Subscribe
Share
Share
Embed
A podcast about the business and market of semiconductors
Ben Bajarin (00:01.398)
Hey listeners, welcome to another episode of The Circuit. I am Ben Beharren.
Jay Goldberg (00:07.49)
Greetings programs, I'm Jay Goldberg.
Ben Bajarin (00:10.518)
This is an earnings week. We love the earnings, the stress and chaos that it provides across the board. Earnings in Semis land, we got AMD, Intel earnings amongst hyperscalers who also make their own semis, Microsoft, Amazon, and a few little kind of tidbits in mobile land, which we'll get to at the end. So let's jump off with AMD.
know, when the numbers came through, it looked good. But the market kind of reacted negatively. And I'm curious on your take on this, because we haven't talked about it right since then. But it was really, is this a case of, well, they just didn't guide to the astronomical expectations, another kind of, know, expectations were too high. And they're just not taking the share of the AI pie that that they hope.
or was something else structurally going on in call it their revenue and their gut.
Jay Goldberg (01:14.802)
I think it's the former, right? I think that's 100 % right, which is expectations among some subset of investors was way beyond what was reasonable and they reported a reasonable number and that wasn't good enough. And so the bloom came off. I think the dynamic around AMD is that there are people who believe that AMD is going to be like this credible second source nipping at the heels of Nvidia.
in AI markets. I don't think anybody rationally can look at it and say that. think that's the wrong benchmark to judge AMD by. And we talked about this before, which is they're not going to displace Nvidia. They just have to carve out a little bit of the market and their numbers work. If you expect them to be like Pepsi to Nvidia's Coke, then you're going to be disappointed when they only do whatever few hundred million
billion in my 300 revenue. But if you're more realistic about it and you look at it and you say, hey, this is a company that had zero AI exposure two years ago, and now they have a decent product that some people are going to buy, then they're doing OK.
Ben Bajarin (02:33.068)
Yeah. I mean, I think a lot of the, the, the post commentary I read, right. was just people saying, well, I mean, this is just more evidence that Nvidia is entrenched. There really is no second source, second source. mean, Lisa and management, you know, talked it up as best they can. I don't think they really moved the pipeline of dollars of that up meaningfully, even though it's tracking to sort of their overall initial estimates. But you, but you're right. think that's the, underlying problem is.
I mean, think if there was a boom to their accelerators as a second source to Nvidia, I think you'd see a little bit more of an uplift than you saw. And so that's what I'm guessing most people reacted to.
Jay Goldberg (03:22.05)
Yeah. Yeah. I mean, Nvidia has just grown at such a crazy clip over the last two years. There are people out there who think AMD should be able to do that too. And I don't think management presents, ever presented themselves as that way. They don't think of themselves that way. It's not going to grow like that, but it's like, so if you, if you could take the AI out of the conversation and just look at it as a semiconductor company that's sort of executing well, they're doing fine.
Ben Bajarin (03:48.034)
Yeah. Yeah. what did you take as I had heard some, some, speculation about this going in and it sounded like it was a little bit confirmed that the general purpose server market, is actually seeing some strength, which doesn't totally shock me. Right. You got just regular old cloud native servers that need to be updated. Obviously the back vast majority of CapEx is going to AI servers, but
It makes it sound like they're seeing some strength in just some of their, their standard, you know, non AI, server parts.
Jay Goldberg (04:25.378)
Yeah, I thought that was encouraging. I'm not quite sure how to interpret all what they said, but it makes sense that there is, I mean, just the world's still getting more cloud, more cloudification, the world's getting more cloudy. don't know. Things are still moving towards the cloud. There's that long-term trend and that's still going on. Like people are still doing regular old Google searches and running their databases and that needs.
Ben Bajarin (04:41.848)
Cloudification.
Jay Goldberg (04:54.76)
sort of playing a little CPU. So, and what I think what interested me about that is that they made that case rather than try to make the case that, we're seeing good CPU demand for AMD CPUs as head nodes for AI clusters. Cause that's, we're going to get to that one soon, but that it's an interesting topic. It's like, mean, people still need CPUs.
Ben Bajarin (05:15.372)
Yes, right.
Ben Bajarin (05:20.982)
Yeah. Right. Yeah. And I think there was a kind of tempered optimism. You know, nobody's pushing this as like, it's a giant refresh cycle, just that they're refreshing some, some, some general purpose servers and we're seeing some strength for that. So I think that's good. Kind of a highlight that is kind of not sexy, but good, good for that business. And they just launched a bunch of new parts, right at their, at their AI day. So
It seems like they're in a good spot for that upgrade cycle as well to whatever, to whatever degree it upgrades.
