Subscribe
Share
Share
Embed
A podcast about the business and market of semiconductors
Ben Bajarin (00:00.967)
Hello everyone, welcome to another episode of The Circuit. I am Ben Beharon.
Jay Goldberg (00:08.46)
I'm Jay Goldberg.
Ben Bajarin (00:10.375)
Jay believes that we have been hacked at this very moment by authorities in the state of China because they want the circuit episode so badly that they need to listen to it the second that it's recording, even though I could theoretically record and publish this live in real time for you all if you wanted to. But that's the background to his actually impeccable Mandarin. But we think we've been hacked. We're just kidding.
Jay Goldberg (00:39.481)
not so impeccable and I didn't name any country in particular, I'm just saying, perhaps a state-sponsored actor, maybe listening in.
Ben Bajarin (00:47.707)
They are, they are, yes, they, there was a little weird static when we started recording. It kind of sounded like when you were recording and then in analog days, you might've had somebody tap into your call. That's why we joked about it.
Jay Goldberg (00:50.158)
I'm.
Jay Goldberg (00:59.707)
I'm wearing the wrong hat. I should be wearing my tin foil hat instead today.
Ben Bajarin (01:02.163)
Yeah, to surround ourselves. All right, so this episode is meaty and there will be disagreement, I think, because we're gonna talk about Intel. And as everyone knows, I'm generally a little bit more positive. Jay is the skeptic, although not nearly as a skeptic as some of the financial people that we know who are generally very, very hard on Intel. But it was a really weird...
earnings call and in a really weird quarter because Intel actually did a little bit better than expected with a number of business groups, but guided lower than expected, which is really, I think the source of the sell off, which we'll talk about. But we just briefly want to highlight, and I'm going to let Jay do this, briefly highlight a couple of the quarterly earnings call outs.
Um, from last, from Q4 23.
Jay Goldberg (02:06.018)
Deep sigh. OK, Intel. So Intel, today is January 26. We're recording it. Intel reported yesterday the 25th. And their results were, I would categorize as not great. The Q4 numbers were good. They're above expectations, $15.4 billion against expectations of $15.1 billion. And EPS profitability meaningfully ahead of the street, I think, it beat by $0.10 or something. Great. Good job.
coming off of last quarter, which was demonstrably like a turning point, which could be seen as a turning point in the stock. Like everything was firing again, they were profitable, process improvements, they were announcing all that good stuff. So it was reasonable to assume that, all right, there's been this long journey, they've turned the corner and are improving. But then yesterday they guided for Q1 and that guide was abysmal. It was revenue, they're sort of forecasting one and a half to $2 billion below where the street was expecting them.
The street was expecting a 30 cent EPS. They're guiding to a 20 cent loss. So big, big drop relative to expectations in both revenue and real concerns about deteriorating margins. So not a good. I mean, I was going to say, that's a terrible guide. I mean, it was one of the worst guides I've seen in a while. And I think investors in general are kind of a little bit spooked about what's going on with semis. We had a really good six months.
everything ran up. And I think some people are worried that we sort of were hitting the top, like it's kind of as good as it can get. And Intel fed into that fear, but also has some very specific issues going around it. Relative to that narrative that they were improving, last quarter they turned a corner and this quarter they seem to have turned back.
And so we're going to debate what to make of all this. I think, where is this headed?
Ben Bajarin (04:07.695)
Yeah. So I think there's, there's a couple things that were kind of weird, but made some sense because, um, they, they came at this from a posture that this off-quarter is explainable in terms of some of the, um, business units underperforming that are not a part of their core businesses. And they kept using this core businesses, which is CCG and DCAI. So consumer, the client group and data center.
which right now will focus on are there two core businesses? I think IFS will be added into that, but it isn't at the moment, in my opinion. So those two, CCG and DCA are sort of the kickers. But within that view, I think you look at the last, I want to say six months, right? And Intel has actually been doing okay as a stock. They were up maybe 15 points over the course of the last six months. I think investors were starting to get some confidence that...
