Question Intel Q4 Results

[jpiniero]

https://d1io3yog0oux5.cloudfront.net/_8020e84ca449012613cb7a51a8b3788c/intel/db/887/8894/earnings_release/Q4+22_EarningsRelease.pdf

https://d1io3yog0oux5.cloudfront.net/_f2b8ef2fed169a734498f0ef90617704/intel/db/887/8894/earnings_presentation/Q4%272022+Earnings+Deck_Final.pdf

Yuck. Revenue down 32% and again they lost money.

CCG revenue down 36%, DCG down 33%... that in Q4 compared to previous year.

Q1 Guidance is even worse... revenue down 40% YoY.

 

[BorisTheBlade82]

Many workers at Intel are highly specialized. Sure they could get a job flipping burgers if they absolutely had to, but most people will want to continue working in their area of expertise.

If the market for their specific job has cratered, it might even require going back to school if they want another skilled labor position. Many people would prefer a pay cut when the alternatives could be a lot worse.

I'm gonna bet several farms that a lot of employees layed off by Intel will be highly appreciated by the industry in general.

 

[A///]

If he is able to succeed remains to be seen.

Investors won't give a damn. They'll vote to have him removed. You're confusing consumers with investors here. They want results fast. His moves will be reaped by the next guy.

 

[moinmoin]

While I don't think it's hard to argue that they're not doing as well as they used to, "trouble" feels like too strong of a word or ignores several years of problems that have led to this point.

The long lead times for products makes it a lot harder to argue that they're in trouble right now. We're just seeing a realization of the problems they were having years ago and even if Intel has made a perfect course correction this very day that fixes all of their woes, we won't see it fully play out in the market for a few years.

AMD was in a far more precarious situation not all that long ago and now we're seeing them reap the benefits of their own successful course correction. Intel is just going to have to deal with unruly investors and that they won't be able to throw their weight around like they used to be able to back in the day.

That's an odd way to dance around the word "trouble".

While what we are seeing now are the result of stuff years in the making, they'll absolutely affect outcomes of what is being worked on right now. Or was, before layoffs and saving plans were introduced.

So it's absolutely justified calling this troubles, since it's absolutely affecting Intel's future. Fabs are being delayed, IFS' services (like the Risc-V programme) are already being reduced before it really started, and projects kickstarted around the time Keller was at Intel are likely stopped to focus on what current leadership considers core products.

Whether that focus is correct and how much lower salaries and layoffs with subsequent specialist employees leaving hurts Intel we will see at the end of the currently starting lead times.

 

[BorisTheBlade82]

Has this great article by Ian already been mentioned here?

What Will Intel Cut Next?

Revenue Down, Margin Down: Q4 2022

morethanmoore.substack.com

My favourite passage:

When Gelsinger initiated this strategy to resteer Intel in the right direction, on the whole it was seen as a positive. I was certainly looking forward to it, however we expected that correction to be as if a boat was steering 30 degrees to the right, and you'd plot a 45-degree correction to slowly get back on track. With the numbers we're seeing, it feels more like a 135-degree drift and it's amazing to see that this is what it's going to take. We've all heard the phrase 'go big or go home', well Gelsinger is going bigger than everyone else it seems. That being said, it has to weather the storm and they've said that it's going to be a rough few years, but the street didn't expect it to be this rough.

 

[scineram]

Mmmm, noice!
Looks like somebody expects this stock to pay a dividend or two.

 

[turtile]

Mmmm, noice!
Looks like somebody expects this stock to pay a dividend or two.

For someone that made over 170 million last year, 250K is nothing...

 

[Mopetar]

OK, I agree with long lead times, but come on, Ryzen 1800x was launched in 3/2017, that almost 6 years ago. They still have not really come out with a compelling server product, and desktop ADL/RKL sucks power like crazy (yes Zen 4 is more than some wished, but still way less than those) So in 6 years thats the best they can come up with ?