Jay Goldberg (05:57.56)
Right, CPU company executes the roadmap.
Ben Bajarin (06:00.616)
CPU, Apple Company, X-ray, so, man, okay. Don't just, they're gonna be mad now that we just called them a CPU. I know. In servers, yeah, yeah. Datacenter, yeah.
Jay Goldberg (06:03.692)
Right? No, that's a compliment, right? If you just look at it from that, they're, like, right? That's a good thing, right? It's like they're, because think about it, five, six, seven years ago, you couldn't have said that about AMD, right? And they've, what have they doubled their revenue in the last three years? I got to check my math on that. But they've doubled revenue in just a fairly short number of years.
Ben Bajarin (06:22.845)
Totally, yeah.
Jay Goldberg (06:33.28)
on the back of having good CPUs, right? So there's no shade.
Ben Bajarin (06:35.522)
Yeah, absolutely. Relevant, relevant to the topic we're going to jump into a good part of this tournament was returning to technical leadership, regaining technology quality was, was relevant to that. will say though, it's just fascinating to me. And now I remember following AMD when, you know, they were roughly five to $6 billion a year and they just did over 7 billion a quarter. So you're like,
Jay Goldberg (07:00.312)
Right. Right.
Ben Bajarin (07:04.888)
Good for him. All right, let's move to Intel. There's there's a lot to unpack here, and we're going to take we're going to take this a couple of different directions. We're going to start with what you and I briefly talked about and then to share with our reader. We're going to start with the Reuters headline that dropped on Monday, basically saying Pat's botching this turnaround.
I might be simplifying it, you can, you can give a better summary, if you will, but it was very sensationalist from a writer standpoint, however, kind of hit dealt some blows. But I want to talk broadly about that. What was your response when you would you What did you think when you read this article?
Jay Goldberg (07:37.349)
That.
Jay Goldberg (07:52.216)
So I wouldn't call it sensationalist. I know you didn't mean it that way, but I'm just putting that out there. I think it was, it painted a fairly bleak assessment of Pat's tenure as CEO. It pointed to a number of missteps, some self-inflicted, some external. And it, you know, was, it pointed out a lot of the...
Ben Bajarin (07:57.868)
Yes.
Jay Goldberg (08:21.326)
the flaws and some of the problems that have happened. It is interesting and I think it's important to note that this again, this is Reuters and they have real, like any serious journalism outfit, they have serious sourcing requirements and I'm just making this stuff up or spreading rumors. They talked about something like 48 sources. They interviewed 48 people for that piece. And that's a lot of people.
Ben Bajarin (08:33.207)
Yeah. Right.
Ben Bajarin (08:37.602)
Yep. Yep.
Jay Goldberg (08:50.094)
I mean, one of the problems that Intel has now is they've, about it in the last 10 years, they've gone through four CEOs and every one of those regime changes means there are a bunch of executives who leave. And as a result, there are a lot of executives out there who have something to say. And you could say one or two of them is biased to try and have an extra grand, but 48 is some serious journalism.
Ben Bajarin (09:09.228)
Yeah, right. Right.
Ben Bajarin (09:15.458)
Yes. But did you think, I guess here's the point. Did you think that the, I agree with you. And I think they hit on a lot of fundamental points, right? But did you think the kind of critique like was, I don't want to say fair. mean, obviously they want to be balanced, but that the critique was kind of accurate to the situation in.
Pat found himself inheriting, I guess, is the question.
Jay Goldberg (09:51.842)
Yeah, I think that is a fair defense. We're going to get into talking about the results in a minute. A lot of the things that they're doing now or things should have been done a year ago, years ago. Some of them, should have done a long time ago, but a lot of them are just like things that needed to get done and it was just left to Pat to shovel it. Can't kill explicit.
Ben Bajarin (10:03.992)
Yes, yes.
Jay Goldberg (10:22.082)
It's a very tough bit of nuance to parse there, which is he's not perfect. He's got a lot of detractors. He's made his share of mistakes and has his share of critics. But anybody who would be in that role would have to be dealing with a lot of these problems, no matter what, because they were created in the past. This is the biggest technical debt that's ever been paid down, I would say.
