Question Intel Q3 Results

[jpiniero]

https://d1io3yog0oux5.cloudfront.net/_2b116bffa4ca8890f09df4510f3c972f/intel/db/887/8873/earnings_release/Q3+22_EarningsRelease.pdf

They actually technically made a profit (albeit an small operating loss) but CCG revenue was down 17% and Datecenter was down 27%.

Says $3B in "Cost Reductions" in 2023.

 

[eek2121]

Can I say this is or will I get booed at?

Intel is dying. No, I really mean it. If their future nodes don't deliver on time they are dead*.

First networks, what really matters in Networking is modems.
Qualcomm is king there, already wifi 7 and 5G modems.
x86 wise AMD is destroying Intel in Datacentre and will do it in laptops next year.

GPUs are pathetic from Intel. I don't care if it's Intel's first time. The market does not care.

Desktop is only the only place where Intel actually competitive now but DIY and OEM desktop has barely any profit. Companies are moving to mobile and laptops. Where AMD, ARM and Apple excel at.

I really think Intel is doomed unless they deliver and execute with NO delays anymore. I want Intel to be great but so far money talks and Intel is failing.

AMD has great Datacentre chips, great desktops chips, great laptops chips and AMD chips are used a LOT in consoles, handhelds and what’s Intel making money in these days is not enough for Intel to survive

This decade a LOT of things will change. Let’s see.

Disagree. Intel won’t die. They will evolve. What that looks like is anyone’s guess.

If IFS misses targets again, one scenario we could see is that Intel spins off IFS into it’s own entity. At that point, IFS either gets a period of exclusivity or they don’t. After that? Intel can use anyone they want.

Seems vague, would be weird if MTL-P (Laptop) releases before MTL-S since MTL-P would end up releasing soon after Raptor Lake-P, though a MTL-S in the Summer or any time soon as possible against Zen 4 V-Cache (which let's be honest even the 13900KS will be left behind in scenarios that love loads of cache).

Intel launched mobile parts first for 10nm (Ice Lake, Tiger Lake) so I would not be surprised if they did it again.

 

[Exist50]

@TheELF

Yeah man. There are certainly uses for the graphics they have developed. Consumer/gaming dGPUs is probably not going to be one of them much longer. He announced up to 13 billion in spending cuts in only 24 months. Gaming dGPUs aren't going to be profitable any time soon. Bean counters axe is coming for it.

Gelsinger has been quite unambiguous in stating that he sees graphics as a vital, long-term investment. It even seems to be personal, with graphics being the last thing he didn't accomplish before leaving as CTO. I think he'll be willing to sacrifice many things before graphics.

 

[jpiniero]

I think he'll be willing to sacrifice many things before graphics.

Definitely not the dividend. Or the fabs. And I think we are headed for both of those going.

As I stated before, the whole reasoning for getting into gaming GPUs has been blown up... and now they can't really afford to blow money on stuff that doesn't make sense.

 

[igor_kavinski]

I think it would be dumb of Intel to dump GPUs after pumping so much R&D budget into it. They will probably continue with their GPU development. Just in "low power emergency lights" mode. If they can start producing GPU dies on 20A and 18A, that should theoretically make them very competitive, at least as far as production cost is concerned.

 

[Exist50]

Definitely not the dividend. Or the fabs. And I think we are headed for both of those going.

If it comes down to the IFS vs AXG, then I agree that Intel will prioritize the former. But I don't think they're that desperate yet.

As I stated before, the whole reasoning for getting into gaming GPUs has been blown up...

I'm not sure what you're referring to, but they certainly didn't get into the market because of crypto. And the need for GPUs remains the same. Not just for gaming, but for AI, workstations, and HPC. That market isn't going away any time soon.

 

[jpiniero]

If it comes down to the IFS vs AXG, then I agree that Intel will prioritize the former. But I don't think they're that desperate yet.

With Server not profitable they aren't going to be able to afford to continue with the fabs. Unless they can turn it around, it's only a matter of time before they are forced to spin the fabs off. And they don't have that much time. For now they are hedging by mortgaging the future (see the 'co-investment')

I'm not sure what you're referring to, but they certainly didn't get into the market because of crypto.

They got into gaming GPUs because they wanted it to fill the fabs (whoops) and they thought Radeon was toast (whoops). BTW, I expect Raja to leave (willingly most likely) in the not too distant future.

