Intel CFO Talks About 7nm Rollout, Delay in 10nm, Increased Competition from AMD

[btarunr]

Intel CFO George Davis in an interview with Barron's commented on the company's financial health, and some of the reasons behind its rather conservative gross margin guidance looking forward to at least 2023. Intel's current product stack is moving on to the company's 10 nm silicon fabrication process in a phased manner. The company is allocating 10 nm to mobile processors and enterprise processors, while brazening it out with 14 nm on the client-desktop and HEDT platforms until they can build 10 nm desktop parts. AMD has deployed its high-IPC "Zen 2" microarchitecture on TSMC's 7 nm DUV process, with plans to go EUV in the coming months.

"We're still keenly focused on gross margin. Everything from capital efficiency to the way we're designing our products. What we've said though, the delay in 10 nanometer means that we're going to be a little bit disadvantaged on unit cost for a period of time. We actually gave guidance for gross margin out in 2021 to help people understand. 2023 is the period that we were ultimately guiding [when] we're going to see very strong revenue growth and margin expansion. We've got to get through this period where we have the 10 nanometer being a little bit late [as] we're not optimized on a node that we're on. But [by] then we're moving to a two to two and a half year cadence on the next nodes. So we're pulling in the spending on 7 nanometer, which will start up in the second half of 2021 because we think it's the right thing to do competitively," he said.



Davis and Barron's also spoke about Intel's 7 nm rollout plans. Davis stated that Intel's 7 nm silicon fabrication node is expected to begin mass-production of chips only by the second half of 2021. Barron's then turned the discussion to AMD and the traction its EPYC "Rome" server processors are gaining with enterprises. "We said we expect to have heightened competition over the next 18 to 24 months. And our outlook reflects that. Our view on the nature of that competition and impact hasn't really changed since we gave [our] longer term forecast in May," he said. You can read the full Barron's interview from the source link.

View at TechPowerUp Main Site

 

[londiste]

Wrong source link?
This should be the one: https://www.barrons.com/articles/intel-cfo-stock-51572840766

 

[Space Lynx]

Selling my Ryzen 3600 next summer, upgrading to Ryzen 4800x then retiring for a good 3-5 years, possibly longer.

Sorry Intel, I won't miss you.

 

[btarunr]

Wrong source link?
This should be the one: https://www.barrons.com/articles/intel-cfo-stock-51572840766

Fixed, thanks.

 

[GeorgeMan]

Selling my Ryzen 3600 next summer, upgrading to Ryzen 4800x then retiring for a good 3-5 years, possibly longer.

Sorry Intel, I won't miss you.

My exact plan. Maybe even a used 4900x on 2021.
And then may consider intel for the next build somewhere in 2025...

 

[birdie]

Another official confirmation that the 10nm Intel node is a failure (as there will be no desktop CPUs unlike rumored earlier) and that we'll have to wait at the very least till 2022-2023 before we get 7nm high performance Intel CPUs.

That'll be such a long wait since 2015 when SkyLake was first released.

 

[ratirt]

I was going to upgrade for 3000 series processor from my 2700x but I figured I will wait for the new Ryzen next year. It will be a nice boost in performance over my 2700x
I'm planning to buy a new Graphics Card next year as well so it will all go together

 

[R-T-B]

as there will be no desktop CPUs unlike rumored earlier

until they can build 10 nm desktop parts.

 

[Space Lynx]

I was going to upgrade for 3000 series processor from my 2700x but I figured I will wait for the new Ryzen next year. It will be a nice boost in performance over my 2700x
I'm planning to buy a new Graphics Card next year as well so it will all go together

Yep that is smart, only reason I am doing it this way is because I really needed an upgrade, and ryzen 3600 wa son sale for $177, and really it was only $160 as I intend to use the 3 free month game pass anyway, saving me money I had already planned to pay for that service.

I figure I can sell ryzen 3600 for 90 easily enough when 4800x launches, and then finally retire for awhile. I didn't want to go balls out just yet, as I feel 4800x might be another 10% IPC gain. I hope so anyway.

 

[Vayra86]

[...]

