News Intel 1Q23 Earnings

[moinmoin]

Meteor Lake on Intel 4's ramp in production wafer starts today

What a very convenient timing. Nobody can ask when it finally starts because it did so today. And nobody can ask how it's faring because it only did so today.

 

[DooKey]

That recession camel's nose is sticking further and further under the tent.

 

[eek2121]

By market share they mean processors sold in the quarter. I'm sure they have ways of finding that out.

Correct. I also speculate they may have an agreement with Microsoft. I'm sure Microsoft will gladly sell Intel your technical specifications.

 

[Hitman928]

Read through the transcript. Gelsinger is a good salesman but I do not like salesmen. If you read the call transcript, he paints the picture that everything is just peachy at Intel, it was a good quarter and everything is going perfect at Intel. I would be much more inclined to believe their process progress claims if they were more upfront/real about their financial performance and position.

We delivered solid first-quarter results on both the top- and bottom-line

This was Intel's worst performing quarter in history and it's not even close. They lost $2.8B on a GAAP basis. The next worst single quarter loss on record was 30 years ago when they reported a loss of $687M. They now have over $50B in debt and are projecting to lose a huge amount of money again next quarter. They are also at historically low levels of gross margin. They are projecting a "modest" recovery in the second half of the year but have given no guidance as to when they expect to be profitable again. Intel is burning through cash like crazy and banking on a recovery (and gov't funds) starting in the second half of the year and building into 2024 and 2025 when they expect to take technological leadership again. If a recession is looming, they are digging a giant hole that the economy might not let them climb out of without drastic measures.

As I mentioned previously, only Mobileye is reasonable profitable for Intel at the moment and even that unit projected trouble in the coming quarters. Every other unit is down, except CCG, which turned a tiny profit (9% margin). Even the data center unit, which used to be their crown jewel, was down 50% Y/Y and actually lost money for the first time in who knows how long. Yet, when asked a question about the DC unit, Gelsinger responded (in part) with this (emphasis mine):

As I highlighted, in my prepared remarks, this was a good quarter for our data center business, a very healthy road map. We did better than we forecast in Q1 market share on track for the Sapphire Rapids ramp. And overall, being able to see these new use cases for AI inferencing in the platform, being able to deploy the increasing security capabilities that we disclosed, all of these are very unique feature capabilities that are both differentiating as well as value add for our customers that helps them to see the value in the underlying platforms that we're delivering. So a very good quarter. Thanks for the question, Tim.

I get trying to put your company in a good light and for Intel that means pointing to the vision you are building for the future, but this is some extreme level gaslighting going on here.

I did catch this fun little exchange too where an analyst basically called out Pat's puffed up remarks in his opening statement in regards to their Gaudi line of products where Pat was claiming that Gaudi2 was able to outperform Nvidia's A100 and that they received endorsements from big players in the support side of the industry which allowed them to address the accelerating AI market.

Pat, you mentioned Gaudi like a very, very positive benchmark on large language models. When I look around me, I don't see, like, good tangible signs of Gaudi, like really gaining traction and getting big, despite the fact that the work today is really starting for more processing power and more capacity to run these models? So my question was, am I just not seeing something that will become apparent very, very soon? Or are there still building blocks and parts and things that Gaudi is missing before really taking this very fast-growing opportunity?

Yes. I think it's a fair commentary that we're only starting to see good positive proof points in the industry. So I think that's a fair critique, Pierre. But I'd point back to the announcement of the Hugging Face, which is the most popular sort of like the GitHub of the AI world. Very positive proof point this quarter, stable diffusion, right, another in Stability AI, important forces also my comments around a rapidly growing pipeline. Obviously, you can't measure that, but I'll tell you, we have many opportunities that we're now engaging in globally. . . So overall, we feel like we're now starting to show up in this space, but we have a lot of work to do to land meaningful revenue customers in this area.

So a lot of fluffy salesmanship, but no meaningful revenue.

I was actually surprised that the analysts didn't address the bad financials and growing debt more in their questioning but it seemed like they were more interested in trying to nail down Intel on their roadmap/future plans.

I don't mean to just trash on Intel here as there are major macro level things happening outside of their control and many companies are (or are going to be) feeling a lot of pain right now but I was really put off by the almost complete lack of acknowledgement of the current state of things by Intel's leadership and, like I said, it makes me trust their comments on the health of their process and product R&D less. AMD reports on Tuesday and I expect it will be a pretty painful quarter for them as well. I hope they are a bit more grounded in their commentary though.