I think their biggest failing with server is not getting E-cores into those products. They've shown that they can actually now compete on desktop with AMD in the heavily multithreaded workloads where they previously got beat so badly that Zen was an obvious choice.

They do need to rework their P core as well, but mostly in a way to reduce the size of it. That's really what's killing them in server. They can't fit as man cores as AMD can and that will limit them.

Power guzzling is mainly a side effect of them trying to hold on to gaming performance and getting MT results that don't look as bad. They aren't quite as efficient at low power levels, which does add another limit in server, but at same TDPs, they perform much more efficiently.

It's pretty obvious that they didn't take Zen seriously and were caught flat-footed by what AMD had done and likely wasted a few years thinking they could just keep on doing what they were doing. Unfortunately for them, AMD improved their core considerably faster than Intel was able to improve their raw compute. AMD's chiplet strategy is also more economical.

That's an odd way to dance around the word "trouble".

Perhaps the reason that I really dislike the term is that it can mean just about anything and isn't well defined. Does trouble mean that the company might go under in the near future? Is it just a short period of bleeding market share and revenue while getting more competitive products to market?

Another reason that I don't like using the term in this context is that what Intel is doing can help the company. I'm reminded of another similar example where a CEO killed off a lot of non-core products in order to refocus the company. It was Steve Jobs when he came back to Apple. I'm sure someone said the company was in trouble for him doing what he did at the time.

Maybe Intel's biggest blunder was not realizing this years ago and taking measures at the time. I've heard that the internal company politics have been awful for a long time, and when AMD was so weak it's not surprising that it happened. Maybe they've cleaned house of that as a part of this, or perhaps not.

Laying off employees that aren't helping a company in some way or being put to good use to generate a profit is no different than removing a tumor. The notion that if the company isn't growing and getting bigger it's in trouble seems as foolish as saying that a rapidly growing cancer and the increasing number of cells in a human body is surely a sign of good health.

Since we're using this metaphor now where is our patient in trouble? Is this stage 1 cancer or are we past that point already? Have we gone as far as stage 3 or has it not progressed to that point yet?

 

[lopri]

Many workers at Intel are highly specialized. Sure they could get a job flipping burgers if they absolutely had to, but most people will want to continue working in their area of expertise.

If the market for their specific job has cratered, it might even require going back to school if they want another skilled labor position. Many people would prefer a pay cut when the alternatives could be a lot worse.

From morale perspective, small wage-cut (especially at the upper level) is way better than layoffs. You lay off a bunch of people, then even the remaining folks do not feel good/secure about their places.

 

[Joe NYC]

I think their biggest failing with server is not getting E-cores into those products. They've shown that they can actually now compete on desktop with AMD in the heavily multithreaded workloads where they previously got beat so badly that Zen was an obvious choice.

Someone on another forum mentioned that since E-Core is 1/4 P-Core and Zen 4 core is 1/2 of P-Core but has 2 threads, therefore 1 thread of Zen 4 core = 1 E-Core. in die area.

It is comparing TSMC N5 vs. "Intel 7", but that's where things stand now.

We will see how E-Cores in Sierra Forrest on Intel 3 compare with Zen 5 or Zen 5d on either N4 or N3E in the future.

 

[Markfw]

Someone on another forum mentioned that since E-Core is 1/4 P-Core and Zen 4 core is 1/2 of P-Core but has 2 threads, therefore 1 thread of Zen 4 core = 1 E-Core. in die area.

It is comparing TSMC N5 vs. "Intel 7", but that's where things stand now.

We will see how E-Cores in Sierra Forrest on Intel 3 compare with Zen 5 or Zen 5d on either N4 or N3E in the future.