as the semiconductor industry had some recovery into 2024, that Intel would benefit from that. And I think that was sort of the source of hope, if you will, coming into this quarter, that they're going to see some of this uplift. And it doesn't even have to be, I'm not even talking about AI uplift at this point. I'm just saying that the semiconductor industry looked poised to rebound in core categories in 2024 because of a fairly rough year last year, and that Intel would benefit from that. And then Intel comes out and is like,
we're going to be low. And now I think the concern comes back to, you have to wrestle with that in now the scope of, we still believe that this industry is going to have strong recovery in categories that Intel should theoretically benefit for, which is also why I think they were relatively confident in their posture saying that all the next few quarters will be up as well as year over year revenue will be up. They seem to believe that.
with a conviction, I'm not sure this first quarter number, because of that guy, gives the street confidence that Intel's correct. I think they're a little bit worried that Intel's not correct in that, and I think that played into it. But the part I'm interested at is that suggests to me that even if Intel grows, they're not going to grow, they're going to underperform the rest of the market. So other parts of the industry will grow more.
Ben Bajarin (06:29.787)
which even if they're growing, which I think is good, you're gonna gotta wrestle with that bogey that sure, you got good growth. Is this high single digit? Is it low double digit? But your peers are gonna grow more as this swing turns around. So I think all of that kinda just impacted now the thinking and the reaction was within some of that greater modeling as well as the way I just take it from a couple of conversations I had with friends on the sell side, some shaken confidence.
Jay Goldberg (07:00.962)
Yeah, so I think there's two types of problems that popped up on the call. One are sort of tactical short-term next quarter kind of problems. And then there's the bigger, longer-term questions. The short-term problem is...
Last year was terrible for consumer semis. Smartphones and PCs were just bad. But as we got through the year, the last few months, everyone sort of had a renewed confidence that PCs and mobile are turning up, which should help everybody. We've seen that from all the mobile ecosystem. We've seen that from AMD and some of the other PC players. But here's Intel, who said they guided down, they cited three things, really. It was a.
Jay Goldberg (07:47.718)
FPGAs, so Altera slash PESG, which is automotive exposed. That's kind of their, that's OK, maybe. They said Mobileye, which again, automotive, big heavy exposure there. So those two are dead. But they also sort of implied that consumers, so the client group, PCs, would see greater than normal seasonality. Q1 is always down quarter on quarter.
Ben Bajarin (07:51.283)
PSG.
Ben Bajarin (07:56.017)
Yep.
Ben Bajarin (08:13.192)
Right. They said their specific words were the low end of that seasonality. So they're saying it's already seasonal, we're going to be at like the low end of that.
Jay Goldberg (08:21.043)
Right. And I think most of, certainly most of the analysts on the call who touched on it, and I think most people in general, were kind of scratching our heads, because this seems like, I mean, $2 billion miss is a big number. I mean, PSG and Mobileye don't contribute that much to revenue. So it sort of sounds like it's a very big miss happening on the client, which again, is counter to the narrative. Shouldn't PCs be doing better?
some at least not as bad. And I don't know how to square that. It really struck me as odd. And then I think the first call was Ross Seymour at DB, who asked about that specific question. And I thought Pat's answer was really weird. It was like he sort of got into the details of how we're going to define seasonality. Is it down this much or down that much?
Ben Bajarin (08:52.659)
Well, not this quarter, but yes.
Ben Bajarin (09:15.247)
Yeah.
Jay Goldberg (09:17.014)
Which is, you don't want to argue numbers on a call with a cell-side analyst who's sitting in front of his Excel model. It was weird. So I think there's that set of tactical concerns that for the next quarter just didn't sound good.
Ben Bajarin (09:24.231)
Agreed. But within.