 

[jpiniero]

Interestingly, Intel blamed the lower ASP on notebooks on OEMs buying more Atoms as well as previous gen. By previous gen I assume they mean Tiger Lake.

 

[Hitman928]

Forgot to mention that Intel is now "paying" fab under utilization penalties with their separated out IFS financial reporting. This is somewhat similar to AMD's wafer purchase agreement of yesteryear with Global Foundries where they had to pay a fee if they ordered less than a certain amount of wafers. Of course, for Intel, it's just an accounting thing, but the point is that because Intel still has foundries in house, they do have to pay for the fabs being operated, whether they are actually running product wafers or not, whereas AMD and other fabless companies do not. Additionally, Intel has stated that they plan on changing how they handle old fab lines by keeping a certain amount running rather than converting almost all of them to the next node (more like TSMC's model). This is what you do when you want other people to use your foundry so they can have access to multiple nodes, but it is also a huge money pit if you can't actually get enough customers to fill those lines. Hopefully Intel can attract the customers it needs.

 

[Mopetar]

A lot of the automotive companies would love to be able to stick with an older node longer. They don't want to have to deal with redesign costs, validating the new chips, and all of the other costs when they just want an $17 part they can continue using.

No one is going to get filthy rich providing that, but if Intel is willing to guarantee that they'll keep a fan online for 20+ years to give that to the companies who want it, they can still earn money on something that should have already paid off the investment cost for them.

There are probably enough cities or states that would give them a lot of tax write-offs for opening a new fab so there's not much downside to keeping an old plant running.

 

[coercitiv]

I don't mean to just trash on Intel here as there are major macro level things happening outside of their control and many companies are (or are going to be) feeling a lot of pain right now but I was really put off by the almost complete lack of acknowledgement of the current state of things by Intel's leadership

Intel's been doing this for years now, when they finally admit something it's really bad... and they only admit it for one second before revealing an exciting come-back plan. We need to learn their patterns if we want a better picture, it's a bit like this movie scene.

I was actually surprised that the analysts didn't address the bad financials and growing debt more in their questioning but it seemed like they were more interested in trying to nail down Intel on their roadmap/future plans.

I think the analysts are more interested in the things they can't gauge by themselves. The financial issues... they can decipher on their own.

 

[aigomorla]

Intel has no idea how to launch anything anymore from what im seeing.
There promise of HEDT, i see the parts, i don't see the supply.

I am assuming the same with the enterprise sector.
I bet there is the parts, but OEM's are probably hording the supply and not letting out.

I don't think i can say anything anymore about the progress on intel without getting political about it, which is not in this section.
But i think all big companies we once knew, have the complete wrong leaders helming them.

Well except a few companies, whose CEO's have flipped them around... wont say the name, but you know which companies i am talking about.

And again, I expect a very bad storm for Tech for at least the next few years.
Until we at least get sourcing and supply from China straightened out.

 

[Hitman928]

I think the analysts are more interested in the things they can't gauge by themselves. The financial issues... they can decipher on their own.

I agree, I guess I was just expecting them to try and dig a little deeper into things that you can't really know just from the filings. One analyst did a little bit by asking why Intel's volume ramping is seemingly having a larger negative effect on their financials than in recent times. Intel gave a very vague answer so maybe someone else could have caught on to that and asked it from another angle. Or simply asking now that Intel is guiding for a major loss in Q2 as well, if they have any visibility as to when they will return to being profitable. It's not that important, just something that surprised me a little when Intel just reported by far their worst loss in company history. Usually analysts will rake a company over the coals for an historically bad result and terrible immediate guidance. Could be they just chalked it mostly up to the macro environment and moved on, who knows.

 

[Vattila]

Usually analysts will rake them a company over the coals for an historically bad result and terrible immediate guidance. Could be they just chalked it mostly up to the macro environment and moved on, who knows.

The short answer to your last part is that they all expected worse. And things are expected to improve from here. In particular, the stuffed channels are clearing up so it is expected that Intel will soon again match the sell-through, which should be healthy sales. Despite the big correction after the pandemic demand, they still expect the PC market to be around 270 million units for the full year, which is not bad at all. Similarly, they expect data centre to improve, with demand returning in China, in particular. Second half of the year is also traditionally cyclically up.

