Fundamentals over everything

0:10Sara Awad from Tech Contrarians. Always great to have you on Investing Experts. Welcome back to another show. Thanks so much, Reina. It's great to be here. It's great to also be on the show in such a special moment for the semiconductor industry and tech more broadly. Absolutely. And I think it's always very edifying for us as an audience, for me as a host, to get your perspective on the tech space. Always a lot going on in recent years. So talk to us about what you're thinking about in this moment, middle of May 2026. What are you most focused on? What are you most thinking about? So I think, you know, center stage for everyone has really been the semis.

0:49And that's been the case for, you know, since this AI cycle really began. but specifically in 2026, we're seeing a lot more momentum leaning into semis. And it's interesting because it comes in contrast to the last time that I was on this podcast, which was, I think it was towards the end of March. We were in a very different backdrop to where we are today. I think on most of the tech sector, if you take a look at the year to date chart, you kind of see that uptick from late March throughout April and then continue into May. You know, the infamous saying of buy in April and sell in May does not apply to 2026, at least not this far.

1:24And in terms of, you know, the setup, what we're really looking at is the backdrop. So you have oil above$100 a barrel and even much higher in terms of the physical market, specifically out in Asia. So there's a huge gap there between oil futures and the spot prices. You also have a very fragile ceasefire, to quote Trump, on life support. right and then you also have markets at all-time highs and i think it was in an episode you had with gary von not too long ago that he kind of said this setup sounds like two truths and a lie but it's all true uh and we agree right it's it doesn't make sense i think that if you told anyone a year ago that we'd be here today they wouldn't believe it um so i think that the market is in a really special place and the context matters a great deal deal here we always go back to the context you know us we really like to kind of lay the ground the foundation there but i think it's specifically important to mention as part of the conversation today because it's all interconnected, right?

2:20No one stock kind of exists in the vacuum outside of this macro backdrop. And so in our last episode, what we were talking about is that, you know, Trump is a businessman. First and foremost, he is very much in tune to financial markets. And I think that's really a big part of the conversation regarding the ceasefire today, because I think part of the back and forth that we're seeing are the lag on the ceasefire and on and on confrontations in the Strait of Hormuz is partially because Trump is not pressured by financial markets, right? They're all at all-time highs. They've already priced in a resolution to this conflict.

2:55And to an extent, there is somewhat of a strong investor confidence in Trump wrapping this up neatly, right? And that's what the markets are reflecting with the stocks up 66 % here today and the S &P 500 up, I think, roughly 8 % here today. So that's kind of the setup, right? Now, if we take a step back, this isn't really a green light that everything's going to be smooth sailing from here. I think the conversation of a correction from Wall Street is picking up more. We're hearing Apollo talk about a correction. We're hearing JP Morgan talk about an exuberant market. there's a lot of conversation building, I think specifically this week.

3:33And something worth flagging on that conversation picking up is that usually the buy and ask questions later activities very often followed with a sell and ask questions later kind of activity, especially as we see more retail momentum entering this market. What would you say about the tech correction for investors following along and wondering when slash if the bottom is going to fall out? What would you say are the points that you're paying attention to? And how are you thinking about it? The way that we're thinking about it really at Tech Contrarians is that a correction isn't going to come out of nowhere.

4:08I think it's going to be forced. And when I say forced here, what I mean is that it's going to be forced by an offshoot of the current macro backdrop, right? The higher oil prices leading to higher transport prices, leading to higher cost of goods, creating stickier inflation and a whole kind of series of events that unfold from that standpoint, right? And if we take that backdrop and kind of how it's progressing, especially after the PPI and CPI reports for April this week, and markets actually reacting to that, something that we haven't seen them react to earlier. If we take that and we couple it with all the supply constraint kind of conversation we're getting on tech earnings this season, I think it sets up for a very rough earning surprise next time around for next earning season.

4:53And what I mean here is that expectations have become uber high. And when it comes to tech, at the end of the day, you have physical constraints, right? You have physical constraints of memory, you have physical constraints of different kinds of components, and you have a timeline, right? You have lead time. So there's a huge structure that the supply chain needs to fit into. And I think that if we have a combination of higher costs coupled with supply constraints, I think this can set up for a very difficult time to see a positive market reaction to print next time around, given the expectations are so high.