Ben Bajarin (09:31.239)
But I did sense, to be honest with you, I think you hit this exactly on the right, that there became a semantic debate about, well, how are you defining that? Because I don't define that way. And now if I have to read within how Intel is defining this seasonality, cyclicality, relative to some unique definition to Intel, again, that makes me hard, you have to believe that their model's right, or that their definition's right, in order to now.
make sense of A, the quarter and B, their growth projections, their confidence in that growth for, you know, called the next three quarters. But I think the issue is the second half of the year should be when we see, and I don't know how to read second quarter. Second quarter is kind of weird, but obviously they got a lot of product wins from Meteor Lake. I mean, that seems evident. Those PCs don't ship though until, you know, late second quarter, early third quarter, and then big into fourth quarter.
So I do wonder if the back half of the year, the question is when, why was data center was what it was in the get like data center should be seeing a lot of that uptick as people are still doing, we know needing to refresh their systems. So that, that was a little bit more of kind of a headscratcher. I get PCC senility that totally happens. The data center though, I think was a little bit trickier trying to make sense of that within what typically will happen in this quarter, because
I would expect that Nvidia and AMD are going to have fairly good data center quarters this quarter.
Jay Goldberg (11:02.838)
Yeah, so I remember on a couple of points during their, when the execs were speaking at Intel, they said, oh, but the year is gonna be up. Like they were trying to say like this quarter is bad, but the rest of the year is gonna be up. And okay, I'm willing to believe that on client, like there's an improvement out there, because they're gonna have a whole bunch of new products on the market. Maybe they can pick up some share from AMD or, you know, and they haven't really led that much to AMD. So they're still competitive there.
Ben Bajarin (11:12.787)
Mm-hmm. Yes. Yes, right.
Jay Goldberg (11:30.646)
But if that's true, if they really think they're going to grow this year, revenue and EPS are going to grow this year, then for the life of me, I could not understand why they didn't give full year guidance. And I've been thinking about it all day. And I had conversations with people on Twitter about it. And it occurs to me, they didn't guide for the full year because they can't. They just don't feel comfortable with how the year is going to look. And I agree with you 100%. That is.
Ben Bajarin (11:41.298)
Yes.
Jay Goldberg (11:59.842)
largely due to data center, because data center is a big problem. But I also really wonder how much confidence they have in client for the year. But it's the data center piece that I think really leads into the second set of problems they have. The quarter is what it is, but the year outlook, I think, is very challenging. And I think that's the narrative on the street is, the data center belongs to NVIDIA now. How is Intel going to be relevant here? And from the beginning, I think,
You know, I think the sentiment is increasingly that Intel is, you know, not that relevant in the data center anymore. I mean, to put it really in strongest bearish terms.
Ben Bajarin (12:39.203)
Yes. And I think that's true. And I think that became sort of conversation from the point of GPUs. And I think because everybody's sitting there saying, well, we know that we're in a training cycle. Training means people want to buy GPUs, which means it's largely going to be Nvidia with a little bit of AMD, not as much in terms of where Intel is positioned there, which is why they kept making the point, which I agree about. And I don't know when this flips over where the dollar buy for data center and on-prem.
shifts to the demand being for inference, at which point that's a CPU conversation and a dedicated accelerator conversation, which should be both an Intel and AMD conversation. But that feels like what they're banking a lot of this on. And again, I believe that thesis, I don't know when that happens because we're definitely still in a GPU market right now. But you could hear them talk about, you know, inference matters, we're going to be well positioned when people start caring about inference.
both in the data center and on-prem infrastructure. But again, I agree with you. I think guiding full year would have helped this bogey. I think I'm with you that they're not sure. I don't think that means they're not necessarily gonna grow. I think it's just a matter of, again, are we talking about, is this a single digit or a double digit type of a situation? And yeah, they may not know which is why. But I do think that would have helped.