Until they can. Will they ever? Davis doesn't say, but we all know the answer. Especially because he's also saying they are pulling 7nm funding in faster. I really doubt they will do 10nm high performance desktop parts for 1 or 2 gens, especially because we haven't yet seen a single indication 10nm can even do that. Mobile and enterprise just screams 'low clocks and efficiency' at us -everything desktop is not.

 

[fancucker]

Oh my goodness Intel is doomed! After Zen finally managed to equal skylake (CFL) and across three iterations and two major nodes. Ice lake is already demonstrating 18% IPC uplift and tiger even further. Behind the process problems genuine uarch dev continued. Zen 3 better be good.

 

[ratirt]

Yep that is smart, only reason I am doing it this way is because I really needed an upgrade, and ryzen 3600 wa son sale for $177, and really it was only $160 as I intend to use the 3 free month game pass anyway, saving me money I had already planned to pay for that service.

I figure I can sell ryzen 3600 for 90 easily enough when 4800x launches, and then finally retire for awhile. I didn't want to go balls out just yet, as I feel 4800x might be another 10% IPC gain. I hope so anyway.

Sure enough. I was going to buy 3000 series but when AMD announced that they already have 3rd gen Ryzen ready and release is for next year I though I can wait a bit longer. My 2700x is in a good shape now so no hurry. Besides there is no point waiting for Intel anyway. 2021 is a long way from now and there might be delays. Anyway I wouldn't go for Intel since AMD and my 2700x is treating me so damn well. It would have been a blasphemy to go Intel these days. I mean for me, you guy do whatever you want.

 

[pjl321]

Don't get me wrong, I love that AMD are teaching Intel a lesson whilst bringing us great chips at amazing prices but I wouldn't write Intel off even in the medium term for 2 important reasons:

Firstly, Intel have only ever NOT delivered on 10nm, on the vast majority of other nodes their time tables have been mostly reliable and their nodes have always had the edge over the competition. Therefore, don't assume 7nm will be late, slow, have issues and not ramp up incredibly quickly. Yes Intel probably should have binned 10nm 2 or 3 years ago and focused solely on 7nm but they were too proud for their own good, they grossly underestimated the threat of AMD and they always believed the solution to 10nm is just around the corner on the next respin/retool, but it never was and still isn't, not truly.

Secondly, at any point Intel can buy capacity on TSMC, Samsung or GF and become the leader again. Let's say another year or so passes, 10nm is vaguely up and running but if you want the fastest Intel CPUs it's still better to make them on 14nm++++, this isn't sustainable. With the lessons learnt from 10nm Intel are much more realistic about the times scans and potential issues of 7nm and if they can't see it working properly until 2023 or beyond then what do Intel do? They can't be still using 14nm when TSMC is using a very mature 3nm that would be ridiculous, so Intel would be forced to make the decision to move Golden Cove to TSMC's 5nm in 2021 and maybe move its GPUs so Samsung's facilities. It will be a massive pride swollowing experience for Intel but it will keep them alive and competitive until they get their own facilities back on track. It would be the smartest thing to do and shareholders would probably demand it. So I'm not saying this will happen, I believe 7nm from Intel will come faster and ramp more quickly than anyone really expects but I'm saying if there are issues with 7nm too then Intel still has the same option as AMD, NVIDIA, Apple, etc to just use a 3rd party provider. 'If you can't beat them, join them!'

 

[yeeeeman]

Intel's biggest mistake IMO is that they didn't got on the multicore bandwagon at the same time AMD has. AMD has long been announcing that multicore will be the next thing so they adopted the chiplet architecture with great success. Intel should follow suit and create something similar with Tiger Lake cores on 10nm+.

Don't get me wrong, I love that AMD are teaching Intel a lesson whilst bringing us great chips at amazing prices but I wouldn't write Intel off even in the medium term for 2 important reasons:

Firstly, Intel have only ever NOT delivered on 10nm, on the vast majority of other nodes their time tables have been mostly reliable and their nodes have always had the edge over the competition. Therefore, don't assume 7nm will be late, slow, have issues and not ramp up incredibly quickly. Yes Intel probably should have binned 10nm 2 or 3 years ago and focused solely on 7nm but they were too proud for their own good, they grossly underestimated the threat of AMD and they always believed the solution to 10nm is just around the corner on the next respin/retool, but it never was and still isn't, not truly.