The bigger story is that Gelsinger is executing relentlessly on his plan to make Intel a player in the foundry space, while seemingly improving product roadmap execution. In particular, Ann Kelleher seems to have executed the new process roadmap well, which by all accounts is now on track. So there is optimism that Intel will make a major comeback in process technology. They may indeed achieve their goal of having density and performance per watt leadership with their RibbonFET transistor and backside power delivery technology by 2025. It is going to be exciting to see how it turns out.

What is lacking though on these earnings calls is probing on the competitive position on current and upcoming products. For example, when they claim market share gains, I presume they mean in units, not revenue. It would be nice if they clarified. AMD has all but abandoned the low end in the PC space, so Intel is serving that end of the market now. Similarly in the data centre, I presume EPYC is commanding a better price than Xeon at the moment. I suspect AMD will continue to erode Intel's revenue share in both the PC and data centre segments. Then there is the GPU and AI offerings. As you pointed out, an analyst had a probing question about the latter (the Gaudi chip), which reminded me about similar questions fielded to AMD's Lisa Su about the progress of their Instinct server GPU, which she has answered in a similar fashion ("not yet material", "early days", "a journey", as she likes to put it). Likewise, I think Gelsinger answered that question nicely and honestly.

On your first point, why analysts on these calls do not field tougher and more critical questions, the answer is that they may be frozen out of future calls! For instance, Bernstein's Stacy Rasgon has not been allowed to ask a question at Intel's earnings calls for a couple of quarters now. Here he is with his analysis of the earnings release:

 

[Hitman928]

On your first point, why analysts on these calls do not field tougher and more critical questions, the answer is that they may be frozen out of future calls! For instance, Bernstein's Stacy Rasgon has not been allowed to ask a question at Intel's earnings calls for a couple of quarters now. Here he is with his analysis of the earnings release:

I had this thought but seeing you post this makes me think this is probably the long and short of it. In the past, I've seen the analysts kind of team up and have each others' back to drill down in the earnings call, but Intel probably is a much more intimidating company to try that. Thanks for the video too, it's a pretty good analysis for being a quick back and forth.

 

[A///]

By market share they mean processors sold in the quarter. I'm sure they have ways of finding that out.

Correct but what I said still stands. Unless all OS are reading a serial number off the cpu and sending it back to intel who's keeping a log of cpu serials online then their rationale has little to no basis.

 

[A///]

From Intel's earnings release:

cheers

to what end is this? are these order downpayments from partners or is ifs in full swing now?

 

[Hitman928]

cheers

to what end is this? are these order downpayments from partners or is ifs in full swing now?

IFS is in full swing, they just don't have much of interest to offer to customers right now.

 

[A///]

IFS is in full swing, they just don't have much of interest to offer to customers right now.

who is the lucky or unlucky customer having a go at it? qualcomm?

 

[Hitman928]

who is the lucky or unlucky customer having a go at it? qualcomm?

I know they have some government or government funded customers because they are a domestic foundry. Beyond that, no idea.

 

[Exist50]

Interestingly, Intel blamed the lower ASP on notebooks on OEMs buying more Atoms as well as previous gen. By previous gen I assume they mean Tiger Lake.

In the Atom thread, a lot of people have been asking why you'd get a new ADL-N laptop when slightly older Tiger Lake ones are around the same price. Well the reality is that's not a steady state reflection of their pricing. It's Intel and OEMs slashing prices in an attempt to burn through a mountain of inventory, and we see it in their financials here. Probably in a quarter or two, those cheap last gen laptops will have dried up.

 

[A///]

I know they have some government or government funded customers because they are a domestic foundry. Beyond that, no idea.

Oh right that must be the ramp-c funding.

 

[jpiniero]

Probably in a quarter or two, those cheap last gen laptops will have dried up.

Presumably Intel records the revenue when they ship the processors to the OEM, not when the laptop sells. They are buying new Tiger Lake. Call it Shrinkflation if you like.

 

[A///]

That recession camel's nose is sticking further and further under the tent.

you must be @biostud's twin

 

[A///]

Presumably Intel records the revenue when they ship the processors to the OEM, not when the laptop sells. They are buying new Tiger Lake. Call it Shrinkflation if you like.

that would be the obvious takeaway. Intel has some oversight about the final form to my knowledge or had in the past. I'm not 100% sure how taxes work for mega corps like intel but they could give away the chips to oems for free and write it off come tax time. they take a loss on paper but also write off the loss and lower their tax burden. i may be wrong ofc. that way they could write the loss of at the full cost per unit in a group of 10k or whatever it is vs selling at a reduced rate per unit of 10k.

