Don't fear AI bubble, there will be winners - Tech Contrarians

0:09Always excited to welcome Sara Awad from Tech Contrarians back to the podcast. Welcome back to Investing Experts, Sarah. Thanks so much. Thanks for having us. It's always a great time chatting with you. It really is. I mean, the tech sector now more than ever is leaving people, and believe me, it's left them, as you know, with many questions up until now, but certainly more questions afoot every day in terms of, is AI a bubble? What's the proper valuation on these stocks? What about these deals that we keep seeing more and more of? You've been tremendously helpful for investors, for listeners, for observers in terms of sussing out and contextualizing what's happening in the tech space.

0:59I'm thrilled to have you back because there's a lot to talk about. I guess the place that I would like to start is with all this talk of an AI bubble, with all these questions around the sector and the long-term nature of how long the bullishness may or can last. What are your thoughts, generally speaking, when you're talking about tech, when you're sharing your insights with investors? What would be the 60-second summation that you would give them to start? That's a great question, spot on. I think that's what's on everyone's mind now, especially after we saw Oracle's earnings earlier this month.

1:38Essentially, if I had to summarize this briefly into a quick overview on how we're looking at things from what standpoint, I think things are pretty fishy at current levels. There's a lot of money moving around, especially this quarter from OpenAI committing money to Oracle, NVIDIA committing money to OpenAI, NVIDIA also committing money to buyout capacity from CoreWeave. It seems like we have you know, multi-billion deals being made all around. And that's great, right? That's bullish. Markets are reacting really well to that. But at the same time, it seems like people are getting really excited without thinking about what's next.

2:15Can we sustain this kind of, you know, this kind of level of spending? Can we sustain this kind of level of, you know, CapEx from the tier one players, Meta, Microsoft, Amazon, and Google? And what next year is going to look like, especially that, you know, there's still a gap when it comes to spend on AI and AI infrastructure versus actual returns on that for, you know, the larger tech sector and even beyond that. So I think that's kind of the standpoint that we're looking at tech through. I think that it's time to have a critical eye to things. There was a lot of, you know, panic around the potential AI bubble after Sam Altman's commentary, after, you know, that MIT report that found that I think something like more than 90 percent of companies trying to spend on AI are really being able to leverage it into growth and see that in terms of, you know, revenue growth and sustainable revenue growth at that.

3:11So there's a lot of questions. I think now we're at a point where there's more questions than we have answers. And it's really dangerous to get carried away with all this news without really thinking about it critically and thinking about how long this can last. And if we're actually headed to a bubble, the more good news that we see. What would you say are maybe the top one to three questions investors should be asking themselves right now about the tech sector? The top question I would say is that where is the money going? and if you trace the money especially when it comes to you know the money that's concentrated with these top big tech players there's something that looks fundamentally wrong so the top question would be like where is this money coming from and is the money that you know these guys are committing are they actually going to be able to to act on it so for example take you know oracle and and the their earnings report right they missed on top and bottom lines and market didn't seem to mind because they had the RPO that jumped something like 359 % year over year to 455 billion.

4:13And it's crazy to think that Oracle jumped something like over 30 % in one day, I think even hit a high of 40 % on that earnings report. But at the same time, if you look just, you know, one quarter prior, they actually, their RPO then was around, if I'm not mistaken, 138 billion. So if you really think about it critically, you know that Oracle suddenly reported this jump in RPO, and you know that they had a$300 billion deal from OpenAI. So the market's really getting excited about something that they think is a sustainable kind of spend in terms of the broader AI narrative, but it potentially isn't because that number only jumped because of that one deal with OpenAI.

4:54And then when you think back to OpenAI, we don't have too much information in terms of the finances that go into OpenAI, because, of course, it's not publicly traded. But what we do know is that the number is something like 700 million active weekly users, and we don't know how many of those are paid. But regardless, committing 300 billion is a very big number. and there's a lot of question marks around how sustainable this is especially now that we have this China alternative to compare to and this deep seek kind of alternative to compare to that's you know also implementing and growing and releasing new models but not with the same amount of spend so there's a question mark about how much of this spend is really needed and how much of this spend is going to be able to actually how much of these dollars committed are actually going to be spent at the end of the day.

5:44So I think people are really getting ahead of themselves in terms of seeing positives or, you know, wearing, let's say, like heart-shaped sunglasses on the situation. It's time to take the sunglasses off because we're entering, in our opinion, risky territory, especially that there's a lot that's not resolved in terms of, you know, the China questions and the broader kind of trends at play. So that's if we just take a look at Oracle and OpenAI, and then if you take a step back and look at the other deals that were happening this quarter. You also have NVIDIA committing$100 billion to build AI infrastructure for open AI, which is also, you know, markets love that.