Well, I TOTALLY disagree with that. 2.4 E-cores = 1 Pcore = 1 Zen 4 core. Yes, the P and Zen has 2 threads, but the above is wrong IMO

 

[Joe NYC]

Well, I TOTALLY disagree with that. 2.4 E-cores = 1 Pcore = 1 Zen 4 core. Yes, the P and Zen has 2 threads, but the above is wrong IMO

That was in terms of die area. That is what Intel tells us, that die area of 1 P-Core = 4 E-cores. And from AMD presentation, 1 Zen 4 core = 1/2 of Intel P-Core die size.

From which, 1 Zen 4 core = die area of 2 E-cores, but Zen 4 has 2 threads. So 1 Zen4 thread = 1 E-core die size.

 

[Markfw]

That was in terms of die area. That is what Intel tells us, that die area of 1 P-Core = 4 E-cores. And from AMD presentation, 1 Zen 4 core = 1/2 of Intel P-Core die size.

From which, 1 Zen 4 core = die area of 2 E-cores, but Zen 4 has 2 threads. So 1 Zen4 thread = 1 E-core die size.

Well, at least that makes sense.

 

[moinmoin]

Perhaps the reason that I really dislike the term is that it can mean just about anything and isn't well defined.

Intel has been in a difficult environment since around 2017-18 when it was clear that competition was heating up again and node progress was stuck at 14nm. Intel is in troubles in the last two quarters where it became clear that the company is ill-prepared for an economic downturn that was essentially delayed due to a previous economic upturn for the PC industry caused by the pandemic, and this external pressure causes Intel to drop or delay projects that have been started just within the previous year under new leadership.

Regarding your imo ill-suited cancer metaphor, after these two quarters it's now clear Intel does have cancer (an external pressure Intel has to react to) and due to the apparent lack of preparation it's not yet clear how malignant that cancer is.

 

[Tuna-Fish]

How are they not doing fine?! Even with the negative cash-flow they have that much retained earnings that they could pay off all of their debt today and still have more than 30 bil in their pockets.

Retained earnings is the wrong metric to look at.

Intel only has $28B of cash, and $42B of debt. And also $28B of other current liabilities they cannot ignore (accounts payable, pensions, etc). They are absolutely not in a position where they can instantly pay off all their debts. In fact, they are now fundamentally living on credit, as they have to constantly take on debt to pay their bills and the investments they are doing.

Yes, they have tons of retained earnings in assets, but those assets are their fabs, and other massive capital investments. They cannot just turn those into cash at will.

Intel is not doomed, but these are not the numbers of a healthy company. I fully expect that they can turn this ship around, but large ships take a long time to turn.

 

[scineram]

According to Dylan the cuts only cover Half of the dividend. I wonder how @TheELF justifies all that. Is keeping that a good move while cash flow negative for quite a while?

 

[Kocicak]

Someone on another forum mentioned that since E-Core is 1/4 P-Core and Zen 4 core is 1/2 of P-Core but has 2 threads, therefore 1 thread of Zen 4 core = 1 E-Core. in die area.

WHAT?! "1 thread" is totally meaningless, it is like you said one engine cylinder. You cannot compare anything about car performance using "one cylinder", because cylinders can have different sizes and and also other parameters in engine determine their performance!

 

[DrMrLordX]

They do need to rework their P core as well

If Golden/Raptor Cove were truly in a healthy and market-competitive state, they wouldn't have even given Gracemont a thought in their flagship desktop products. Take a look at their enterprise offerings to see what is really going on: Intel's processes and designs are falling behind their competition, simultaneously.

 

[Kocicak]

They do need to rework their P core as well, but mostly in a way to reduce the size of it. That's really what's killing them in server.

Losing say 10% of performance for 40% area reduction makes a lot of sense for server products.

Losing 10% in desktop CPUs makes them to lose the only area where they are at the moment better than AMD - single thread performance. They cannot afford to lose even 1 % of it.

High single thread performance made me to try Alder lake when I was having problems with my 5700G system. Now I am happy with 13600K. AMD cannot touch its six fat cores running all at 5500 MHz.

The fatter P core in a desktop CPUs the better. At least I do not need many of them, and many other "normal people" too.