5:26And I think we've seen part of this already if we contrast the first half of earnings season to the later half of earnings season so far. So in the first half of earnings season, we had stocks flying double digits post-Brit. We saw that with Texas Instruments, and we saw that with Intel, of course, we saw that with a couple of different names. But then towards the end of the season with Western Digital, with Monolithic Power, that where I would more accurately say mid-season. But with those guys, we had great reports. We had very strong, you know, upbeat outlook. But the market didn't react. In fact, you know, Monolithic traded down after the print, as did Western Digital.

6:02And so I think that what this is reflecting is that a lot of the positives are getting priced in into print. And that creates a huge risk on the print itself. So instead of earnings acting as a catalyst, earnings are more acting as a test now. And I think that's something that will become more pronounced in Q3, because that's when we're really going to see a lot of these expectations be put to the test. And although there are companies that are doing great and the fundamentals check out, I think the expectations are the point of concern moving forward. We have NVIDIA earnings coming up next week.

6:34What would you say in that context? and also maybe just anything to note there as we're speaking of earnings? Yeah, I think that's a spot on question, especially with Jensen making his last minute trip to China, which is exciting because I think there's a lot that'll come out of it. And I'd love to get into that a bit more towards the end of this podcast. But heading into NVIDIA print, we were bullish on NVIDIA, I think when the bulk of the market was kind of losing face, when the stock was pulling back into the 170s, high 160s kind of range. We saw that there's more upside ahead based on the fact that at the end of the day, there's no actually fundamental problem with NVIDIA, with the business itself or with the stock itself, at least not for 2026.

7:16It's not that they're going to lose share to ASIC in 2026. I think that's more of a 2027 conversation. And that's when it'll be time to really take a step back and reassess NVIDIA and where the upside surprise could come from. But heading into print next week, the stock is at all time highs. I think most of the positives are well recognized by the broader market. And the stock reflects just that. I think the only kind of offshoot variable when it comes to NVIDIA's print, because, you know, a$10 billion beat to expectations or sorry, sequential growth is not undoable for NVIDIA. It's very manageable.

7:54It's something they've done before. But I think that what NVIDIA needs now is a wild card. And that's why I mentioned Trump and China and this visit that's going to happen over the next two days. It's because I think the upside surprise for NVIDIA could come from regaining that China market. And that's something that we've talked about a lot on and off, that as the US kind of advances in terms of advanced nodes, the gap between the US and China when it comes to basically tech capacity and tech credibility widens, right? China lags behind because there There are a lot of constraints as far as the semi-cap equipment goes.

8:32There's a lot of constraints in terms of their ability to access UV lithography machines from ASML, which they do not have access to, even advanced UV machines they don't have access to. And these are the machines that you need to use to progress technology to advance the smaller and smaller nodes. So our expectation was as the gap widens between the US and China, we're going to see Beijing eventually allow H-200s back into China. And we know that they're not against this idea entirely because we've seen AMDs and MI308s make their way to China. And so that's something that we're, you know, we're staying in tune to specifically with this visit.

9:05But none of this is priced in, which is why to us it's so appealing or attractive, because that's the one thing around NVIDIA that the market really has not priced in yet. And then if we take that and we think about what Jensen said in terms of the market share in China at the moment recently, I think it was earlier this month. He said that NVIDIA had, you know, 90 something percent of the world's total market share. It has today 90 percent of the world's market share. But today the share in China has dropped to zero. And so China is huge, right? They're the biggest AI. Sorry, they're the biggest smartphone market, their biggest EV market, their biggest PC market.

9:44And they're a huge market when it comes to AI and AI servers. So I think that would really be what would push the needle on the print itself. But outside of that, I think most of the positives have been priced in. I think a lot will come down to commentary around Jensen and the Vera Rubin ramp that's expected to be scheduled for Q3. So we're optimistic about that. We think the Vera Rubin ramp is the last big hooray for NVIDIA, at least in terms of its current outperformance cycle around the AI GPU. And then after that, it'll be time to take a step back and reassess specifically as we're seeing ASIC gain so much adoption.