But still what's odd to me is that again, I've thought a lot about this. So I've, you know, I've, I've rationalized it to myself. Like I can, I can believe in these next quarters. What I don't understand is like what happened that other people just didn't understand it. Like it just constantly got, got hit and brought up about like they, they tried until tried their best to, to soften that blow. And it just didn't, it just didn't work out. And
I realize it's a complicated issue because like I said, I did have to think about it for a while to understand the explanation, but it just didn't come across. It didn't land. Now investors look at this and they go, we had hoped for a nice clean recovery and that this year would be good. Now they're like, it sounds like it's going to be more bumpy, which again, shakes my confidence.
Jay Goldberg (14:55.594)
Yeah, I've listened to what a few hundred conference earnings calls in my life. And after a while you get a sense of the, I'm sure you've done this too. There's like a certain pattern. There's patterns you pick up on. And the patterns I was hearing from the answers they were giving to their questions eliminated any confidence I had in their answers, right? They just weren't giving confident answers. They weren't talking to the...
subject at hand. They didn't really, I don't think they fully grasped the extent to which they're still very much a show me, prove it to me story.
Ben Bajarin (15:28.049)
Mm.
Ben Bajarin (15:32.559)
Yes, right, right.
Jay Goldberg (15:34.906)
And their answers were, they wouldn't necessarily answer the question given, or they would sort of go off on these very narrow, specific critiques of the question. And that's not how you address this big issue. And I don't know if they didn't want to talk. I mean, they gave the sense that they didn't want to talk about these issues. I don't think that's necessarily the case. I don't think they fully grasped how much people don't believe them about a lot of this stuff. And I think that's the most concerning part to me, is just sort of.
a big disconnect. And I wrote about it in my, I published a note on it today where I was, I said, it feels like, this feels like the Intel of the past, right? 10 years ago, Intel could ignore a bad quarter because they just say, oh, the new, the new node will be here soon and that's gonna fix all the problems. You know, we're Intel and we're confident. And back then they could carry that, right? They could walk that, they could walk that walk. And today I don't, you know, it's been years since they could do that though. And
The answers they were given very much harken back to those bad, those old days. And it's just, it's a different world. And I think they really mishandled the communications on this call. For some reason, it didn't seem to anticipate the tone of the questions they were gonna get.
Ben Bajarin (16:38.585)
Yeah.
Ben Bajarin (16:49.763)
Yeah. No, agreed. I think like I said, like I said, it was a head scratcher to me because just listening to the whole thing and kind of how it played out and then just a handful of notes that I read afterwards, it was like, it just feels like, I don't know, I feel like this could have been avoidable if, if it was just articulated a bit of a different way, but we are where we are. You're totally right. This Intel needs to proof in the pudding. I think
Hopefully their next three quarters are what they are. I do, you know, it is interesting, right? To see what happens with Mobileye in terms of, like you said, automotive exposed. Is that just a couple quarters or is that better? And then what happens with PSG revenue wise, because as you know, we looked at yesterday, it drew a lot of semi down. The only other, you know, it's now secondary stock on the market that's FPGA specific, which spooked the whole market of SPGAs, right?
Jay Goldberg (17:45.718)
Yeah, it was interesting because AMD was only down a point on the day. AMD was down a lot at the open, but it closed only up down a point. Lattice got hit pretty bad just because, like you said, it's the only other FPGA stock out there. But it does raise an interesting question, where you have Intel is sort of trained in how it speaks to the street. And I wonder how that's going to carry over to Mobileye and ultimately.
Ben Bajarin (17:49.935)
Yeah, Nvidia was down slightly.
Jay Goldberg (18:14.114)
PSG slash Altera when they get spun off. Mobileye was public for a long time, and they're public again now. So I think they can probably speak for themselves, but PSG is like, that's a whole new thing. I hope they don't pick up those habits is what I'm kind of saying. They need to find a new way to communicate.