Ben Bajarin (10:49.954)
Hmm. Okay. I would agree with that. So, so I'm just, I'm just going to lob, I'm going to lob out my take and then I want you to come at it from the latter point that I'm going to make, because I don't know the answer. know you don't know the answer, but, essentially that you'll get where I'm going. So, so first of all, I think a lot of the problems that, that happened with Intel was because the two prior CEOs to Pat Gelsinger.
were a lot more financially minded, meaning that, you know, they took a little bit less of a technology approach and more about financial prudence, you know, returning to shareholder value, et cetera. And there's nothing wrong with that, but that led to a technological deficit in product and foundry. So I was always on board, like with basically, right. The plan he sort of said from the beginning,
was we're going to regain, you know, technology leadership technology is at the center of this turnaround. What I think nobody kind of grasped was that, mean, not, not that he said this, but that it kind of came with a, we're going to spend whatever it takes to get there and, then hope that this all fixes itself. And then financial, our financial futures change. Now the financial futures are taking a lot longer.
But when you kind of hear them talk now, it's like, well, we needed to spend this money. And that's, that could be part of the debate. I don't know the answer, but the question is, could you have gotten back to technological leadership and had some financial prudence at the same time? I don't know. But the point is now they're sort of saying we've spent that spend, we are on track technologically. And now phase two is bringing back economic prudence, right? Doing our financial plan.
that was laid out. I see these two things, my personal take, and I think you might agree is there might have been just a little bit more balance that could have that could have happened, you know, that maybe do it but also look where you could cut costs at the same time, and still reach your objective, maybe a more exhaustive financial analysis, right, could have could have come from that. But but I don't
Ben Bajarin (13:12.502)
I'm not faulting them for doing what they believe they need to do to regain technology leadership again, or at least become competitive technological because that's 100 % step one that needed to happen. Per right my point about what what Lisa Sue did with AMD. The main difference is AMD was not saddled with a gigantic foundry that was high fixed cost business. Pat was dealing with far more as you said, tiny financial debt than Lisa had to deal with.
Ben Bajarin (13:44.204)
So yeah, I think that's, anyway, that's my rant.
Jay Goldberg (13:47.786)
So, so I'm, I consider myself a finance professional first and foremost, who works in tech. I've been CFO of companies before. I will say this. I want to, I want to be very clear on this point. The CEO of a semiconductor company should never be a CFO. I have been the CFO come from the CFO side. And there are people out there are going to be very mad at me for saying that. there are
Far more people are going to be like, yeah, right on. So I absolutely agree that the past management's over focus on financial metrics is what got them into this mess. There's other factors as well, but that was a big part of the problem. In terms, of this transition from overspend, know, spend whatever it takes to now rationalizing your spend,
I think the picture is much more, it's not as clear. I think there are many, many things that should have done earlier on this side. And we've talked about them before. And they're just the fact that like one step after another, they're like, we should have done this a year. The dividends, the obvious one. But there's lots of other things. And like last quarter when they had that big blow up, they started putting in all kinds of things.
Ben Bajarin (14:57.144)
Mm.
Jay Goldberg (15:14.614)
really should have done years ago. And one of the things that stood out for me on this call last week was they came with a clear plan, right? They had really thought through everything. And I think what happened last quarter was they got surprised by it. I remember thinking that everything they've done since then has been like, they only found out about the shortfall really soon before earnings.
Ben Bajarin (15:35.033)
Mm-hmm.
Jay Goldberg (15:43.488)
and they were in a panic, not a panic, yeah, you know, they, they, every, and everything seemed very rushed. The way they laid people off seemed very rushed. There's all kinds of very obvious signs of last minute ad hoc planning, as opposed to this quarter when they very much more had their act together. They had answers to questions. They could think they thought through a lot of those things. And I think that was telling, because I mean, Intel's done big reorgs in the past. They pivoted in the past and last quarter looked kind of sloppy and they've, they've caught up for it now.
That doesn't mean they're out of the woods, it's certainly, it's a positive sign. I think what worries me though, is if you look at the story they present in their earnings call, or they presented in their earnings call, and you compare that to the story from Reuters, it's like two stories about the same thing.
just with slightly different interpretation of past events, like one of those avant-garde French movies. Same story. One's very class half full, other's class half empty. And I think the problem that Intel faces right now, is going to face for the next year or so, is they burned a lot of credibility last quarter with that blow up. This quarter certainly helped regain it, but there's lots of people who do not trust it. Right?
