Question Intel Q3: Ouch

[jpiniero]

https://d1io3yog0oux5.cloudfront.net/_c675d259e2193ee11667e182e2fab5ca/intel/news/2020-10-22_Intel_Reports_Third_Quarter_2020_Financial_1424.pdf

Desktop volume down 16%, Notebook up 16% but ASP down 7%.

Q4 forecast is bad too, they are expecting 14% decline in revenue compared to Q4 last year.

 

[esquared]

Ladies and gentlemen, please keep the politics out of this thread.

<--------------P&N is thataway.


esquared
Anandtech Forum Director

 

[Ajay]

Oops, Okay.

 

[Systems analyst]

The top three risks Intel faces seem to be:

1. TSMC have a superior production process.

2. AMD have superior micro-architectures for X86.

3. In the medium term, ARM's architecture may evolve to be an increasing threat to X86.

Intel may end up like IBM, with a legacy role, and some interesting new developments, but no longer dominant. They probably need to focus on fewer things.

 

[gorobei]

I Have to say I'm surprised seeing AMD still having only 20% of the desktop market and even more surprised mobile where the advantage is much less pronounced has the same marketshare. Since there is no DIY mobile market AMD's OEM market share for desktops must be even lower than mobile. Can't wrap my head around that.

if we assume the DIY market is indeed about 80% AMD and 20% Intel (as MindFactory and Amazon figures suggest) this means, if we assume OEM vs DIY volume is about 90% vs 10%, 87% of all OEM desktop systems is still Intel. if the DIY fraction is larger Intels OEM desktop market share goes very quickly towards 100%.

Since a large proportion of OEM desktops are gaming PCs I find this very hard to believe...something is off in that desktop market share chart.

oems will lag on market trends by at least a year or more due to the time it takes to prep and validate a mobile design. desktop can be faster but customer/tech support still has to set up training. ryzen 3000 has been out a bit over a year and dominates diy. the oems may have needed to see consistent customer demand to justify shifting purchasing amounts/sourcing changes. with zen3 seemingly taking the gaming single threaded crown there is no way for oems to ignore that. but that is still only in the consumer market.

in the business market it doesnt matter how good or better amd is, oems have to buy from someone who can meet demand. intel supplies ~90% of the chips that go into those machines. even if amd is better, the oem cant commit if amd cant provide enough cpus. intel cranks out tons of chips, nothing any of us are looking to buy but business arent that discerning. tsmc would need significant capacity expansion to makeup the volume of intel losing share. it will take time.

 

[Thala]

The top three risks Intel faces seem to be:

1. TSMC have a superior production process.

2. AMD have superior micro-architectures for X86.

3. In the medium term, ARM's architecture may evolve to be an increasing threat to X86.

Intel may end up like IBM, with a legacy role, and some interesting new developments, but no longer dominant. They probably need to focus on fewer things.

Actually both, focus on fewer things and at the same time diversify. As example they could the Atom line of cores finally discontinue and use the freed engineering resources for other things.

 

[jpiniero]

Actually both, focus on fewer things and at the same time diversify. As example they could the Atom line of cores finally discontinue and use the freed engineering resources for other things.

Atom does make them a ton of money though, or at least did. The IoT business I believe mostly uses Atom, and it was a decent money maker before the virus.

Core is too big for what those markets need.

 

[maddie]

The top three risks Intel faces seem to be:

1. TSMC have a superior production process.

2. AMD have superior micro-architectures for X86.

3. In the medium term, ARM's architecture may evolve to be an increasing threat to X86.

Intel may end up like IBM, with a legacy role, and some interesting new developments, but no longer dominant. They probably need to focus on fewer things.

Ha, that brought a laugh, "They probably need to focus on fewer things". I see stores in my country claiming that they specialize on about 20+ lines of products.

I do understand what you mean however.

 

[chrisjames61]

Selling off parts of the Company to fluff the Stock Price seems ominous.

Buybacks were illegal for a long time for good reason.

 

[chrisjames61]

This is getting dangerously close to some form of burning platform memo...

Kind of like Nokia and we know how that turned out.

 

[chrisjames61]

TSMC is the last game in town for leading edge digital nodes. If I were the US Government I'd be sending more soldiers to Taiwan. Although I'm not sure anyone in D.C. has even an inkling of how important this shift is... Intel missing the mark on fab tech so badly is becoming a national security issue. While it's good for competition now (AMD, Arm, etc), it's not going to be good for the US in the long run. It must be pretty weird being the CEO of TSMC, which now also means being the head of (possibly) the most important private company on earth.

Thank Apple for this. Also Intel has much of the blame telling Apple they weren't interested in making the processor for the iPhone.

 

[ondma]

The top three risks Intel faces seem to be:

1. TSMC have a superior production process.

2. AMD have superior micro-architectures for X86.

3. In the medium term, ARM's architecture may evolve to be an increasing threat to X86.

Intel may end up like IBM, with a legacy role, and some interesting new developments, but no longer dominant. They probably need to focus on fewer things.