Ben Bajarin (18:34.007)
Yeah, agreed. And it's tough. Like FPGAs are, as we've talked about, hard to wrap your brain around because they're very blank slate. They can be lots of different things. And so it's kind of hard to really grasp the categories and the productization when, you know, they're, they're a blank slate of programmability. You can make them do lots of different things. They're general purpose. Um, so yeah, that will be an interesting, interesting to watch how, how PSG, uh,
Jay Goldberg (19:02.774)
Yeah, yeah, yeah. You and I have had this conversation. I don't know if we've had it on the podcast before, but like modeling an FPGA company, if you're on the street trying to build a financial model around the FPGA company, it's impossible because it's too many end markets and you can't keep track of them. And it's really hard to get data for them. So it's a modeling challenge just from the get-go. And I think that's part of the reason that Altera and Xilinx were kind of undercover for a long time is just because nobody could build
Ben Bajarin (19:03.591)
gets spun off.
Ben Bajarin (19:31.018)
Hmm.
Jay Goldberg (19:33.078)
financial models that gave them any confidence around it. I think it's a problem for lattice too. I mean, they're doing fine, but it's a tricky problem to solve. And that's just strictly a Wall Street interpretation problem. I think FPGAs are super interesting and they're gonna have their place in all these markets. FPGA market is fine, but it's tough for the street to really dig into those names.
Ben Bajarin (19:56.795)
Right, agreed. So let's shift the conversation now to Foundry because interestingly, even though this wasn't relevant to necessarily Intel's earnings, they did announce a new partnership with UMC. And there's a couple things about this that I thought was interesting. So the first, the broader background and context for this was that there's...
driven by UMC, but there's demand to move to from, I have no better way of saying this other than the trailing edge. So all the strength that we see in 28 nanometer and all these other things that do a good job. There are elements of those manufacturing capabilities that want to modernize. So move from 28 to something else, bring a little bit more of a compute capable node and start building your products on again, a little bit more compute capable node.
That's not just right. 28 nanometer. So, so part of this drive is, is that right? Intel wants to compete in what is now a collaborative. So there's, there's assets being brought by both Intel, by UMC in terms of capital people tooling, um, in order to build this specialty node that's being assisted by not necessarily for UMC, UMC is going to use it. But.
as it pertains their customer base as a whole, who wants these sort of specialty nodes. So A, I find that interesting about, okay, so the trailing edge moves from 28 to something else. So now when we say the trailing edge, we might mean something in the teens where there's interesting things happening. Secondly, it's interesting that to me that Intel is taking a page from TSMC's book finally and using depreciated assets for this. So their cost structure.
should be good when this hits in 2026, which means, and again, US soil, right? So part of the story is, oh, we can do this now. So we have defense. We've got companies that might want more modern, again, modern trailing edge nodes, which is feels like a paradox, but that's how I'm saying it. It, in order to move their products kind of to more compute capable and do so on us soil, which then brings up interesting questions about global foundries in, in light of this in terms of.
Ben Bajarin (22:20.439)
Is that bad for them? Is this something they need to do to bring more compute capable nodes at call it between 10 and I don't know, 24, 22, wherever those things fall. Um, but I thought, I thought this was an interesting signal is kind of my, my point about how the trailing edge will evolve and then who's poised to compete there. That's the point I was trying to get to with that ramp.
Jay Goldberg (22:47.59)
Yeah, so there's a lot of interesting things happening at the Trailing Edge, and interesting good and interesting bad. And I mean, we probably should do an episode on Trailing Edge at some point soon, because there's a lot of...
Ben Bajarin (23:01.919)
Remember, I just diversified the trailing edge to the trailing edge, the modern trailing edge. I've brought two new definitions to the table. Mid edge.
Jay Goldberg (23:06.596)
That's right.
Jay Goldberg (23:11.242)
Mid edge. But no, there's a lot of stuff going on there. I thought the UMC deal was interesting. Since path taken over, he's made it very clear that he wants to increase Intel's trailing edge capacity. He recognizes, he's been very public about this, as TSMC has these big profit centers because they use these fully depreciated fabs. And Intel didn't do that.