Secondly, at any point Intel can buy capacity on TSMC, Samsung or GF and become the leader again. Let's say another year or so passes, 10nm is vaguely up and running but if you want the fastest Intel CPUs it's still better to make them on 14nm++++, this isn't sustainable. With the lessons learnt from 10nm Intel are much more realistic about the times scans and potential issues of 7nm and if they can't see it working properly until 2023 or beyond then what do Intel do? They can't be still using 14nm when TSMC is using a very mature 3nm that would be ridiculous, so Intel would be forced to make the decision to move Golden Cove to TSMC's 5nm in 2021 and maybe move its GPUs so Samsung's facilities. It will be a massive pride swollowing experience for Intel but it will keep them alive and competitive until they get their own facilities back on track. It would be the smartest thing to do and shareholders would probably demand it. So I'm not saying this will happen, I believe 7nm from Intel will come faster and ramp more quickly than anyone really expects but I'm saying if there are issues with 7nm too then Intel still has the same option as AMD, NVIDIA, Apple, etc to just use a 3rd party provider. 'If you can't beat them, join them!'

I agree. The majority of people put an equal sign between Intel not being able to ramp 10nm and Intel not being able to increase IPC, design a good CPU, create great technology, etc, etc, which is plain wrong. The simple fact that an architecture from 2014 (Skylake) and a process from that same year can have so much life left in it and it took AMD 7nm, Zen 2 uArch, chiplets, infinity fabric and so many years of R&D to finally beat it, is the testament of the fact that Intel had and still has great talent and how good Skylake was and is. One other way of looking at this is taking what AMD had in 2014 and compare it to Ice Lake. It would be garbage, unlike Coffee Lake/Comet Lake/Ice Lake. Also, if 10nm would have worked, now we would have had 7nm with products like Alder Lake, Meteor Lake and most probably chiplets on HEDT platform (see EMIB), which would have beaten Zen 2 by a mile. Even Ice Lake has better IPC then Zen 2.
The fact that they screwed with fabrication tech, that is an entirely different thing, but don't you ever go into thinking Intel is staying idle on their designs.
Just as a clarification, I do love what AMD achieved with latest gen of products, I love their pricing and hope they push the boundary even further. But I also tend to see things in a realistic manner.

 

[pjl321]

Intel's biggest mistake IMO is that they didn't got on the multicore bandwagon at the same time AMD has. AMD has long been announcing that multicore will be the next thing so they adopted the chiplet architecture with great success. Intel should follow suit and create something similar with Tiger Lake cores on 10nm+.


I agree. The majority of people put an equal sign between Intel not being able to ramp 10nm and Intel not being able to increase IPC, design a good CPU, create great technology, etc, etc, which is plain wrong. The simple fact that an architecture from 2014 (Skylake) and a process from that same year can have so much life left in it and it took AMD 7nm, Zen 2 uArch, chiplets, infinity fabric and so many years of R&D to finally beat it, is the testament of the fact that Intel had and still has great talent and how good Skylake was and is. One other way of looking at this is taking what AMD had in 2014 and compare it to Ice Lake. It would be garbage, unlike Coffee Lake/Comet Lake/Ice Lake. Also, if 10nm would have worked, now we would have had 7nm with products like Alder Lake, Meteor Lake and most probably chiplets on HEDT platform (see EMIB).
The fact that they screwed with fabrication tech, that is an entirely different thing, but don't you ever go into thinking Intel is staying idle on their designs.

Intel's biggest mistake IMO is that they didn't got on the multicore bandwagon at the same time AMD has. AMD has long been announcing that multicore will be the next thing so they adopted the chiplet architecture with great success. Intel should follow suit and create something similar with Tiger Lake cores on 10nm+.