Ben Bajarin (17:10.807)
Hmm.
Jay Goldberg (17:12.462)
And the rotor story and a lot of other things that have been coming out about them provide a very compelling counter narrative to what management says. And they're going to have to walk a very tight line to keep that going. Because as much as there's a lot of positive stuff on the call yesterday, was still like pretty not a great quarter. There weren't good numbers. Just wasn't as bad as everyone feared. So that's a tough spot they're in.
Ben Bajarin (17:32.504)
Yeah.
Ben Bajarin (17:37.002)
So, all right, so let's talk about that, right? So it felt to me like the big push in prepared remarks and sort of their supporting narrative was we're regaining fiscal responsibility and we're executing this path. They listed all the things they've been doing, right? Shedding headcount.
They listed the cost that that was right. mean, to two over 2 billion in severance, right, to call it out. So they're, they're, they're, they're showing their costs, but also, you know, they're, they're executing all the things that they said they wanted to do. one of the more kind of disappointing ones, and I, just find it interesting that nobody reacted negatively to this was that they basically said, you know, Gaudi is not going to hit its, its price target, which really shouldn't shock anybody. I'm guessing it didn't shock anybody.
But as we've said before, an AI lift is not priced into Intel. It's kind of void. There was the announcement with Amazon, which we kind of knew already was scuttling bug around networking and then also some quasi customization of the CPU part, you know, it's fine. No real update to Foundry and...
Jay Goldberg (19:00.214)
No, had two new customers. yeah.
Ben Bajarin (19:01.878)
Well, to customers, but the revenue pipeline didn't move. They didn't make a projection. said, yes, customers, but that number didn't change. think that's kind of what people were hoping was sure. Move it from 15 billion to 17 or something. Right. But, that number stayed the same. and then a little bit technical clarity, which I thought was interesting on kind of the TSMC margins to our costs with lunar Lake, where they, obviously we know, right. They're paying,
a higher cost than they would to Intel foundry to TSMC, but they made the point to say, one of those one of those costs was a design decision to bring memory onto package. And we're not going to do that going forward with some percent of tiles coming back to TSMC or to Intel. I think people are thinking it's roughly 70 to 80%, probably not 100%. But the vast majority coming back to Intel for 18 a
Jay Goldberg (19:49.218)
Yeah.
Ben Bajarin (19:57.712)
So structurally, I think you're right, right? People were like, okay, it sounds good, but at the same time, exactly what you said, right? I don't think people are like, whew, the worst is over. Like they're still unsure, you know? However, going back to what we've talked about before, I feel like somewhere around this time last year, people really did feel the worst was over. And that was why the stock gap got to the 40s. And we don't have that confidence right now.
Jay Goldberg (20:22.926)
Right.
Jay Goldberg (20:26.498)
So I think the problem is that all those things you just pointed to could almost point for point be countered in the Reuters article. You think about like Gaudi, Reuters asserts that Pat oversold Gaudi to the street. Customers for AWS, sorry, customers for Foundry, for IFS.
there's still, lots of, Reuters has done other reporting saying that's not going so great. Relationship with TSMC, they had a discount and then Pat botched that somehow. I, I don't, I don't, you know, I'm not taking sides on any of these. think there's nuance and gray in all of this, but it's interesting that like a lot of the things that they held up as positives, probably in part to recognize that the Reuters story was out there, are sort of.
Ben Bajarin (21:09.304)
Sure, sure.
Jay Goldberg (21:21.294)
spun very negatively in the Reuters thing. know, it's a tricky one. I think their big issue is credibility with the street. And they're gonna struggle with that. It's gonna be a while. And you know what, I kept like, when I saw the stock was up so much and I was sort of reading before I read the whole transcript, I kept thinking like, this is the perfect like meme, right? This is like,
Ben Bajarin (21:21.558)
Yes. Right.
Ben Bajarin (21:37.386)
Agreed. Agreed.
Jay Goldberg (21:50.412)
the guy walking down holding hands with his girlfriend and turning to look. And it's like, the girlfriend is Intel's falling apart. And the new hot thing is Intel's doing really well. just like, just complete.
Ben Bajarin (22:01.764)
Hey, the guy looking back and forth. Yeah. Yeah. Okay, so so I do want you to clarify from the CFO perspective, because maybe this is building back confidence. I asked you this briefly, and I want you to expand on it. So what? Why did they report gross margins the way that they did? I felt like
Jay Goldberg (22:05.186)
We're doomed. We're saved. We're doomed.