Not sure I would agree with number 2. Intel has some good architectures, they just lack the process tech to utilize them.

 

[uzzi38]

Not sure I would agree with number 2. Intel has some good architectures, they just lack the process tech to utilize them.

Well Superfin and Enhanced Superfin is supposedly fixing that, so all we have to do is monitor what happens over the next year and a half or so to decide whether or not process node is Intel's only problem.

 

[uzzi38]

Actually both, focus on fewer things and at the same time diversify. As example they could the Atom line of cores finally discontinue and use the freed engineering resources for other things.

Dear god no, please not the one team at Intel that have an actually interesting roadmap and can actually execute. Anything but them!

 

[Thala]

Dear god no, please not the one team at Intel that have an actually interesting roadmap and can actually execute. Anything but them!

We got Jasper Lake and Lakefield from them this year - something you would not even want to put into a phone 2 years ago - performance wise - and power wise.. This is as far as the consumer front is concerned - maybe the Atoms do better in IoT.

 

[cellarnoise]

I may need to cough up and invest in Intel soon! Now that I have a few non-interest bearing green backs...

I followed cpu's hard back in the the 80's and 90's. Amd made me good money (could have been better) by reading up on the coming of the Athlon. Mid 1999 to Mid 2001, and I was late and sold out early, but it was a good ride with not enough Moolaw! I also purchased a home full tower Athlon 1200, overclocked to 1333 and then made a few 80 mm to 60 mm heatsink adaptors. I should have gone into the air cooled heatsink biz... Still have my Athlon tower taking up storage space, though I think it is AT case.... GPU's (graphic cards... vs. other cards like audio) where just starting to be interesting.

Also was a bit late to the Intel comeback...., but made good there also. IF I had been riskier probably could have retired on this comeback, that was in the works at the time (new cpu architectures then took like 5 years from initial thought to consumer product - kind of like now? - I don't really know more than off the common interwebs sites like Anand.)... It seems to get hard to track cutting edge tech, without following rumors on 7 different anti-social media platforms including video based... Who has the time without inside information?

I really messed up (f) and do believe I saw the Ryzen wave coming.... But I did not have much mooney... and did not get on the ride. Wasn't AMD like $2.50 then, OMG! Though I now have two - soon to be 3 AMD desktops.... All consumer and zero investor For the Loss!. And maybe an AMD video card? Crazy... So I am now watching my AMD tech "investment" decline in value faster than water over a fall, but having fun!

World politics as they are, and the extreme value of leading edge transistor manufacturing and the value that maintaining nationalities now equates to military / geo-political power (currently based on cutting edge, but largely iterated upon, old silicon tech).... I am now thinking Intel might be a good deal to invest in. National silicon-wave maybe coming? I'm not talking breast or other implants as that has been a thing also long as Intel Core?

Just need them down another 10 or maybe 20% and then IN! Or we lose our shirts and WW3? Haha? Haha??

 

[cellarnoise]

I'll share another thing about underdog tech companies, that are well established like AMD. Generally when the leadership seems confident or cocky, but not overly so, they tend to be on a upswing for a few years. New AMD graphic silicon announcements comes out soon and they still seem confident about that....

They also tend to get bought out over time. Crazy. Though monopolyism does come into play.

 

[uzzi38]

We got Jasper Lake and Lakefield from them this year - something you would not even want to put into a phone 2 years ago - performance wise - and power wise.. This is as far as the consumer front is concerned - maybe the Atoms do better in IoT.

Gracemont getting AVX2 support will open up a lot of doors and also lead to some very interesting products. Just give them some time

 

[TheELF]

So, you are the only one here that thinks Intel is doing fine. I won't argue that they are making money, but with their product stack, that can't go on forever. Especially when Zen 3 and Milan come out. Right now at least they have the gaming crown. That will end very soon, then what do they have ????

One weak quarter doesn't mean anything, intel had similarly weak 2nd and 3rd quarters in 2019 but yearly they still raked in the mullah.

I won't argue that they are making money, but with their product stack, that can't go on forever.

And it doesn't need to go on forever, why would anybody even think that? I know that you think that intel can't tell a computer chip from a potato chip but I'm sure they can.
If you have a business and you keep making the same amount of money (twice actually) without changing anything, why would you change anything?

Especially when Zen 3 and Milan come out. Right now at least they have the gaming crown. That will end very soon, then what do they have ????

Intel will still have the volume crown and with the new fab now working they will have even more volume.

 

[TheELF]

are now net negative in terms of liquid assets to debt (i.e. they've burned through their cash reserves to prop up their stock price and are now officially overall in debt). They are now down to $18.3 billion cash on hand with a total debt load of $36.6 billion.

That has been so for a long time.

So, now they are selling a piece of their company in order to be able to afford another $9 billion stock buy back to try and keep their stock price up.