 

[Mopetar]

Losing say 10% of performance for 40% area reduction makes a lot of sense for server products.

Losing 10% in desktop CPUs makes them to lose the only area where they are at the moment better than AMD - single thread performance. They cannot afford to lose even 1 % of it.

They essentially already have lost that battle because the v-cache Zen parts largely overcome that advantage where people actually care about the performance it brings.

Unless they're going to try to find a way to release a 450W CPU to try to squeeze out another 3% performance, they need to rethink what they're doing.

Even being the top gaming CPU may not be all that valuable outside of the bragging rights and any mindshare it gets. Almost anyone who's running a 13900KS could switch to a 13600K and not see a noticeable difference in performance. So is it really worth trying to cling on to something that doesn't really matter in sacrifice of everything else?

Regarding your imo ill-suited cancer metaphor, after these two quarters it's now clear Intel does have cancer (an external pressure Intel has to react to) and due to the apparent lack of preparation it's not yet clear how malignant that cancer is.

Your post shows why it works though. If someone tells you they have cancer, there's a pretty stark difference between stage 1 and stage 4. This fact is why we have different stages of cancer and why a word like "trouble" doesn't convey meaning without a thorough explanation, at which point the word becomes unnecessary. Same idea we see elsewhere (e.g. DEFCON levels, EF scale for tornadoes, etc.) when there's a wide spectrum with vastly different implications at either end.

 

[Kocicak]

They essentially already have lost that battle because the v-cache Zen parts largely overcome that advantage where people actually care about the performance it brings.

Unless they're going to try to find a way to release a 450W CPU to try to squeeze out another 3% performance, they need to rethink what they're doing.

Even being the top gaming CPU may not be all that valuable outside of the bragging rights and any mindshare it gets. Almost anyone who's running a 13900KS could switch to a 13600K and not see a noticeable difference in performance. So is it really worth trying to cling on to something that doesn't really matter in sacrifice of everything else?

I do not think that Intel desktop CPUs are doomed after 3D Ryzens come, even if they lose by few percent, they may be better in performance per price, and the Raptor refresh may retake the crown again.

The point is that Intel desktop/laptop CPUs are still competitive, in these markets Intel should have 50% of the market, and in server CPUs 5. Or 1%. They should be very happy if their real market share is larger than that. Honestly I have no idea why they bother with so many SKUs of server CPUs, when they are so bad compared to AMD products. Do they do it as an image thing?

I see a lot of posing and pretending in what Intel communicates and does, I believe everybody would welcome more honesty.

 

[Doug S]

Retained earnings is the wrong metric to look at.

Intel only has $28B of cash, and $42B of debt. And also $28B of other current liabilities they cannot ignore (accounts payable, pensions, etc). They are absolutely not in a position where they can instantly pay off all their debts. In fact, they are now fundamentally living on credit, as they have to constantly take on debt to pay their bills and the investments they are doing.

Yes, they have tons of retained earnings in assets, but those assets are their fabs, and other massive capital investments. They cannot just turn those into cash at will.

Intel is not doomed, but these are not the numbers of a healthy company. I fully expect that they can turn this ship around, but large ships take a long time to turn.

Most companies prefer to operate that way. Look at Apple, who generates more free cash flow than anyone - they have set a goal to be cash neutral and are slowly but surely making that happen. When Apple reaches that position they might have $100 billion in cash and $100 billion in debt (or more if you look only at current liabilities and not current assets) but that won't mean they are "fundamentally living on credit".

If Intel has negative cash flow for a sustained period THEN they will be living on credit, and might have to make some hard decisions about their dividend.

But I do agree that looking at retained earnings makes no sense as a way to judge the health of a business. Some businesses by their nature carry a lot of retained earnings on the books, others very little, but they can be equally healthy or equally unhealthy regardless.