Ben Bajarin (22:29.6)
the analyst takes to this was like, well, I get it. But why? Like you didn't need to do that. And so I'm still perplexed with this other than it's it's deliberately being conservative and transparent. But I feel like there's more going on. But I'm not. I'm just not sure because it seems like an odd decision because it looks like on paper if you didn't. man, they had 30 something gross margin last time and now it's 18.
Holy cow, burn the world, right? But they didn't need to do that because it wasn't actually 18. There's a reason for it that's a fine financial reason. So I just want you to elaborate on this from a CFO perspective.
Jay Goldberg (23:12.196)
Yeah, so this is a good example of things they should have done last quarter. I think what Intel did this quarter was they kitchen-sinked the numbers. They took every possible potential bad news or bad number they could, and they threw it into this quarter because they took a massive restructuring charge. I don't even remember, $10, $15 billion, massive charge. And they wrote all this stuff down, and it's like,
It's everything, it's taxes, it's Gaudi inventory, it's all these other things thrown in there. And the idea is behind kitchen sinkies, you throw all the bad news, you lump it all together in one quarter, and then that eases comparisons in the future, right? Because they warned that next quarter gross margins aren't gonna be great, it's not gonna, because it's below 40%. But I guarantee you next quarter, people are gonna say, my goodness, their gross margins are doubling. They were 18 % last quarter, and now they're 39%, like,
Ben Bajarin (24:06.04)
How do you not know? How do you not know this? Like if you follow this company? That's what I'm weirded out by.
Jay Goldberg (24:12.472)
So they are giving themselves a lot of wiggle room because they know they're going to need it. the thing is when you start pulling apart, and I haven't done this yet, I haven't read the 10Q, but I'm sure when I read through the 10Q, you look at all the things that are in the restructuring charge and that there's this big weird tax thing that I have to parse. But those are all things that are sort of taking place now. It gives them insurance in the future when the other things are going to be bad. They'll get a point or two gross margin, two, three quarters from now.
when they have something else that's going wrong that they need. And believe me, that'll work because it's a joke among sell-side analysts. We spent all this time writing long notes and 99 % of the investors who read our note will read the headline and that's it. Maybe that's 90%, another 5 % will only read the front page. And getting into all that murky detail is just like...
know, there are fund managers who have to cover a thousand stocks and they don't have time to get into all the murkiness. It is from an investor relations standpoint, it's a pretty smart move. And in terms of that gross margin calculation, they could have absolutely sort of non-gapped that away, adjusted gross margin, they chose not to. I think it's very, a big part of that is just because they gives them some wiggle room and they can always, you know, a year from now.
when their gross margins are not quite 40%, they'll be like, we're great. Look, we're so much better. It's an easy comp next year.
Ben Bajarin (25:46.27)
Okay, well, we will we will all remember this. I brought it out. So everybody will never forget that Intel reported lower gross margins on purpose didn't need to. And yes, anyway, there are there were strategic reasons. In fact, I saw people diagnose this too, that like, they could have foregone all of the things that drew it down to 18 and just said, well, let's just take a
a 3 billion hit off that 41 and say it's 37. And that would have been reasonable to like, I get it's going to be low, like it's only so I looked this up, it's only been in the 40s. Grosso Mendes have only been in the 40s twice in the past eight quarters. So theoretically, they've been around that range, like they need to be in the 50s. That's what they used to be, right? That's to get as far as semi companies should be everybody else is in the 50s, unless you're Nvidia. And, but
Right? They could have made it a little lower. Like I said, they chose to take what you said, the kitchen sink option, take it to 18. But but I say that because a lot of people looked at that number and didn't know why. And we're like, it's 18. That's, that's horrible. Like, yeah, if it was really 18. That would be really bad. It's it's really 18. But it's explainable, right? That's the difference.
Jay Goldberg (27:05.122)
Yeah. Yeah. I, yeah. I mean, I saw an exchange on a discord server that, that's filled with semi's people and they're like, one person's like, brother, this company is going out of business. They lost $15 billion. And we're like, did they, I mean, they, generated cash in the quarter, which was, which was good. I, but I guarantee you a year from now, they're going to be ramping AT &A, the volume production. They're going to need all the help they can with optics around gross margins.
Ben Bajarin (27:19.202)
Right. Yeah. Yeah.
