0531_01.htm )
 
Last edited:

Hitman928

Diamond Member
Apr 15, 2012
5,160
7,595
136
"Intel now plans to make its first 7-nanometer chip shipments in late 2022 or early 2023, CEO Bob Swan said on a conference call with analysts on Thursday."
Source: CNBC

Man, that is going to be tough sledding for Intel. AMD and ARM will most likely be on TSMC's 3 nm by the time Intel gets 7 nm out the door in volume.
 

JasonLD

Senior member
Aug 22, 2017
485
445
136
"Intel now plans to make its first 7-nanometer chip shipments in late 2022 or early 2023, CEO Bob Swan said on a conference call with analysts on Thursday."
Source: CNBC

Ah, I see, thanks. Since it looks like Intel is going to use external fab for Ponte Vecchio, not sure if that is going to be CPU or GPU. If they can't get 7nm on sufficient volume by then, Intel should look for TSMC/Samsung.
 

Hitman928

Diamond Member
Apr 15, 2012
5,160
7,595
136
Intel said on Thursday that its gross margin fell to 53.3%, the lowest in 11 years.

Looks like Intel ran out of accounting tricks to play to keep their GM up. If this trend holds up, AMD's GM will be within 10 pp of Intel's at the end of the year. That's not great when Intel has to pay for all their fab R&D and maintenance themselves.
 

Hitman928

Diamond Member
Apr 15, 2012
5,160
7,595
136
Ah, I see, thanks. Since it looks like Intel is going to use external fab for Ponte Vecchio, not sure if that is going to be CPU or GPU. If they can't get 7nm on sufficient volume by then, Intel should look for TSMC/Samsung.

The last I heard (may be dated at this point) was that Intel was reluctantly willing to fab GPUs and other things at 3rd party fabs, but that there was no way Intel would allow their CPUs, especially high performance lines, to be fabbed anywhere but Intel. It does make some financial sense as you have huge investments and infrastructure all built around your fabs that gets wasted if you start producing even your CPUs elsewhere, but there's probably a more ideal strategy of mixed fabrication plans. From what I heard, it is largely ego pushing this Intel fab or bust strategy for CPU fabrication.
 

Doug S

Platinum Member
Feb 8, 2020
2,191
3,377
136
I believe that I’ve read in multiple locations that TSMC is still very capacity constrained on their leading edge node products. For AMD to get more capacity, it will take a bidding war with Apple and Nvidia for it among others.

The way Apple is getting first dibs on TSMC's new processes is reportedly the same way they got access to huge volumes of flash back in the day when they went all flash on the iPod, and the iPhone started exploding. Prepayments. If you send a few billion TSMC's way to help them fund the rollout of 5nm, of course you going to get a guaranteed allocation of those early wafers.
 

Doug S

Platinum Member
Feb 8, 2020
2,191
3,377
136
Man, that is going to be tough sledding for Intel. AMD and ARM will most likely be on TSMC's 3 nm by the time Intel gets 7 nm out the door in volume.

Yes, and anyone still wondering why Apple is going their own way on the Mac instead of sticking with Intel should consider that based on what happened with them at 10nm this may be only the first of several 7nm delays.
 

JasonLD

Senior member
Aug 22, 2017
485
445
136
The last I heard (may be dated at this point) was that Intel was reluctantly willing to fab GPUs and other things at 3rd party fabs, but that there was no way Intel would allow their CPUs, especially high performance lines, to be fabbed anywhere but Intel. It does make some financial sense as you have huge investments and infrastructure all built around your fabs that gets wasted if you start producing even your CPUs elsewhere, but there's probably a more ideal strategy of mixed fabrication plans. From what I heard, it is largely ego pushing this Intel fab or bust strategy for CPU fabrication.

If they can actually get CPUs based on 7nm by late 2022/early 2023, it is probably going to be fine, but if they can't, this is probably the time they should be preparing plan B (using TSMC 5nm or Samsung 5/3nm). Not having competitive product out there is more devastating than worrying about wasting previous investment.
 

uzzi38

Platinum Member
Oct 16, 2019
2,556
5,531
146
Ah, I see, thanks. Since it looks like Intel is going to use external fab for Ponte Vecchio, not sure if that is going to be CPU or GPU. If they can't get 7nm on sufficient volume by then, Intel should look for TSMC/Samsung.
Some clarificatio: Intel are producing tiles both internally and externally, and will be packaging both together in-house.