I agree. The majority of people put an equal sign between Intel not being able to ramp 10nm and Intel not being able to increase IPC, design a good CPU, create great technology, etc, etc, which is plain wrong. The simple fact that an architecture from 2014 (Skylake) and a process from that same year can have so much life left in it and it took AMD 7nm, Zen 2 uArch, chiplets, infinity fabric and so many years of R&D to finally beat it, is the testament of the fact that Intel had and still has great talent and how good Skylake was and is. One other way of looking at this is taking what AMD had in 2014 and compare it to Ice Lake. It would be garbage, unlike Coffee Lake/Comet Lake/Ice Lake. Also, if 10nm would have worked, now we would have had 7nm with products like Alder Lake, Meteor Lake and most probably chiplets on HEDT platform (see EMIB).
The fact that they screwed with fabrication tech, that is an entirely different thing, but don't you ever go into thinking Intel is staying idle on their designs.

Yeah, that is another point i didn't stress enough. If Intel said ok lets just release Golden Cove tomorrow on 14nm++++ it would have the best CPU by quite a margin. Ok they can't do this as it's not close to ready but my point is the main reason Intel has lost so much ground is not because 10nm is having massive issues and 7nm is still 2-3 years away, the reason is actually because they didn't have a new architecture to release on 14nm since Skylake in 2014. If they saw the problems they were having on 10nm and so released Ice Lake on 14nm 2+ years ago then AMD would still be playing catch up in most situations. Also, if Intel dropped the iGPU and used the whole die only for CPU then it could easily have a not too large 16 core die, clocking high and very good yields.

So in a nut shell, Intel didn't need to lose it's crown just because of the 10nm issues, it was the poor decision making after realising 10nm had issues that really did the damage.

 

[Vya Domus]

Firstly, Intel have only ever NOT delivered on 10nm, on the vast majority of other nodes their time tables have been mostly reliable and their nodes have always had the edge over the competition. Therefore, don't assume 7nm will be late, slow, have issues and not ramp up incredibly quickly. Yes Intel probably should have binned 10nm 2 or 3 years ago and focused solely on 7nm but they were too proud for their own good, they grossly underestimated the threat of AMD and they always believed the solution to 10nm is just around the corner on the next respin/retool, but it never was and still isn't, not truly.

It doesn't work like this, nodes are increasingly more difficult to make, if you screwed up your current node and since this relies a lot on previous developments the chances of getting the next right drop dramatically. This is a industry that doesn't wait for anyone, you can't make a good leading node on time you're out, look at GloFo.

20 years ago there were dozens of IC manufactures that had leading nodes or close to them. Now you can count them on the fingers of one hand, you think anyone cared that all of these guys had reliable nodes up to the point when they couldn't compete ?

Secondly, at any point Intel can buy capacity on TSMC, Samsung or GF and become the leader again. Let's say another year or so passes, 10nm is vaguely up and running but if you want the fastest Intel CPUs it's still better to make them on 14nm++++, this isn't sustainable. With the lessons learnt from 10nm Intel are much more realistic about the times scans and potential issues of 7nm and if they can't see it working properly until 2023 or beyond then what do Intel do? They can't be still using 14nm when TSMC is using a very mature 3nm that would be ridiculous, so Intel would be forced to make the decision to move Golden Cove to TSMC's 5nm in 2021 and maybe move its GPUs so Samsung's facilities. It will be a massive pride swollowing experience for Intel but it will keep them alive and competitive until they get their own facilities back on track. It would be the smartest thing to do and shareholders would probably demand it. So I'm not saying this will happen, I believe 7nm from Intel will come faster and ramp more quickly than anyone really expects but I'm saying if there are issues with 7nm too then Intel still has the same option as AMD, NVIDIA, Apple, etc to just use a 3rd party provider. 'If you can't beat them, join them!'

The moment Intel resorts to other manufacturers it's over. There is a reason they are avoiding this like the plague and prefer to do nothing with regards to the shortages that they face and it's not due to their pride. The reason it's simple, so much of their cash and future product development is invested in their foundries it would probably sink them if they decide it isn't worth it any more, it's the same issue that nearly killed AMD many years ago. TSMC is the only viable manufacturer for leading nodes and there is no way they can sustain Intel's volumes, Intel is in really deep shit despite their efforts to appear calm.

 

[londiste]

This is a industry that doesn't wait for anyone, you can't make a good leading node on time you're out, look at GloFo.