Ben Bajarin (27:33.496)
All right. Good, good, good. All right. We'll pin that for this quarter next year. Set comp. That will be fun. Put a pin in it. I will say, go ahead. No, no, no. I will say the other thing. Hold on. gotta get a prop for you, but it's something that... I'm gonna get the prop. You talk.
Jay Goldberg (27:43.854)
I'll tell you that... Go ahead.
Jay Goldberg (27:52.974)
So one of the things that really continues to worry me is Intel's culture, right? Because all the stuff that we're talking about really comes down fundamentally to culture. I'm encouraged by how they took their approach to this quarter, because it showed a much more coherent approach to all of this. I think I actually wrote a note from this right back when they blew up last quarter. Like, here are all the things that they should be doing. And they actually, you know.
Ben Bajarin (28:00.222)
yes, let's talk about that, yes.
Jay Goldberg (28:23.246)
It's no great inspiration on my part, it's no genius on my part, it's just like very straightforward investor relation steps. And they started doing all those things, which makes me feel better that they're engaged and thinking about this coherently. But I'm still like in the back of my head, like a lot of their problems were just the fact that they weren't being fully honest with themselves as an organization. And that there are little, I mean, that's a big, think sort of underlying theme of the Reuters piece was deep down there's like just
there's a mismatch between what the organization feels it has to be doing and the sort of dire straits that the investors see the company is in. And that's never gonna be easy to solve, it's gonna start becoming more important, not just for survival, but as they become a foundry, like they gotta fix that. And I'm not convinced that they've embraced that yet.
Ben Bajarin (29:17.986)
Yeah. So I think if I was to just assess what you just said is I actually think the, right, they said that Foundry is going to be its own subsidiary that had kind of pre-announced that going into earnings. I actually think Foundry is culturally more aligned to what it needs to be in terms of its goals, its technology, its process. I think they're being smart about how they spend because they know they just can't.
But the cultural part you brought is a really interesting one because Pat Gelsinger went on CNBC, had a whole sort of ramp. But one of the things he said was, you know, look, spending was a little bit like we, we, did, we had some stuff we didn't need. We had machines we didn't need. So he was kind of saying people on these teams were sort of just buying stuff that was maybe of excess. Right. And so I think the point that you make is, is needs to be even more hit home with the product side.
Jay Goldberg (30:01.643)
Yeah.
Ben Bajarin (30:15.768)
to still recognize what they need to do to be competitive, but not be loosey goosey with the budget, right? Still say, do I really need this? Is this essential? If I can do without it, still hit my goal. Let's do it. Let's save the company some money. And that's a, again, I'm not, that's not typically the way engineers love to work. So, so I get it. I'm like, no, no, of course I need it. I have to do this. But to some degree, that's gotta be in people's minds, managers minds, upper, like just to bring a little bit more of that.
What you want? How would you describe it? Fiscally responsible mindset to product.
Jay Goldberg (30:54.04)
they are exercising their fiduciary duty to shareholders.
Ben Bajarin (30:59.116)
Yes, yes. So for everybody, I'm holding up a visual of what is probably going to be a historical relic, which is also what I gave to your son in the Intel discrete GPU. I'm holding an ARC A770. Nobody's said this yet, but I think it makes a ton of sense to not keep spending on this product. You are not going to be Nvidia or AMD in discrete GPUs anywhere. You're not going to word them in servers.
Jay Goldberg (31:01.304)
the old term.
Ben Bajarin (31:28.98)
either. So discontinuing that business, I think, is inevitable, in my opinion.
Jay Goldberg (31:36.504)
So they actually got asked on the call, like, you've just said you're done with Gaudi. You've written down all your inventory. You're not going to hit your targets. And somebody asked, can you compete with NVIDIA in AI? And they didn't really directly answer that. And so what they did say was CPU were the preferred head node for CPUs. Sorry.
We have the preferred CPU for AI head nodes, which is okay, fair. And then they started, he sort of pivoted to talking about inference and, you know, head nodes become more important and CPUs become more important in those workloads. But he never directly answered the question of like, are you going to compete in AI? And that raises to me, which is a question I keep trolling you with, which is what is, comes, what does come after Gaudi? Right? Cause on paper roadmap, it's Falcon Shores.