10:18And that's where I think we turn to look at these new and upcoming guys in ASIC, not necessarily new, but you're kind of gaining more traction with the hyperscalers like Broadcom, which is really the staple name for ASIC with six of the largest customers under its belt, and Marvell, which is really up and coming, competing with Broadcom and really benefiting from what looks like it's growing into a dual source kind of ASIC market, right? So for some context here, Google really led the ASIC adoption with the TPUs. And that was the reason we were bullish on Google early last year. We thought that their TPUs would become more operational and efficient faster than everyone else.

10:59And that proved to be the case. And so Google, which works with Broadcom, has now decided to also work with Marvell. So it's pursuing a dual source strategy. And I think that we'll see more hyperscalers pursue a dual-source strategy as ASIC adoption increases. And so as that happens, I think NVIDIA's share is pressured. They're forced to share the AI market with ASIC. And I think that's when we turn from NVIDIA and we look at these other players because NVIDIA will need to reset. It will need to re-rate on this new reality. So who are those other players? Yeah, so that's where we have Marvell and we have Broadcom.

11:36These are two of our top names when it comes to the semi-space. And so when it comes to Marvell, Marvell's done exceptionally well year to date. And they've done exceptionally well as of their last earnings print, which was in early March. So, you know, they guided for the April quarter to be$2.4 billion, outpacing Kazas is for$2.28 billion. And they gave investors their fiscal year 27 and their fiscal year 28 outlook. And I think that those are conservative, considering everything that's happened intra-quarter and the visibility that we're having on ASIC adoption as we move throughout the year.

12:13So I think there could be a lot more upside still for Marvell. And in the same extent to Broadcom, which is a more, I would say, broadly recognized ASIC player in the space. So Marvell has, I would say, more of an advantage being able to, you know, revise upwards its full year guidance to reflect its current reality and to also secure more of these customers as hyperscalers go for a dual source kind of strategy. But when it comes to Broadcom, the results have also been stellar. For the April quarter, they're guiding$22 billion ahead of$20.4 billion consensus. And they're expecting essentially$100 billion in 2027 from across their six AI ASIC customers.

12:56And so we could even see that number revised up as well, because there's time to see this adoption gain more momentum. And so I think that for NVIDIA, ASIC is now a very relevant part of the conversation, looking out towards the next four quarters. But I would say for the next two quarters, especially into that Vera Rubin ramp, that's really the armor that NVIDIA is shielded with. They have the hottest products in the market still. They still have the superior designs to AMD. And they are experts at what they do. And I think that goes a long way, especially in the kind of supply chain backdrop that we have today, because the supply chain is half the equation.

13:33You have to get the supply chain ready. You have to get all the components ready. and you have to make sure that you can secure that capacity at TSMC. A few stocks that have been in the news recently that have had nice runs and that we've been talking about on the podcast, stocks like AMD that you've talked a little bit about today, Micron, Credo. We had Steve Cresson talking about that Credo versus AMD, although recently AMD has shot up. Quite a bit. Maybe provide some context for those names and how you're thinking about them and feel free to kind of bridge the gap with any other players as well?

14:10Yeah, definitely. I think the names that you just mentioned are some of the hottest guys today, especially when it comes to semis. For Credo, last time we spoke, the stock was around$93. And that was when Credo was coming under a lot of heat, this whole idea that the industry is going to move from copper to optics and that Credo would be lost because they really are the king of AEC. And that's really their advantage. They're the king of copper. But the thesis that we had for Credo is that it's not going to disappear overnight. Copper is not going to go away. The industry roadmap is built around the efficiency that copper provides.

14:46And as we progress, we're going to need to use optical, but it's not going to be an overnight kind of transition. And Credo was already with a zero flap trying to situate itself in what's expected to be the industry's future. And then we saw them really push towards that with the Dust Photonics acquisition and the stock really break out from there. For Credo, we think that they still have a nice growth run right ahead. The stock has definitely priced in the positives of that. But I think towards, you know, second half of the year and then 2027, as they build up their expertise and integrate the acquisition, I think they're in a much healthier position.

15:20So for Credo, I think there's a lot more to love and that we could see this, you know, acquisition really position them favorably in the next leg of where the industry is headed, specifically when it comes to AI servers. And then if we switch from there, sorry, yeah, if we switch from there to AMD, so AMD has been on a hell of a run. And so, you know, they reported earnings in earlier this month, if I'm not mistaken, around May 5th. And the stock really shot up. I think part of why this happened really links to Intel. So I'm going to kind of discuss the two side by side, because these are the two players that are really riding the CPU renaissance moment, along with ARM, of course.