So he's wanted to add trailing edge capacity. They had the tower deal, which got blocked. But I think that was clearly the motivation behind that deal. I think he'd buy UMC if he thought he could get it past the regulators, which of course he can't. But right, so he's and what two, three months ago, they signed a very similar deal with tower, similar to the UMC. I think the UMC deal is even broader. I assume that those deals have been both have been in the works for a long time. So I mean, he's.
Ben Bajarin (23:53.669)
Yeah.
Jay Goldberg (24:10.27)
So he's got a lot of, a lot move, he's really putting a lot of his time and his effort into getting Intel's trailing edge capacity built up, which is important, it's a good idea. It was, it's really interesting to me though, because last week we talked about TSMC's earnings, and I think it was one of the last things we talked about was they actually mentioned that they are, I think they're building a new 28 nanometer fab, or they're revamping one.
Ben Bajarin (24:35.52)
Mm. Yeah, because they need the capacity. Yeah, yeah.
Jay Goldberg (24:38.07)
Which is weird, right? It sort of throws the whole story up on its end because the whole point of you keeping those old fabs open is that they're fully depreciated. But now they're putting more capital into them because there's some need there. And I think part of the motivation for what's going on here in the industry broadly is China, right? China has a huge amount of trailing edge capacity.
coming on stream. We'll probably see some fabs in other parts of the world too. Europe probably has some too. Lots of interest in building fabs in every country. A lot of that's gonna be, all of that's gonna be a trailing edge. And China in particular is gonna bring this big wave of capacity. And I think the best way that the global leaders can defend themselves against that is to have specialized something really special about their process. So it's not plain vanilla 28 or plain vanilla 16. It's something.
a little bit more, a little fancier, which, like you said, that leads directly to Global Foundry's front door, because that's their whole value proposition, is that they're all into these specialized trailing edge processes. And so I don't think the GF people are that worried quite yet. I think we haven't seen a lot of detail about what this UMC process is gonna be. Is it gonna be SOI?
Ben Bajarin (25:38.727)
Exactly.
Jay Goldberg (26:05.726)
Are they going to go after the RFSY market? That's a big deal for GF. But I, uh, there's, there's lots of room there for specialization and everybody's could be a little bit different, but it is like, there's a, there's a lot of moving parts and I, I think this is actually a pretty smart strategy. I think this is, like I said at the time, Intel losing tower is a blessing. They didn't need to buy tower. Now they can just sign this deal and you know, it's a little more cumbersome. They don't have control over it. So there's a little bit of margin stacking.
That's all fine. Like now you can go to a customer like the US government and say, we'll build you a whole system with the trailing edge and the leading edge parts. And I think that's really what they want to do. Fantastic. We're going to have all these rad hard military grade semis coming out of a whole host of nodes.
Ben Bajarin (26:43.085)
Yep.
Ben Bajarin (26:53.223)
Yep. Well, and I think you see where this becomes interesting and I haven't fully worked out, you know, kind of how this plays out, but, but similar to a theme we talked about last week about these very specific AI architected and designed pieces of silicon, you know, for things like let's call it in, let's call it a dedicated AI silicon security chip. That doesn't need to be on three nanometer or two nanometer, right? That could be plenty good at 10 or, or something else or 10, right?
those things. And I wonder again, like just how much this is a signaling of a trend to more again of that modernization of some of these mid edge, mid edge boundaries, because it does feel like there's an element of AI that will necessitate that. And you're then you're exactly right when you've got local people. So again, right. China will do this for it does again have to be at the leading edge for Chinese companies. Samsung will do this for Korean companies. There's Japan semiconductor companies who've already made
investments there, Europe, et cetera, for their sort of domestic needs around, again, call it a mid-edge specialty node that's got mix of some AI architectures or something else. Anyway, I like it. I thought it was very clever. And you're definitely right. I think the one thing that Intel commentary wanted to sort of make clear though, in terms of the tower deal versus this one, like they've positioned this tower deal really as just kind of contract manufacturing.
This is a collaborative effort where they're working with UMC on this kind of specialty process to again, go and solve problems for them and their customers that they're hearing. What that was to me, the kind of part of this where I was like, that's actually, that's really interesting as a signal. So.