And they didn't mention that once yesterday or yeah, on the call today. They didn't mention it once. Like where's what, what is the plan for Falcon chores? Is it still on the roadmap? Cause you know, if they really, really want it to be like brutal about this and like batten down all the hatches, they would push Falcon chores out and say, we're not even going to bother with AI until we get everything else sorted out. How's that for a bull plan?
Ben Bajarin (32:46.765)
Yeah.
Ben Bajarin (33:01.366)
Yeah.
Yeah. No, again, I just think this underscores.
this underscores the challenge, hard decisions are going to have to be made very, very strategic on where they're, where they really have, I guess they need to come to grips with really where can they compete? Like, and I've started thinking about this just for all of our listeners. I think a lot of, a lot of companies, like, and you, ask business analysts, you know,
whoever, right? We'll have a thesis on a company and it's generally a market thesis with a business thesis. What rarely exists is a technology thesis, meaning like, where like, what is the strength of your technology differentiators? And I think that that helps define where people will compete. like, for example, if we just say that we were sitting there with with Gelsinger or others on the product team, and we said, Okay,
What's your tech thesis? Like, what are you betting on technologically? I think he'd say, right, a handful of things, right? Product, process, et cetera. So again, if that's really your thesis, then have that moment of, we can't compete else. Like, we just can't. Okay, focus on what you can. Focus on the bet that you're making technologically. And I think that's the right approach. The smartest companies in the world, in my opinion, know that. Like, they know.
Ben Bajarin (34:31.98)
where they can compete, what their assets are, what their strengths are, what their people is, they know it. And they just, double down on that. And that's, I feel like some companies really struggle with just going outside their swim lanes. And it's unfortunate for Intel, right? You've missed AI and that might lead actual to an actual disruption moment over time for them. That doesn't mean that they're not going to continue in some fashion or, or wait out 10 years and hit something else.
I don't know. But my point is, I think that's gotta be where they focus, right? Is, okay, we can compete here. Focus on that. And that will be one point of helping you navigate through very choppy waters of the next few years.
Jay Goldberg (35:17.41)
Yeah, you know, just thinking about what I said a second ago, I like that idea a lot more. It's like, you know what? We're not going to compete in that. We're going to stop spending a billion dollars or whatever it is to develop out our accelerator roadmap. We're going to make CPU. We're going to focus on CPUs as head nodes. We're going to focus on Foundry and on the CPU side, let's, let's go partner with Nvidia. Wouldn't that be wild? Like stop making an accelerator, go, go, go talk to Pat and Jensen up a conversation.
Ben Bajarin (35:38.722)
Yeah, absolutely. Yeah.
Jay Goldberg (35:48.098)
Jensen invests a little bit of money and agrees to use 14A. I don't know. I'm just making this up, like, that'd be huge for both.
Ben Bajarin (35:54.744)
Well, but exactly what you said, though, like, well, why not take so I keep coming back to this just from a technical standpoint, right, that the great some of the greatest threats competitively to Intel's not just Nvidia, it's the hyperscalers making their own stuff. And so so why not do more of what be more liberal with customization like they're doing with Amazon like much more liberal.
If in fact, if there's the threat that you're actually going to lose, become a minority share in AI headnote or host CPU, then just go, I'll do whatever you want to not get designed out. I'll do whatever you want. I don't know, custom make them x86. Do you know what mean? Like, but I feel like that's like the moment to say, if we believe we've got technical chops to compete there, like we're, we're, we're willing to be liberal. Like we'll take a semi much more semi-custom approach to merchant Silicon than we've ever had before.
Because the threat is you just get designed out by their own internal efforts.
Jay Goldberg (36:57.015)
Well, I think if you look at it, that's what Foundry is. Really unleash Foundry and say, go help all the hyperscalers make their competing CPUs.
Ben Bajarin (37:09.772)
Well, mean, I think they'll do that. My point is do it with x86. Do it with product as well.
Jay Goldberg (37:15.906)
Yeah, or not. Let him stand alone.
Ben Bajarin (37:18.689)
Or not.
Ben Bajarin (37:22.232)
All right, let's move on. Now is our rant on Intel. Okay. So I'll let you sort of jump off on the mobile insights, starting with Corvo, which we'll kind of talk about MediaTek and then give a preview to Qualcomm next week.
Jay Goldberg (37:24.226)
Yeah, let's move on.
Jay Goldberg (37:37.622)
Okay, so this week we had MediaTek and Corvo report results. MediaTek, as far as I can tell, printed a pretty good quarter. Stock was off a little bit because again, it wasn't up to expectations, but numbers looked fine otherwise. Corvo then reported the next day and their stock is off 25%. And the culprit was they lowered their guidance for both the quarter and they gave a lot of warning signs about next year.