6:23They reacted very positively. But there's a lot of money going around and there's a lot of question marks about how sustainable this spend is going to be. And the use of this infrastructure that's being spent, sorry, this infrastructure that's being built, if we still are in the very early innings of AI and we still don't have, let's say, a good command about where things are headed. You know, I revisit what we spoke about, I think, in a prior podcast, which is that Apple research paper on AI that said that the more complex inferencing workloads are, the models are actually collapsing on themselves.

6:58So if we're not able to have, you know, clean execution on higher complexity models, then how are we able to justify such a high spend for low to mid-complexity inferencing, right? So that's kind of the question I think that people should be looking at. You know, take the news, but also ground it in logic, because at the end of the day, logic is really what's going to show where AI is headed next, where tech is headed next, where potentially we might see this bubble deflate. I wouldn't say pop, because at the end of the day, we still are in the very early innings of AI, but we might see a bit of a deflation or a correction, because things are really getting ahead of themselves and moving very fast.

7:35Are you of the opinion that the dot-com era is worth a look back in terms of teaching us some lessons about how to avoid being on the wrong side of the switch, whether it's a deflation or a slow deflation or a big pop? Like some people think, are there lessons to be learned or is this time its own thing? I don't think there's any right or wrong answer as of now for this. But what I would say is I'm pretty confident that currently we're in uncharted territory. Right now, the combined weight of the top five companies, I think, in the S &P 500 is almost at 30%. So that's higher than in the late 60s, early 70s.

8:18And it's also higher than what we saw in the dot-com bubble. So there's really something that I would say is a red flag about how things are playing out, especially when you take that and you put it side by side or pair it with the fundamentals that we talked about in terms of Oracle and the sudden jump in RPO and how markets are reacting to things. It doesn't seem that people are really reacting to fundamentals. It seems like they're reacting more to what they want to see and not what is really happening. So I do think there are lessons to take from that dot com bubble and apply here in terms of being more wary.

8:51The fear, however, is that people are actually being very avoidant of drawing that comparison when it comes to practice and kind of just thinking, OK, markets may keep going higher. I think we're at the fifth month in a row of closing the month higher. so I don't blame those people because everyone at the end wants to make returns but it'll become increasingly risky to keep this up uh down the line because we're at a very fragile point where any slip-up could really cause the kind of reaction especially from you know the retail investor that's you know press sell now and then ask questions later uh so it's I think it's fundamental that if the dot com bubble really taught us anything it's about asking questions now before taking action and understanding really who are the guys that are going to stay on top and who are the losers in this play because not everyone is going to come out of this a winner you know we talked last time about Nvidia we also talked last time about Broadcom so Broadcom is one of those players that I'd say is going to be you know one of the top winners moving forward into 2026 but at the same time we're still pretty bearish when it comes to guys like Marvella or Nvidia, sorry, not Nvidia, Marvell or AMD, the little brother of Nvidia.

10:02So when it comes to guys like Marvell or AMD, I pair those guys in the same place because at the end of the day, they're kind of trailing behind if you really look at the performance and the fundamentals, and they don't have the positioning that Nvidia has in comparison to AMD or that Broadcom has in comparison to Marvell. So I think it's time to take a critical lens to things and really make sure that if you're going to put money into this AI play, you're putting money where the fundamentals do check out. And you're not just chasing this broader narrative that everyone has the potential to be the next Invidia, to be the next Broadcom.

10:36Yeah, I think context is definitely key. And I think to your point of people looking at things and seeing them how they want to see them, as opposed to how they are, is a very astute point and something I think we would all do well to be checking ourselves on. I mentioned at the start, and for those that have been following these conversations that you and I have been having together over the past year, your investing group on Seeking Alpha is called Tech Contrarians. You just mentioned Oracle as a bit of a contrarian take given how much it's risen lately and your more bearish take, let's say, if not bearish outright.

11:15Right. What would you say are your most contrarian takes right now? So I think Oracle's definitely makes it to that list. Aside from Oracle, we do have some contrarian takes. For example, another one that I just mentioned is AMD. We're buy rated on AMD just because we think the stock has come down so much. I think it returned to the 160s, even high 150s level, which makes it attractive just for the bounce back because people want that AI narrative. They're going to, you know, it's going to be a self-fulfilling prophecy. They're going to see it even if it's not there. So that's one of the contrarian takes where we don't really think the fundamentals support more upside there.