Jay Goldberg (28:38.186)
Yeah, I didn't mean to imply that the tower and UMC were the same. There's lots of differences in this deal, but it's sort of thematically, they're very tightly coupled.
Ben Bajarin (28:45.411)
Yes. Yeah. And I think it puts Intel in an interesting position, right? To say, look, where might more of these collaborative efforts start to show up with depreciated assets of theirs at, you know, kind of this mid-edge scenario, right? And especially as you move to, again, over the course of the next, I don't know, three or four years, and we move to, you know, 20A, 18A, 16A, 14A, I don't know if those are the names, but I'm just making them up. Well, then Intel 5,
Those things become actual validating, you know, again, mid nodes to go and you depreciated them. So, like you said, Pat wants to monetize these depreciate assets like TSMC. So that becomes some of these opportunities over the next, you know, call it decade that I think are interesting.
Jay Goldberg (29:33.323)
if they make it that long.
Ben Bajarin (29:35.251)
They'll make it that long, Jay. I'm not.
Jay Goldberg (29:39.614)
I'm not going to let you end this on a completely upbeat note.
Ben Bajarin (29:41.111)
I know, I know, I know, I know you're not. They have to survive. Like, I kinda know, I feel like this is a, they have, like, for the sake of the industry, they have to survive.
Jay Goldberg (29:53.866)
Yeah, they'll survive. But in what shape that will take and what the company will actually be producing at the end of it is still very much to be debated.
Ben Bajarin (30:01.615)
Right. But I like, but, but within that point, like, I think what I'm positive about is that there has been a lot of, so yes, the turnaround has not been rosy. And I know we're not necessarily saying it's a full turnaround yet, but we are in the midst of this turnaround. I think things could have gone a lot worse. I think they're actually going better than honestly, I, I thought in terms of execution, but I like the, I like the, the nimbleness of some of these things that Intel's doing.
Like you wouldn't have scripted that right a year ago. So it feels like, it feels like they're very open to lots of different changes that are happening, both at the manufacturing and the product front. Cause again, they're making a number of tiles at TSMC for this upcoming product launch, which super interesting. Um, I don't know. I just, I like that it's, it just doesn't feel like they're stuck in their ways. They might be still beholden to some beliefs, but I just don't feel like this is the same.
we're going to live or die by this thing that prior management would have. And so I like that there's diversity, willingness to think out of the box. Now again, does that mean that they, all things remain rosy in the end? I don't know, but I just feel like that's a better posture than some of the posture I've seen, which is live or die by a very specific philosophy.
Jay Goldberg (31:20.706)
So for years now, I've used this sort of framework of Intel has to solve three problems in serial. One has to be solved before the next. And the first was fix their manufacturing process. And they've probably done that. We don't know for certain, but it's looking pretty good. So that's one problem one. And I think that was the story of 2023, is their journey through getting 20A and 18A to production. And we're almost there.
Ben Bajarin (31:40.219)
Yep.
Jay Goldberg (31:50.878)
Okay, problem number two is they have to now go out and actually make products that people want to buy. And that's what I think that's what we're in right now is at this exact moment. They don't have a lot of products that people want to buy. And we're going to see through this year, how that changes. The new processes will come on that will make their products more attractive. They're right, but we don't know. And, and the whole shift to AI and Nvidia's rise.
makes that story, solving that problem much harder. And then the third problem they face, which we really are just scratching the surface on is, then they have to get IFS to work so that they can fill their fabs and keep, and stay the course and maintain, have enough volume going through their fabs in order to support the R&D to keep on this curve. They gotta keep the treadmill going. And that one, that one's
Ben Bajarin (32:50.479)
Yes, longer. Right, right. Exactly.
Jay Goldberg (32:50.513)
still to be determined, right?