16:04And so these three guys are the guys that are focused on the CPU. Intel, AMD, and ARMS earning calls were very much centered around this agentic AI moment that's going to require a different kind of ratio of GPU to CPU that's more favorable for these players. And that's definitely true. With agentic AI, we are going to see a ratio shift. The only catch for guys like AMD, I would say, if we're looking specifically at AMD first before moving to Intel is that AMD is somewhat of a chameleon, right? So when the AI boom began and everything was about AI GPUs, AMD was racing to compete with NVIDIA on the AI GPU front.

16:43And there, of course, they are, you know, they do have part of that share, but they did overall lag behind the kind of momentum that we saw NVIDIA deliver. And now that we're seeing the CPU conversation pick up, we're seeing, you know, Lisa Su shift modes and focus on the CPU catalyst under this CPU agentic AI moment that the market is having. And so I think that in that sense, it's important for investors to know that because it's different than being a first mover and capitalizing on the strengths that you know that you have. For AMD, the strength for us over the past two years has really been their positioning when it comes to CPU and the gain share there against Intel, right?

17:23They were executing very well, but their AI GPUs were not executing as well. their report in Q1, you know, they missed data center expectations. Without the China added shipments, they would have missed data center expectations. And then this quarter, again, on the earnings call, they got a great question specifically about whether AI GPU sales grew sequentially. And management kind of just beat around the bush. They didn't want to answer. They don't want to give a clear answer. And the interesting thing is that into their data center sales segment, they bake the CPU server numbers and the AI GPU numbers.

17:58And so they looked extremely strong, but that strength, I don't think reflects AI GPU strength. Instead, I really do believe it reflects the CPU strength that we're seeing. And that's something that we have to know when we look at AMD because expectations are pricing upside for both, not just one or the other. And so I think that's important to look at when we're talking about AMD. And I think given the stock's run up, it is in an extremely more risky territory right now because there's a lot more proof points that we need around agentic AI for AMD to really have that catalyst that's going to drive financial outperformance because the expectation of the benefit from agentic AI is already comfortably priced into the stock.

18:36So we need more tangible proof points. And as of now, what most people don't realize is that agentic AI is still not a reality today. And so that's the case when we're looking at AMD. Now, if we shift to Intel, Intel has a similar situation, although it gets a bit different with the foundry side of their business. So when it comes to Intel, they entered the quarter with very low expectations. So Wall Street's expectations were pretty low, especially after last quarter's print. And so we turned bullish on Intel actually into print because we took a look at expectations. We took a look at the CPU pricing benefit that Intel has, that AMD has benefited from.

19:13And we thought that, you know, expectations are low enough for a beat here based on the CPU shortage and higher pricing that comes with it. And so that's exactly what happened. The stocks surged, you know, they beat the consensus by far. They're guiding for Q2 sales to be, you know, 2 % to 9 % up sequentially to$13.8 to$14.8 billion, which is comfortably ahead of consensus. That's around 13.1 billion. And their gross margin surge for the quarter. Their margins expanded 310 basis points to 41%. And so while this looks great at face value, and the stock did react very favorably, I think there's two data points that need to be taken into account to digest this properly.

19:58The first is that part of the CPU shortage is actually internally created from the supply chain from Intel itself as it transitions from seven nanometer nodes to, you know, three, four and three nanometer nodes and skips those in between. So they're trying to compete with TSMC. So they're moving to smaller nodes. And because there was, you know, a lag and they trailed TSMC for a while, they're having to skip ahead. And with skipping ahead, that comes with a shortage on the older CPUs, right, the less advanced CPUs and higher pricing on the more advanced CPUs. And so that setup created a shortage that was, you know, from the supply chain rather than a shortage that's created because of overwhelming end demand.

20:39So I think that's really important to get out there because it speaks to how this might continue over the next couple of quarters and the momentum that we can see there. Because that, you know, that CPU shortage, that pricing benefit is not going to sustain forever. Intel has said for this year that they're aiming to raise their CPU prices every quarter, which is, you know, crazy because that really means that the shortage will continue. But at some point, I think expectations catch up and that becomes a riskier setup. And so that's on AMD. And of course, the same idea of proof points for agentic AI need to, you know, need to become more tangible for us to see Intel really benefit from that or the x86 market in specific to benefit from that.