11:55But we do think the market will push more upside there. But we think those names like AMD, like Marvell, are not stocks that you want to be invested in when they report, because that's when really investor scrutiny and all these, you know, hopes and dreams of AI growth are really going to be put to the test. so that's you know we have um oracle amd morvel um core weave is a stock that we've you know been watching closely and this you know growing neocloud peer group so we've got core weave nebius um i ren is also becoming very you know in the public eye after they doubled their their gpu inventory from nvidia i think it was earlier this month if not last week so those guys are also very attractive in terms of you know the hype around them and the hype around them being the extension or the next generation of you know ai names to play at the same time for those stocks we think they're just trading stocks not stocks to be invested in in the longer term because as you know tech advances when we get the next generation the vera rubin from nvidia and the one after that since we have you know a four-year roadmap outline the tech is going to go um out of date it's going become outdated and then these guys are going to need to put a whole lot of upfront capex to be able to buy the next generation and then one after that to be able to keep up a higher ASP and keep you know the business running so this peer group is attractive to trade but I think our contrarian take there is that the business model in itself is fundamentally flawed when it comes to the world of tech right and you can see that because the guys for example that run and CoreWeave are more finance guys, not tech guys.

13:35And they're acting like finance guys with a whole, you know, renting GPU as a model without looking at the nit-gritty, which is that if you don't have the latest generation, you're not going to be able to rent it out as high, and you're going to be losing money. You're not going to be seeing the returns on the investment that you put there. And that kind of loops into the idea that NVIDIA is buying out CoreWeave's extra capacity. So we're seeing customers and NVIDIA, CoreWeave is supposed to be a customer of NVIDIA, but NVIDIA now is going to be a customer of CoreWeave. So we're seeing weird kind of dynamics and relationships forming.

14:10So if you take that kind of lens and you look at anything in the tech space, you start to see the cracks in the bigger AI narrative. Yeah, it's definitely interesting watching. I mean, nobody would accuse the tech sector of being a burgeoning sector, of course, but this whole idea of how AI is going to be, how it is already and how it will be incorporated into our lives is very much a developing story and one that has been very interesting. if you're just an observer to observe. Obviously, if you're invested, it's a lot more concerning at times. What would you say in terms of the China story, in terms of NVIDIA, in terms of these deals Intel has done the past couple of weeks?

14:58What would you say about all of that in context? That's the natural next place to go. I'm glad you're asking about that. So when it comes to China, I think China is becoming actually a bigger part of the AI conversation. And where I don't think in general, in the discussions happening around China, we're giving it enough credit in terms of the AI advancements, or let's say the technological advancements that are happening. And instead, China's consistently being discussed within the context of geopolitical risk, which is really overshadowing what is potentially, what potentially could be actually the biggest AI market, right?

15:35Because at the end of the day, you know, China is the biggest smartphone market. They're the biggest EV market. They're the biggest PC market. So there's a very big potential that there's a huge AI opportunity there. And we're kind of just thinking about this in terms of national security risk. And that's overshadowing the technological advancements that are happening. And I really think that's what Jensen is trying to shed a lot of light on. I think it was, you know, earlier this week that he said that China is, and I quote here, nanoseconds behind the U.S. when it comes to AI and that the U.S.

16:08really needs to compete and that all this geopolitical jargon is stopping NVIDIA from tapping into what is and could be an insanely big opportunity. I know that Jensen kind of put out an estimate there that it's a $50 billion opportunity. I think that's extremely like a modest estimate. It could be much bigger than that, considering the fact that, you know, China is the biggest market in terms of PC, smartphone, EV, etc. So when it comes to China, I think we really do have an alternative to what AI could look like and what the alternative to what AI spend could look like. And at the same time, we're seeing China kind of trying to move towards self-reliance with, you know, the latest news on this being Huawei and the new plans to prep a ramp in production for its most advanced AI chips because they're really trying to steal market share from Nvidia and, you know, kind of show the world that China doesn't need Nvidia, although the US is now going to allow these H20 chips to be sold.

17:05We don't need these H20 chips. And in fact, you know, the government is advising Alibaba, Tencent, all these players to stay clear from buying those because they have this new domestic alternative that should be able to compete. So I think that should be a bigger part of this broader narrative because what China is doing is important to contextualize U.S. spend and if that spend is misplaced or if it's actually going to pay off in terms of returns. So what we know is that essentially China, NVIDIA's market share in China is shrinking, right? They've been banned from selling these H20 ships for a while.