Ben Bajarin (37:57.24)
Thank
Jay Goldberg (38:05.772)
And what they said was that they're seeing a big shift in the Android market to low-end, low-priced phones, which don't need as much RF content. And so there's two ways to read this. One is you can say, the whole Android market itself is moving to lower price points at a greater clip than it has been for years. The other thesis is that Corvo is just losing critical share in places.
I tend to think it's the latter. I Corvo is, I think they're in, they're losing out in flagship, think, right? Samsung is moving pretty hard towards Qualcomm and their next flagship line, Galaxy. And I think that probably comes with some RF as well, some Qualcomm RF. And then Apple itself is likely doing a modem next year and that's going to have more Broadcom RF in it. And that's bad for Corvo.
And so it is, yeah, the, the, I think the, the S smartphone market is there's nothing really wrong with smartphone market that hasn't that nothing new there, at least we're not seeing everyone moving to $400 to $200 Android phones. but I think Corvo's losing a fair amount of share.
Ben Bajarin (39:26.914)
So doesn't that also mean Skyworks probably is too? I haven't followed them up in a while, so I don't even know where they stand, but those two companies typically jockeyed for some level of RF content. wouldn't the same, if what you're saying is true about Qualcomm, picking up RF, which I'd agree with, Broadcom likely, with Apple, unless Apple just completely designs them out and does something else. But Skyworks typically would have some of the same dynamics, right?
Jay Goldberg (39:53.186)
Yeah, so I think in the current cycle, Skyworks has less to lose than Corvo does. In the current iPhone, Corvo had done well last year because they picked up a bunch of sockets. It looks like they've lost those or they think they're going to lose those. And they didn't lose them to Skyworks. mean, the problem with both of those companies' stocks is that they trade almost entirely based on year to year their position in the iPhone. And that's not been great for either.
Ben Bajarin (40:15.352)
Yeah, right, right, right.
Jay Goldberg (40:22.958)
Yes, I think Skyworks had less to lose and I think it picked up a couple of things here and there. The reality is if with Qualcomm having an RF line and Broadcom having ambitions to grow its RF line, that just doesn't leave enough market for both Corvo and Skyworks.
Ben Bajarin (40:43.426)
Yeah. Yeah.
Jay Goldberg (40:46.51)
And so it's like, mean, ultimately I think one of them will exit or get bought or like my crazy idea is that they merge and they split the two companies into mobile and non-mobile lines. That'd be a good idea. I don't know what's going to happen there, but it is, I think it's the market in its current state is not tenable long-term.
Ben Bajarin (41:10.616)
Yeah, agreed. But you'd say that's within some of those, some of those dynamics that might maybe, maybe, I don't know, Qualcomm's RF business or more importantly, I think people are seeing strength to the high end ASP of Android, might be a, might be an okay, some, a bright spot for Qualcomm next week.
Jay Goldberg (41:30.528)
Yeah, I mean, I think I Qualcomm.
Qualcomm is in a good position for the current cycle because they picked up Samsung, because they picked up some of the other China, some of the China high-end flagship lines. I don't think it's RF that's driving it. Their RF stuff is not particularly exciting at this point, but it's there and it's okay. And so I think what's mostly happening is it's bundled with the modem.
So I think we're at a point in the cycle where Qualcomm looks okay. Next year, we'll see what Apple does with their internal modem. That gets a little dicier there.
Ben Bajarin (42:14.317)
Yeah.
Jay Goldberg (42:14.786)
Right. And who knows? We'll see if Samsung gets its act together with Exynos and maybe comes back next year.
That's an if. Big if.
Ben Bajarin (42:23.542)
Nope. Zero chance. All right. All right. Good talk. Next week we've got Arm and Qualcomm, probably something else because this industry never sleeps and hopefully no more drop bomb articles about Intel. They need a little bit of a break. If whatnot. Anyway. All right, everyone, until next time, we appreciate your feedback, your notes to us.
Jay Goldberg (42:45.55)
you
Ben Bajarin (42:54.144)
Send us questions, et cetera, and we will talk to you next time.
Jay Goldberg (42:57.858)
Thank you everybody, tell your friends.
Ben Bajarin (43:01.496)
See you later.
Jay Goldberg (43:02.35)
See you later.