I will absolutely grant your point though that I think Pat is doing as good a job as he can given the cards, the hand he was dealt. I can't think of anyone else who could have done this, maybe a couple people, but it's like, he's the right person. I have faith and you're right. He is willing to break the old ways of thinking and sort of, you know, there are no idols, no sacred cows, so to speak, that he's beholden to and he's...
He's doing a good job. And so that's why I was a little bit concerned by the way they were talking on the call yesterday, because it seemed like reverting the habit. And not necessarily just him. Like I don't wanna single out him. I think it was just their whole tone overall. So yeah, they're not doomed, but I do think what really upset people about this quarter was the fact that we thought we had gotten further along their trajectory than we...
Ben Bajarin (33:25.483)
Sure, sure. Agreed.
Yes.
Ben Bajarin (33:48.675)
Yes, right. I agree.
Jay Goldberg (33:50.282)
seem to have. And so we still have, we're kind of, you got to prove it to us again.
Ben Bajarin (33:55.375)
Yeah, I think that's exactly the point and which we'll end on because that's again, just the summit it all up. We thought we were really through the schloff. They had some really hard quarters last year. Everybody had hoped that was the bottom, right? When margins hit, I forget the bottom was like 32, 34%, something like that, some really bad number. Granted though, gross margins were good and they're predicting gross margins to stay good, which does suggest a strong product backing. But I think you're exactly right.
Jay Goldberg (34:13.6)
Yeah.
Ben Bajarin (34:25.599)
People had hoped that we were through the worst and having a Q1 like that didn't, didn't help that any. And so it's back to, you got to prove it to me again. And these next three quarters will be, will be very telling. So, um, show me the money.
Jay Goldberg (34:32.46)
Yeah.
Jay Goldberg (34:38.11)
Yeah, and there are more problems. Show me them.
Ben Bajarin (34:45.234)
There you go.
Jay Goldberg (34:45.598)
Yeah, I mean, the thing is there are more problems out there. We know there's more bad news coming, right? For one, next quarter I think they're going to break out IFS, or next month they're going to break out IFS' revenue model. And they've been telegraphing pretty clearly, like, don't get your hopes up. We're losing a lot of money. And I think we all know that, but I think once we see just how much money IFS is losing, people are going to flip out. That's one. So you know.
Ben Bajarin (34:56.219)
We think, I don't know, maybe we think, hopefully.
Yeah, it's...
Ben Bajarin (35:11.384)
Yeah.
Jay Goldberg (35:14.738)
If you're listening to this, get ready. Those numbers are going to be deep, deep in the red.
Ben Bajarin (35:18.883)
And we're excited to talk about that when we learn about it after their event on February 21st or 22nd, one of those days.
Jay Goldberg (35:25.546)
Yeah. And the other thing that concerns me is, like, David Dinser, their CFO, mentioned a couple times like, oh, we'll be cashflow break even this year with a little asterisk next to it every time you say, you know, but remember, we still have to spend a lot of money, the transition, and we got to do a lot of capex. So like, maybe they're not going to be cashflow positive this year.
Ben Bajarin (35:43.414)
fixed-cost business.
Ben Bajarin (35:50.967)
Yep. That was my favorite commentary from Dave in the bad quarters, the dark days of the growth, low 30% margins when he just kept saying, this is a high fixed cost business. And it is, you got to spend 30 to $45 billion per fab now. When I first started covering Intel, this was like 8 billion. Now it's 30 to 45 to get competitive at the leading edge. But we hope.
those ships, right? They're going to need some big whales of a customers to help push that forward and get back to full capacity and at some point profitability. But as we talked about before, I don't think we're going to know any of that really until 26 or later. So, but it'll be fun to talk about IFS once we have IFS day at the end of February. So stay tuned for that because that will be interesting. So on that we will wrap. Thanks for listening everybody. As always, your comments are appreciated.
The kind words we get via email, like, subscribe, review, as Jay liked to say, tell your friends and we will talk to you next week.
Jay Goldberg (36:55.862)
Thank you very much, everybody. Thanks for listening.