 

[eek2121]

The last time I saw this much pessimism regarding Intel, they dropped a product that nearly killed AMD.

Correct but what I said still stands. Unless all OS are reading a serial number off the cpu and sending it back to intel who's keeping a log of cpu serials online then their rationale has little to no basis.

Microsoft does not need serial numbers since they have unique device identifiers that are far more accurate, and they would be the likely part to sell that info.

Note that even if you pirate Windows, you may still be giving Microsoft enough data to fingerprint you.

The only way you might get around this is to avoid Microsoft products, however, if you use a product such as Spotify, your data will again be roped up…

If you are in the EU and think the GDPR will help you, no, it will not.

 

[A///]

The last time I saw this much pessimism regarding Intel, they dropped a product that nearly killed AMD.



Microsoft does not need serial numbers since they have unique device identifiers that are far more accurate, and they would be the likely part to sell that info.

Note that even if you pirate Windows, you may still be giving Microsoft enough data to fingerprint you.

The only way you might get around this is to avoid Microsoft products, however, if you use a product such as Spotify, your data will again be roped up…

If you are in the EU and think the GDPR will help you, no, it will not.

there is still a finite amount of processors sold that are being added to the pool. a lot are being taken off line. that is the crux of what I'm trying to get across. there is only so much you can add to a pool when the global numbers wouldn't add up for continuous marketshare gain.

gdpr and any privacy by the eu is a joke. companies will find a way to squirrel away that data regardless of whatever a user says. the end fine is a tiny amount vs the profit made from selling user data. legal and gray area data mining is highly profitable.

 

[maddie]

The short answer to your last part is that they all expected worse. And things are expected to improve from here. In particular, the stuffed channels are clearing up so it is expected that Intel will soon again match the sell-through, which should be healthy sales. Despite the big correction after the pandemic demand, they still expect the PC market to be around 270 million units for the full year, which is not bad at all. Similarly, they expect data centre to improve, with demand returning in China, in particular. Second half of the year is also traditionally cyclically up.

The bigger story is that Gelsinger is executing relentlessly on his plan to make Intel a player in the foundry space, while seemingly improving product roadmap execution. In particular, Ann Kelleher seems to have executed the new process roadmap well, which by all accounts is now on track. So there is optimism that Intel will make a major comeback in process technology. They may indeed achieve their goal of having density and performance per watt leadership with their RibbonFET transistor and backside power delivery technology by 2025. It is going to be exciting to see how it turns out.

What is lacking though on these earnings calls is probing on the competitive position on current and upcoming products. For example, when they claim market share gains, I presume they mean in units, not revenue. It would be nice if they clarified. AMD has all but abandoned the low end in the PC space, so Intel is serving that end of the market now. Similarly in the data centre, I presume EPYC is commanding a better price than Xeon at the moment. I suspect AMD will continue to erode Intel's revenue share in both the PC and data centre segments. Then there is the GPU and AI offerings. As you pointed out, an analyst had a probing question about the latter (the Gaudi chip), which reminded me about similar questions fielded to AMD's Lisa Su about the progress of their Instinct server GPU, which she has answered in a similar fashion ("not yet material", "early days", "a journey", as she likes to put it). Likewise, I think Gelsinger answered that question nicely and honestly.

On your first point, why analysts on these calls do not field tougher and more critical questions, the answer is that they may be frozen out of future calls! For instance, Bernstein's Stacy Rasgon has not been allowed to ask a question at Intel's earnings calls for a couple of quarters now. Here he is with his analysis of the earnings release:

I did not watch before writing this but am proceeding on your post.

Anyone mentioning Geopolitics? Did they, at all acknowledge, the wider "issues" unfolding? China? Business as usual? I understand most experts are extremely narrow in focus, but wow. The world is splitting and things aren't going back to "normal".

There was a member here yrs ago, who, in exasperation, mentioned the changes coming. I got the impression that he worked in the intelligence community, so couldn't speak freely and was trying over many posts to hint. He, in some PMs, claimed that the ruling/elite class had all decided that China was the enemy and would be "challenged", and if I wanted, these commodities would outperform. This was pre-pandemic and everything has, and is happening as stated.

The good old times isn't happening, and I'm not depending on that interchange alone. This year will reveal a lot more. This is me trying to be diplomatic.