They wouldn't need to sell that, they made 50bil net income in the last 3 years, they can easily afford to pay the buybacks from that.
They could just as well be selling nand tech because they think that it will be much less worth until 2025, ssd prices are already dropping hard.

 

[TheELF]

Why aren't they investing more into manufacturing to take the lead from TSMC?

Because "lead" is just a gimmick to sell more product and intel is already selling all of the product they can make.

 

[TheELF]

The only thing I have seen as new is a limited quantity of Rocket lake laptop chips that have mixed reviews. They are losing badly in server, HEDT, and in 2 weeks, desktop. And their laptops are fighting with the new Renior chips.

If they are losing in the same way they lose in desktop then they are losing in things nobody cares about or in things that people have migrated to specialized hardware gpus or arm.
In desktop intel loses in 3d render video transcoding and 7zip which are things that no desktop user will ever even run.
Well mabe 20% of them will and that's why AMD has 20% market share, who knows.

 

[turtile]

Because "lead" is just a gimmick to sell more product and intel is already selling all of the product they can make.

They can't make more product because they haven't shrunk their process. 10nm's yield/cost is so bad that it isn't viable until now.

Maybe Intel should just stay at 14nm forever. Everything lower is a 'gimmick'. Yeah, tell that to server customers...

 

[TheELF]

However, Intel has to fund their entire fab operation and R&D from that as well whereas AMD doesn't, that's already built into the price they pay to have their chips manufactured.

Yes AMD is paying for TSMCs fab operation and R&D and margin on top of everything else while intel doesn't.

On top of that, when Intel fell behind in fab technology to the point that they now struggle to make a competitively performing product, all of a sudden having these fab facilities costing tens of billions of dollars starts to look more like a liability than an asset. Intel is certainly not a lost cause at this point, but they are deep in the weeds and (personal opinion) I don't think they have the leadership to find their way back right now. We'll see what the next few years holds, but I'd be pretty nervous right now if I was a key stakeholder in that company.

This would make sense if intel wouldn't be making 4 (to 6) bil a quarter, clean money in the pocket, from these fabs.
You are talking like intel is losing money from the fabs but intel is making a lot of money.

 

[SAAA]

The top three risks Intel faces seem to be:

1. TSMC have a superior production process.

2. AMD have superior micro-architectures for X86.

3. In the medium term, ARM's architecture may evolve to be an increasing threat to X86.

Intel may end up like IBM, with a legacy role, and some interesting new developments, but no longer dominant. They probably need to focus on fewer things.

1. They have a greater capacity and execution, not so sure on superior process aside from density (thinking 5nm).
14nm ++++, 10 super fin and the successor from Intel are/will be definitely great nodes performance wise. High power is 99% due to higher clocks, lower core counts on most of the lineup and… 14 nm being 14 nm so way too old by now. Hopefully obsoleted in a year with all the products on 10nm minimum.

2. AMD have a comparable architecture today, after Zen 3 releases. First time in years it beats 5 years old Skylake at gaming, cool. How much of that is due to 70 MB or so of cache?
Tiger Lake is out, Rocket lake in a few months, then Alder Lake will show up and it can't be a small step.
I mean Atom cores have pretty much reached IPC parity with Skylake, and those are tiny cores 1/4th as big as Tiger/Ice lake who are merely 20% faster for clock (same goes for Zen 3...)
If Golden Cove isn't 50% faster per clock at several times the area they better kill the design now and start back from the Atom line, pushing clocks/IPC on those.

3. ARM is a threat not design wise but product wise. The rise of smartphones and chips produced outside Intel, so giving money to TSMC/Samsung etc. I agree Intel messed up big here 10+ years ago not seeing the future of consumer products and the billions they have lost trying to get in too late.

One weak quarter doesn't mean anything, intel had similarly weak 2nd and 3rd quarters in 2019 but yearly they still raked in the mullah.

And it doesn't need to go on forever, why would anybody even think that? I know that you think that intel can't tell a computer chip from a potato chip but I'm sure they can.
If you have a business and you keep making the same amount of money (twice actually) without changing anything, why would you change anything?

Intel will still have the volume crown and with the new fab now working they will have even more volume.

That'ss Intel's saving grace currently. If AMD could fab 10X the CPUs tomorrow they would happily do so.
Ryzen 5000 could be twice as fast as Intel best but it means nothing if they can fab a few of them and price shoots up when stock reaches 0.
Market is both product quality -and- quantity/availability.

 

[TheELF]

They can't make more product because they haven't shrunk their process. 10nm's yield/cost is so bad that it isn't viable until now.

Maybe Intel should just stay at 14nm forever. Everything lower is a 'gimmick'. Yeah, tell that to server customers...

10nm (well the CPUs they are going to make on 10nm) will have more IPC (more transistors) and more cache (more transistors) so overall it's probably going to be very close as far as overall volume goes.