GloFlo CapEx was 4-5 times smaller a couple years ago when they were still planning 7nm process. TSMC CapEx is $15B this year, Intel's is in the same range and Samsung's semiconductor unit is at $20+B. GloFlo had knowhow and with the new TSMC cross-licensing has even more now but they are not likely to have the cash to compete.

The moment Intel resorts to other manufacturers it's over. There is a reason they are avoiding this like the plague and prefer to do nothing with regards to the shortages that they face and it's not due to their pride. The reason it's simple, so much of their cash and future product development is invested in their foundries it would probably sink them if they decide it isn't worth it any more, it's the same issue that nearly killed AMD many years ago. TSMC is the only viable manufacturer for leading nodes and there is no way they can sustain Intel's volumes, Intel is in really deep shit despite their efforts to appear calm.

Foundry business is still a huge part of what is Intel. I think Intel not outsourcing core things like CPUs is a simple money question. Outsourced production is a cost while in-house manufacturing effectively increases margins. Margin that outsourced manufacturing partner would cash in sounds a lot better going to the in-house manufacturing unit. Especially if they have/need the up-to-date manufacturing process anyway.

Intel has a lot of fabs and judged by manufacturing capacity they should not have shortages that we see. However, with what Intel is saying about their plans (and not commenting on whether these will materialize as expected) they are moving their fabs over to 10nm and considering the timeline in plans are probably starting to move fabs to 7nm or are already in process of moving. That would explain 14nm shortages.

 

[laszlo]

so i re-read the article:

"Intel CFO George Davis in an interview with Barron's commented on the company's financial health, and some of the reasons behind its rather conservative gross margin guidance looking forward to at least 2023 " : sht happens

 

[pjl321]

It doesn't work like this, nodes are increasingly more difficult to make, if you screwed up your current node and since this relies a lot on previous developments the chances of getting the next right drop dramatically. This is a industry that doesn't wait for anyone, you can't make a good leading node on time you're out, look at GloFo.

20 years ago there were dozens of IC manufactures that had leading nodes or close to them. Now you can count them on the fingers of one hand, you think anyone cared that all of these guys had reliable nodes up to the point when they couldn't compete ?



The moment Intel resorts to other manufacturers it's over. There is a reason they are avoiding this like the plague and prefer to do nothing with regards to the shortages that they face and it's not due to their pride. The reason it's simple, so much of their cash and future product development is invested in their foundries it would probably sink them if they decide it isn't worth it any more, it's the same issue that nearly killed AMD many years ago. TSMC is the only viable manufacturer for leading nodes and there is no way they can sustain Intel's volumes, Intel is in really deep shit despite their efforts to appear calm.

You could be right, who knows what the future will bring. I still believe Intel are making it more difficult for themselves than they needed to have but they are a long way from down and out. The next few years will be crucial, what shape 10nm is really in? Will they bring an entirely new architecture to 14nm as a stop gap? How quickly can they roll out 7nm? Time will tell but I'm just enjoying some real competition in god knows how long and with ARM and RISK V on the horizon things are looking very exciting.

 

[londiste]

How quickly can they roll out 7nm?

I am willing to bet this is the key question.
With every passing quarter and year, 10nm is getting more irrelevant and does not justify additional cost of widespread adoption in Intel fabs.

 

[ppn]

10nm is for chipsets. 7nm is for CPU. 7nm Intel is atleast 50% more dense than 5nm TSMC, so very close to 3nm TSMC. And at the time TSMC will still be using 5nm mainly for arm and such.

 

[ratirt]

Yeah, that is another point i didn't stress enough. If Intel said ok lets just release Golden Cove tomorrow on 14nm++++ it would have the best CPU by quite a margin. Ok they can't do this as it's not close to ready but my point is the main reason Intel has lost so much ground is not because 10nm is having massive issues and 7nm is still 2-3 years away, the reason is actually because they didn't have a new architecture to release on 14nm since Skylake in 2014. If they saw the problems they were having on 10nm and so released Ice Lake on 14nm 2+ years ago then AMD would still be playing catch up in most situations. Also, if Intel dropped the iGPU and used the whole die only for CPU then it could easily have a not too large 16 core die, clocking high and very good yields.

So in a nut shell, Intel didn't need to lose it's crown just because of the 10nm issues, it was the poor decision making after realising 10nm had issues that really did the damage.