17:41And then even when that ban was reversed, they're still not able to get the products in. And that is going to, you know, leave a big gap, which Huawei is going to naturally fill, even if their products don't live up to the performance of NVIDIA's AI chips, and even if their products aren't going to be able to be produced at the same scale, right? Because now everyone's getting excited about this Hawaii story, but at the same time, in theory, it sounds great. In theory, we have this roadmap from them that's aimed at getting China to be self-sufficient. But the problem is really about production, how realistic this is going to look.

18:17So the question is that eventually China is going to need in the short term NVIDIA and the hardware and the software that it has. And if not even the software, at least the hardware, because on seven nanometer chips, it'll be hard, very hard. You know, and I'm cautious with the use of the world impossible, but quite impossible to compete with NVIDIA's next generation because they don't really have the hardware for it. And with, you know, the restrictions and the really tight grip that we have from the U.S. on China's access to tech, we're likely, in our opinion, this is one of our takes that I'd consider to be contrarian, to see China actually somehow come to a deal with the U.S., maybe one that coincides with the U.S.

18:57and China's trade deal, to allow the sale of NVIDIA's hardware back into China. and that that moment i think is coming and i just pay a lot more attention to china and that market and what's happening there because not only is it a huge market that nvidia that nvidia could tap into but it's also a huge market that shows us an alternative reality of what ai spend could look like whether that's from deep seek or you know alibaba committing a lot more in capex to ai but still being one of the lowest spenders when you compare it to the tier one players in the U.S. And what would you say about the deals that Intel's done recently?

19:36So Intel is actually one has been one of our favorite names. For those of you who check out our articles on Seeking Alpha, I've listened to this podcast before. We've been really preaching, screaming, buy Intel under 20 bucks. It's a no brainer. And the logic behind that at the time was that Intel deserves a larger market cap. than what it had and that it really had bottomed out. I think it was last year in 2024, August, that Intel took a nosedive and didn't break the$30 per share level sense. That is until, of course, we saw it break it earlier this month. And so when it comes to Intel, these deals that are coming through, I think are a direct result of Trump's involvement in Intel in this kind of effort to make US chip making great again, right?

20:28And to make Intel great again. And what we saw, this is a divergence from what we saw from Intel's own earnings report earlier this quarter, where, you know, Li Bhutan's tone regarding the foundry side of the business, at least, was pretty cautious. It was along the line, especially if you take a look at the 10Q, I think it's on page 26, that they very clearly disclaimed that they're not going to be moving forward with their foundry plans unless they can guarantee and lock in the demand just because it's been such a high cash burn and such a weight on the company itself. So we're seeing Intel close all these deals.

21:02We have a stake from the US government. We have a stake from SoftBank. We also have a stake from Nvidia, you know, what people are calling now the Nvidia effect, which really moved the stock. And now we have, you know, Intel looking to get some money as well from Apple. So all of this is great to kind of boost optimism when it comes to Intel, but the fundamental story, which is the Foundry side of the business is still has a huge question mark around it. So while we've been preaching buy Intel under 20 bucks, we think that at the moment, it's the question people should be asking themselves, whether invested in Intel or not, is what's gonna happen to the Foundry side of the business.

21:38There we see two scenarios playing out. The first is, you know, Intel doubling down on IFS and really going for the 18A and the 14A after that, especially that they're already committed and bought, you know, ASML's high NA tools, which are the most expensive tools. You know, TSMC still hasn't bought those or committed to buying those. And that would only happen if we get external customers, which is why to us, you know, the deal with NVIDIA was a big positive. but it was also about what was missing from that deal, which was NVIDIA committing to being a Foundry customer of Intel, even if in small part.

22:14And we didn't get that. The fact that we didn't get that and we only got NVIDIA put in, I think it's around a minor stake compared to the amount of money that NVIDIA generates, it seems more of a show for Trump, in our opinion, or a show to spice up the competitive environment, especially as it relates to ARM and AMD, than an actual commitment to kind of bring Intel's Foundry to its full potential. So that's the first case scenario that we see Intel's Foundry really get external customers. And then the second case scenario is that we see it spun off. And I think in either one of those case scenarios, Intel stock is going to move up from there.

22:55But there is uncertainty around which will happen. And the stock now is moving up on these investments, but we still don't have a guarantee about how they're going to show up on the financials. And the guarantee about how financials can get healthier is getting an answer about what's going to happen to foundries. I appreciate that. We saw a report, a denial of a report from The Wall Street Journal this morning. I don't know if you've if you've already seen it about Taiwan Semiconductor saying that they're not doing a deal with Intel. Are these the kinds of headlines, the kind of yes, maybe there's a deal in the works?