 

[Joe NYC]

Regarding your imo ill-suited cancer metaphor, after these two quarters it's now clear Intel does have cancer (an external pressure Intel has to react to) and due to the apparent lack of preparation it's not yet clear how malignant that cancer is.

Less than a year ago, Intel was saying it will spare no expense to regain leadership. Intel was on a hiring spree, poaching employees from other companies.

I wonder about the last people to join Intel, who left their original companies for higher pay, I wonder how they feel like after Intel took away what they came to Intel for (few extra bucks).

If they still have a job, because Intel has been on a stealth firing spree.

 

[maddie]

Less than a year ago, Intel was saying it will spare no expense to regain leadership. Intel was on a hiring spree, poaching employees from other companies.

I wonder about the last people to join Intel, who left their original companies for higher pay, I wonder how they feel like after Intel took away what they came to Intel for (few extra bucks).

If they still have a job, because Intel has been on a stealth firing spree.

Intel is the inverse of "speak softly and carry a big stick". Next few quarters/years will be interesting.

 

[moinmoin]

Your post shows why it works though. If someone tells you they have cancer, there's a pretty stark difference between stage 1 and stage 4. This fact is why we have different stages of cancer and why a word like "trouble" doesn't convey meaning without a thorough explanation, at which point the word becomes unnecessary. Same idea we see elsewhere (e.g. DEFCON levels, EF scale for tornadoes, etc.) when there's a wide spectrum with vastly different implications at either end.

That's not really applicable to Intel as a whole though, rather to different departments and projects at Intel. The Risc-V Pathfinder program is dead for instance etc. You are essentially forcing to play Nostradamus with future oriented speculation like these. I guess we could rather liken Intel's "troubles" to a(n equally stupid) boxing metaphor: In the last two rounds Intel got close to losing by KO, let's see if it gets better or worse in the next round with the changed strategy. The insistence to pretend to the audience that everything's fine and still throwing extensive confetti (dividends) is no good look.

 

[IntelUser2000]

The insistence to pretend to the audience that everything's fine and still throwing extensive confetti (dividends) is no good look.

I know. There's an argument that maybe they should go private.

So the point of publicly traded companies is to hold them accountable right? But the majority can be retarded. So it can go either way. The actual benefit of a publicly traded company is that they can possibly get more funding/exposure.

I held their stock for few months way, way back. I got an email telling me I can approve/disapprove various high level leadership. That's the reality for these companies. Like someone pointed before, Gelsinger isn't in full control, the Board of Directors have lot of that.

Also the shareholders do have control. Unfortunately the majority can be very stupid and short sighted. Also when it comes to investment it's said "majority must be wrong", and you need to be able to go against that tide to win, which will be massive.

Many, many investors put money in stocks to make money. Short term, Otellini is the best CEO, and Kraznich next. But we know the latter was a disaster and Otellini affected the company long term, despite doing ok everywhere else.

So I can see them dropping it by 25%, and that'll "save" them $1.5 billion annually. But cutting it entirely is unreasonable. Unless your hidden argument is that current management should be kicked out. Then I have to question your motives. You are saying they shouldn't go into new fields, don't invest in fabs and cut down on R&D. Then I think you are advocating for the demise and the argument you make is invalid, so is everyone else in that line.

I am not saying Pat is the guaranteed way of success, because past performance isn't necessarily an indicator of future performance. However, Intel's problems aren't just 5 years old. It's likely 25+ years old, starting from when the last Founders quit being CEO.

You aren't looking at them needing to change a flat tire. Or even do a massive tuneup. You are looking problems at a structural level. Engine, Transmission, Body. Frankly, even the Engine isn't necessary(of course it's VERY nice). You can replace it with a pedal, or foot like Flintstones. But Gears, Body, they are absolutely necessary. And that's the problem Intel has.

Again, the height of their profits and margins were during the Kraznich era.

Also if the economy continues to worsen and inflation goes higher and higher, I think the problems we see may be just tip of the iceberg. Then there's literally no one in the world that can fix it.