I don't think it is that simple dude. Remember that Intel has had a problem with 14++++ nm products this year. It was basically the same node with some minor improvements. Besides the CPU design for 7nm or 10 nm may not be transferable to 14+++nm just like that so that Golden Cove for 14nm tomorrow's release, is more of a miracle than any possibility. Intel wanted to be innovative and keep the marketing going so they did release 10nm processors which was delayed anyway.
Call me crazy but you make it sound like Intel can just decide we will make iceLake 10nm and they do it. It took time for Intel to get the design of the CPU to work with the node. Believe me, if Intel could release the Sky Lake earlier, that would have happened earlier.

10nm is for chipsets. 7nm is for CPU. 7nm Intel is atleast 50% more dense than 5nm TSMC, so very close to 3nm TSMC. And at the time TSMC will still be using 5nm mainly for arm and such.

How can you say it is more dense when it hasn't been released yet? Where do you get that information? How is it denser than 5nm TSMC when desktop 5nm from TSMC is not released yet as well.
What a crap you say. 7nm is like 3nm what is this stuff? 7nm is 7nm 5nm is 5nm what are you talking about?

 

[pjl321]

I am willing to bet this is the key question.
With every passing quarter and year, 10nm is getting more irrelevant and does not justify additional cost of widespread adoption in Intel fabs.

But surely 10nm only becomes irrelevant if you are ready to use someone else's fab or close your doors, if neither of these options are viable then you can only work with what you have. TSMC will 100% have 5nm before Intel has 7nm and they really could be on 3nm before Intel is on 7nm so in a way 7nm will be irrelevant very soon but it's all Intel has so it's not irrelevant to them.

The key now for Intel is it brings cutting edge architectures to whatever node they have, even if that is 14nm or bite the bullet and move production of some of their chips to another fab just so they don't lose too much market share.

 

[kapone32]

Fact: The last time AMD was faster than Intel; 2005
Fact: The last time Intel had to worry about AMD (Or anyone else for that matter) 2017

That is 12 years of complete dominance in marketing, sales and hardware. The other thing it can and has done is stagnate. I look at where Intel is today the same way I looked at Tahiti when it was finalized by ATI/AMD. That series of GPU was so popular and nice for the consumer that I am sure there are still a ton of people with 7950, 290,280,390 cards but what it did was stagnated the GPU division of AMD (by then) because they could not come up with anything better without a heavy investment of time, money and brain power (which became Polaris and eventually Navi). Intel is huge and I mean huge they have more foundries around the world than most of us are aware of. The problem with being that big is it will take Intel longer to respond to AMD's (well TSMC's) 7nm and so on as not only will they have to make changes to their current 14+++ but also try to refine the albatross that is 10nm. Unfortunately, based on the timeline provided by Mr Davis they may be too late to take the wind out of AMD's sails fully as we have been getting more concrete announcements from them. Every news article from Intel lately has been about promise and nothing more. They are simply not nimble enough to respond to AMD in even a year's time. I have zero doubts though that when Intel does release 10NM nad 7nm desktop parts they should absolutely fly but who knows where AMD will be then too.

 

[pjl321]

I don't think it is that simple dude. Remember that Intel has had a problem with 14++++ nm products this year. It was basically the same node with some minor improvements. Besides the CPU design for 7nm or 10 nm may not be transferable to 14+++nm just like that so that Golden Cove for 14nm tomorrow's release, is more of a miracle than any possibility. Intel wanted to be innovative and keep the marketing going so they did release 10nm processors which was delayed anyway.
Call me crazy but you make it sound like Intel can just decide we will make iceLake 10nm and they do it. It took time for Intel to get the design of the CPU to work with the node. Believe me, if Intel could release the Sky Lake earlier, that would have happened earlier.

I didn't explain that very well, you are right they can't click their fingers and change a chip that was designed for 10nm and now make it on 14nm. What I meant was Intel should have seen how bad things were with their 10nm and re-worked Ice Lake to work on 14nm. My understanding is that Intel has learnt this lesson and all architectures moving forward are cross node compatible, meaning if they did have an issue or shortage on a certain node they can use another node to plug the gap.