23:27no, there's not a deal. Are these the kind of headlines we're going to see in the next meanwhile, while Intel kind of does its media tour? Not really, but surreptitiously? It feels a bit funny to see Intel approach so many of the big players and kind of try to leverage capital. I think the number that was being floated around is that Intel needs at least 50 billion to really get into its next phase of growth. So even with all this money that's been committed, We're still, you know, we're still looking at something that's much lower than where we need to be for Intel stock to really hit that next level of growth.

24:05And when it comes to TSMC, I think TSMC is a big factor of this conversation, especially as it relates to the to the foundry side. And it's like that we're going to keep seeing these kinds of headlines, because for TSMC, if Intel really does, you know, walk the talk when it comes to the foundry side of the business, then TSMC actually has competition, which is something we haven't seen TSMC have to deal with really seriously up until now. So if I was to, you know, paint a picture here, if Intel does take that option, the first option that we outlined and is able to get an external customer to, you know, take a bet on their foundry, then TSMC would actually have to buy ASML's high NA tools, right?

24:47Which is, if I'm not mistaken, something in the range of like 400 million euros. So TSMC would have to do that because they'd have to keep up with the technological advancements. As of now, with Intel kind of still, you know, flopping around in terms of what they're going to do next on the foundry, TSMC hasn't needed to commit that kind of money. So this is actually what happens with Intel will be a ripple effect within the broader sector of tech, which is why we're seeing so many, let's say, tense and like urgent headlines around it. because if Intel commits to the foundry, TSMC has to buy high NA tools.

25:21And that's a massive tailwind for ASML, which has really been going through it in terms of, you know, what's going to happen after they said they're not sure if they can guide for growth in 2026. So there's going to be a really big ripple effect from Intel, ironically, considering that it was a player that was kind of more under the radars or in the shadows for the greater part of this year. Although the stock actually, if you take a look at the stock, you know, quarter to date, And in Q1, Q2, and even in Q3, it's been one of the best performing stocks if you're trading it, you know, buy below 20 and then sell when you're comfortable above that and then hold into Q4, which is what our original thesis there was.

26:00TSM also, by the way, one of six tech stocks recently highlighted by Steve Kress, our head of Quant and Data. He recently highlighted it as one of six strong quant buys. For those interested, I'll leave a link to that article in our show notes. But anything that surprised you recently, whether it be in the earning season out of the conference calls or what you've seen from these development deals or companies pivoting? I think the biggest surprise, and I think a lot of people share this, and it's one of the surprises, there's a couple more to discuss, is how much involvement we're seeing from, you know, the Trump administration within tech.

26:40I think, especially when it comes to, you know, diplomatic talks and what we saw in the UK, in terms of more investments, it seems that the government and the Trump administration is very closely tied to the trajectory of tech and the trajectory in specific of AI. So we're seeing these big tech players not only consolidate in terms of, you know, the being a large weight within the S &P 500, but we're also seeing them play a bigger role when it comes to trade talks, especially in how this relates to China. so and especially how this relates to China and and the the next thing that comes up there is really the TikTok deal so I think the TikTok deal and and seeing that resolve this year is something that was expected more or less with how you know determined Trump was to to get a deal there but what wasn't as expected from our end is that Oracle is going to be such a big part of that conversation and that we're going to see this new you know spin-off in which Oracle is going to be working on an American version of TikTok's algorithm.

27:41And I think that's going to have interesting ripple effects into Snapchat, into Meta as well, and into where advertisers choose to put their dollars spent. So this is actually the impact when we look at it through the lens of ads could even impact Amazon. It could even impact Google. So that's one thing that I think is interesting. A lot of people were very excited about that news, but I think that the ripple effect hasn't gotten enough attention because if we have an American, which we will have an American version of TikTok, there's going to be a ripple effect in terms of advertisements that needs to be taken really seriously.

28:13Before you go on to the second one, can you explain a little bit in more detail about what those ripple effects may be? Yeah, of course. So, you know, if you look at Google's earnings, if you look at Meta's earnings, a lot of the bulk of the revenue really comes from advertisement. And this year, earlier this year, we had a lot of panic around you know if the economy if the situation worsens then we're not going to have as much ad dollars spent uh especially if there's no unfavorable trade agreement with china a lot of the the products advertised do come from china and it'll be harder to have you know the capital to advertise these and so if we have a new player that's already very well established coming into into the the u.s um and coming into the u.s kind of permanently which hasn't been the case for TikTok, we could see the behavior from ad spend differ and maybe circulate more towards TikTok over Meta, over Snap, over Google.