 

[Exist50]

With Server not profitable they aren't going to be able to afford to continue with the fabs. Unless they can turn it around, it's only a matter of time before they are forced to spin it off.

Sure, their financials are bad, but they're not yet losing large sums of money, and they still have quite a bit of cash on hand even before getting into loans (A+ credit rating gives them reasonable options). Long term, yes, obviously the current trajectory is unsustainable, but that assumes no broader economic improvement, no product execution improvement, and no returns from IFS. I'm certainly not going to claim that any of those are guaranteed, but that's basically the most pessimistic outlook possible.

They got into gaming GPUs because they wanted it to fill the fabs (whoops) and they thought Radeon was toast (whoops).

Where did you get either of those things from? Their GPUs don't even use the fabs, nor are they rumored to in the near future. And the public rhetoric, if nothing else, is about the need for parallel compute, which seems like a perfectly sane argument.

 

[jpiniero]

Where did you get either of those things from? Their GPUs don't even use the fabs, nor are they rumored to in the near future.

When they started, they were intending to use the internal fabs. And that was given as a big reason as to why they got in the market in the first place.

 

[IntelUser2000]

Gelsinger has been quite unambiguous in stating that he sees graphics as a vital, long-term investment. It even seems to be personal, with graphics being the last thing he didn't accomplish before leaving as CTO. I think he'll be willing to sacrifice many things before graphics.

Intel also said dGPUs allows Intel monetizing the area they weren't able to previously make money on. They mean that with iGPUs they don't get money from all the R&D they do(technically I disagree but I also see their point). It's client gaming GPUs that do this, not HPC.

 

[Exist50]

When they started, they were intending to use the internal fabs.

Sure, for Ponte Vecchio and Arctic Sound, but DG2/Alchemist was never going to use Intel fabs, and being on an Intel process is clearly secondary to being in the market at all.

And that was given as a big reason as to why they got in the market in the first place.

Where?

 

[jpiniero]

Where?

Raja said that like 5 years ago when they started.

 

[Exist50]

Raja said that like 5 years ago when they started.

I don't remember that. Would you happen to have a quote?

 

[moinmoin]

AMD had contracts with MS and Sony and Apple during the bulldozer era/early Zen.

Indeed those were life savers for AMD, especially since they covered important R&D AMD likely wouldn't have been able to afford otherwise. But note that for years on people argued that those contracts were bad business for AMD due to the low margins.

Intel is not yet at the place that something like that is conceivable for it. All of Intel's decisions are still about keeping margins high, not for keeping the lights on as was the case with AMD. The two are just not comparable (yet?).

"Right-sizing". What a BS word for downsizing/layoffs.

"improved sales and marketing efficiency"

Those are just tacit admissions that the core business is crashing and the whole sales and marketing overhead that Intel previously served well (to the point it turned into a Xerox 2.0 for a couple of years, recall that Jobs quote) is no longer affordable to today's Intel.

I'm not sure how much of it Pat planned to do anyway at some point in the future, but the current macro situation most certainly sped up a lot of those decisions and developments.

 

[jpiniero]

I don't remember that. Would you happen to have a quote?

Probably not. After all it was 5 years ago. But that's why it took so long for anything to come out. When the PV 10 nm disaster happened, they also decided to port the gaming GPUs to TSMC as well. IMO they should have canned the gaming GPUs then, but that's water under the bridge now.

And don't underestimate the AMD part. There's only so much room in the market... and even AMD is going to have problems if they can't convince people to buy $1k+ dGPUs regardless of it's competitiveness.

 

[Exist50]

When the PV 10 nm disaster happened, they also decided to port the gaming GPUs to TSMC as well.

As far as I'm aware, DG2 was never planned to use an Intel process. It was really only Ponte Vecchio that they were forced to port because of the 7nm issues.

And don't underestimate the AMD part. There's only so much room in the market... and even AMD is going to have problems if they can't convince people to buy $1k+ dGPUs regardless of it's competitiveness.

The market is huge and still growing; just look at Nvidia. And they don't even need to make a large profit on gaming GPUs so long as they can justify them as a complement to their CPUs/overall ecosystem.

 

[A///]

intel has a lot of groundwork ahead of them. Raja took sabbatical 5 years ago and never returned. it's of my opinion new leadership at amd slowly began pushing raja out roughly a year before then.

 

[IntelUser2000]

The market is huge and still growing; just look at Nvidia. And they don't even need to make a large profit on gaming GPUs so long as they can justify them as a complement to their CPUs/overall ecosystem.