21:19And then, you know, on the CPU kind of narrative, if we look at AMD and we look at Intel and we look at the CPU tailwind, neither is actually our favorite for the CPU tailwind. That's when I would actually tell you, no, I'd prefer ARM. Because when it comes to ARM, ARM has already kind of cozied up with a lot of the hyperscalers using its models, using its designs. ARM-based CPUs are used by Google, by AWS. and I think meta uses a dual source across AMD and ARM. But when it comes to this ASIC buildup, there's a lot more dependency on ARM than there is on AMD or on Intel. And I think that as ARM comes out with its first in-house CPU, that's going to be a huge tailwind for them because they're already the cost-efficient pick for a lot of these hyperscalers that are leading the buildup whose CapEx dollars are really going to be the dollars that matter depending on who wins them.

22:15And I think that ARM is better positioned to kind of ride the CPU renaissance moment than at the moment Intel or AMD. Fittingly, the story that I was reading before I got on with you was that UBS just said AMD and ARM gain server CPU share at Intel's expense in Q1. So much to that point, you guys are in agreement there. I just want to add something about Intel. I know I go on about the semis quite a bit, but for Intel, I think the real conversation we need to be having is around the Wall Street Journal headline from last Friday about Intel and Apple arriving at a preliminary agreement for Apple to use some of its chips and its devices, some of Intel's chips and its devices.

22:59and Intel surged from there. Because it's interesting because we had this report come out. We didn't get a comment from either Intel or Apple and the stock still surged and the market is pricing this in as a reality. But there's a lot of gaps in the viability of this deal in itself that I think are really important to address. So when it comes to Intel's, you know, foundry business, a couple of things to point out. The first is that what most investors don't realize is that if you add up Intel's client computing group and their data center and AI group, you arrive at Intel's basically total revenue, right?

23:37The Intel foundry sales, which increased 20 % sequentially this quarter to 5.42 billion, aren't really added into the top line. And that's because essentially Intel is its own customer and its only customer as far as the foundry business goes. And so what we've said about Intel before is that we're going to turn very bullish on Intel once we know that they can, once they secure an external customer for their foundry. Prior to that, last year, when it was sub 20 levels, we said Intel's a no brainer because at around$89 billion market cap, it doesn't make sense for Intel. It's undervalued. And then once the stock soared and specifically at the levels that we're seeing today, comfortably ahead of 100, I think the real conversation now what investors should be paying attention to is if Intel can actually secure that external customer, because we know from management that they're not going to commit CapEx spend on 14A until they've secured that.

24:32And even with the Musk-TaraFab kind of conversation going on prior to print, on the earnings call, we didn't have them commit any CapEx. So I think that's the data point to watch. Once we see Intel commit CapEx, that's when we know that they've actually secured that external customer. And Apple makes for a logical potential customer, at least a lot more logical than previous headlines we got around Intel that suggested that AMD may be a customer of Intel's. Because that kind of setup wouldn't make sense, right? At the end of the day, Intel would be supplying its competitor. If there were any supply constraints, Intel would prioritize its own over its competitors.

25:11And so it creates a very messy setup. So for Apple, it's not competing directly. It makes a lot more sense. At face value, it looks like a nice setup. but then if you look more closely there's a lot of question marks that we have around a potential Apple deal and the fact that it hasn't it's only been reported and not confirmed is also you know a huge red flag and I think the biggest question mark if I was to really boil it down it's around yield because as of now you know Apple makes everything with TSMC even the guys that Apple buy from make everything at TSMC like Qualcomm like Broadcom and so it would be a huge shift It wouldn't be just a one-off shift.

25:46It'd be a very big shift for Apple to decide to go with Intel on some of its devices. And for that shift to be warranted or to be justified as a move from Apple, I think we'd have to see really great yields from Intel's foundry business. And we know already on 18A that the yields weren't expected to reach industry standard until 2027. We still don't have a lot of color on what the yield looks like with 14A. But just logically for a foundry like Intel to supply a customer as large as Apple, which makes up, I think, around 24, 25 percent of TSMC's total sales, they would have to have very strong yields, which we haven't seen from them yet today.