They no doubt do have ambitions as they said the dGPU part was to "monetize what previously wasn't monetized".

With Battlemage having up to 2560EUs, and if they keep up on improving the drivers, things can change quickly.

 

[moinmoin]

Oooof, I missed that Intel is back to BS economics:

Intel Cuts Fab Buildout by $4B To Pay Billions In Dividends

First Net Loss In Over 30 Years, Cutting Fab Buildouts, But “Committed To Growing The Dividend”

www.semianalysis.com

At least stocks are seeing a great upswing for that ludicrous dividends promise paid with future competitiveness.

 

[Markfw]

Oooof, I missed that Intel is back to BS economics:

Intel Cuts Fab Buildout by $4B To Pay Billions In Dividends

First Net Loss In Over 30 Years, Cutting Fab Buildouts, But “Committed To Growing The Dividend”

www.semianalysis.com

At least stocks are seeing a great upswing for that ludicrous dividends promise paid with future competitiveness.

Makes you wonder how they ever got as big as they are with stupid decisions like this.

 

[IntelUser2000]

Makes you wonder how they ever got as big as they are with stupid decisions like this.

They get big because of good decisions and then they make bad decisions and become smaller or bankrupt. Ultimate big corp is government. Such is life.

 

[lightisgood]

Probably not. After all it was 5 years ago. But that's why it took so long for anything to come out. When the PV 10 nm disaster happened, they also decided to port the gaming GPUs to TSMC as well. IMO they should have canned the gaming GPUs then, but that's water under the bridge now.

Agree.
I remember that Intel canned Xe HP built by 10nm SF.

 

[TheELF]

Oooof, I missed that Intel is back to BS economics:

Intel Cuts Fab Buildout by $4B To Pay Billions In Dividends

First Net Loss In Over 30 Years, Cutting Fab Buildouts, But “Committed To Growing The Dividend”

www.semianalysis.com

At least stocks are seeing a great upswing for that ludicrous dividends promise paid with future competitiveness.

Intel clearly stated that they are going to shut down some departments, these departments will not need to be build out anymore in the future, so that is cap expenditure they can safe on.
Unless they can show some analysis of exactly what and where these cut backs are, then this could just as well be 4bil that were planned to expand on making more optane (and whatever else) in the future.

Clickbait is gonna clickbait...

The FABs for the CHIPs act are dealings with the government and they will not risk anything going wrong with that.
Don't be fooled, intel increased their FABs and everything else by 12b in the last year and still put 3b under their mattress.
They are not losing money, they are using money, instead of having it as an earning they are putting it into FABs.

Intel Reports Third-Quarter 2022 Financial Results

www.intc.com

(In Millions; Unaudited)

Oct 1, 2022​

Dec 25, 2021​

Property, plant and equipment, net

75,763​

63,245​

Retained earnings

71,024​

68,265​

 

[DrMrLordX]

Amazingly the stock is up after hours. Probally because of the layoffs announcement.

The street loves layoffs.

Seems vague, would be weird if MTL-P (Laptop) releases before MTL-S since MTL-P would end up releasing soon after Raptor Lake-P, though a MTL-S in the Summer or any time soon as possible against Zen 4 V-Cache (which let's be honest even the 13900KS will be left behind in scenarios that love loads of cache).

If Intel can pull in Meteor Lake-P quickly enough, they can ignore the Raptor Lake mobile suite and just roll with that. Meteor Lake-S hasn't been demonstrated to be anything special anyway.

If Intel were to go through the same we won't look at just 20% layoffs but a complete new kind of Intel.

We'd be looking at a fabless Intel.

 

[poke01]

The street loves layoffs.



If Intel can pull in Meteor Lake-P quickly enough, they can ignore the Raptor Lake mobile suite and just roll with that. Meteor Lake-S hasn't been demonstrated to be anything special anyway.



We'd be looking at a fabless Intel.

If Intel goes fabless. TSMC will be the crown jewel of semiconductors then.

 

[moinmoin]

Intel clearly stated (...)

Clickbait is gonna clickbait...

The problem is not that Intel is cutting back, that can be a valid economic necessity. It's that Intel in that situation still promises dividends and continuously rising ones at that. The latter comes at a huge cost which means Intel will have to cut back well beyond economic necessity, it's not unlike the stocks buyback program Intel had going at a time node development needed much more investment than it got.

So giving up future competitiveness for some short term shareholder profiteering.