Chip Stocks Are Crashing! Here’s the Truth About the AI Trade

I texted one of my friends who works at a large cap semiconductor company. He's the same friend who made seven figures off micron calls this year. And I said, "What are you thinking?" He says, "Quote, I don't know, man. I feel the problem is everything is too bullish and semi stocks are still not going up." I said, "I agree. It's a weird spot." He said, "It literally can't get more bullish than this." So his argument is if the stocks aren't reacting to what are the most bullish capex prints, earnings prints, free cash flow, EPS, buybacks, etc., and we're still depressed, is it a function of these companies genuinely have topped and you're just not seeing so you're not going to see the semi-names and when I mean companies, I mean semis obviously Google, Microsoft AS, right? These are long long long-term investments. But these picks and shovels bottleneck trades, have they legit topped to the point where AOI is not going to see 250 again like it did in June? Or is this once again sentiment, geopolitics, bond yields, Iran war, everything's red, so everything's red and there will be a chance for these names to meaningfully recover. And and and if it's the the the the former over the latter, that's an issue because you're not buying a bottleneck trade for the long term. I will say this again and again again no one here I think is really buying Micron for the next decade everyone's buying it for the next 20% up 30% up and if everyone thinks that the bottlenecks are now not going to be something they can rely on for even the next 20 30%. Then what's the point of buying the bottom? What? Like I'd rather buy Amazon or Google. I mean, I'd rather buy Meta down 30% from earnings than take a chance on a photonix play even though I know photonics and optics are so important for data centers. But how can I trust the sustainability cyclicality and ultimately the uh sentiment around these types of companies when effectively their entire livelihood will be destroyed once the capex does come to an end. So, so I mean how are people thinking about that? Is it sentiment or is it genuinely we have the most bullish news on semis and them not pumping means that they might have topped out in the short term? >> Yes and no. We have bullish reports. I don't think anyone can make the case that we've seen bullish sentiment towards these names. like it doesn't seem very bullish to me when when you have this semi sector that is trading at a low 20 times low 20 times earnings maybe even lower than that right now forward earnings so I I don't know how much bullish sentiment there is there are definitely bullish comments from a lot of these companies which again business-wise that's very good that means you're buying an undervalued company not everything's undervalued but on average across the board the semi trade seems to be much cheaper today um than than a couple of months ago As for the bottleneck trade, you you talked about one specific the AI AOI1. Does it need to go to 250? No, it's at 80 or 70 something. So even if it goes to 150, you've doubled. And like you said, you take the bottleneck names because there is a lot of growth ahead in the short run. You make that trade and you're done. So it doesn't go back to all time highs. >> So you're admitting these are all trades. These are not investments. bottleneck. >> The bottleneck on I've said it time and time again. You're not investing in bottlenecks for the long run. You're investing bottlenecks for oh this is going to go up substantially because the business is exploding. Once it does, you're done. Goodbye. Thank you very much. And you move to something that's way more stable for the long run, which I've started to notice that maybe Meta is the best business around. So I'm happy you said that because that means okay if you're saying the bottlenecks are trades which I would put micron in that category right >> I mean micron and and uh opto whatever ai is not the same like one is >> but is micron an investment >> billion dollar company >> no but still is is but still is micron an investment to you or is it a trade >> for me it wasn't it's it wasn't it still is an investment buying it today would still be an investment for Micron. It's a company that's going to generate hundreds of billions of dollars in revenue and probably free cash over the next couple of years as well. These small bottlenecks play are not coming close to that at all. No, not I mean not even close. So the reason I'm asking is because if these are traits that means people have to be good at timing. That's my point. And most people including myself suck at timing. that that like a lot of what we all preach is you find good companies fundamentally undervalued you buy them hold on for the long term because you know they they will appreciate over if you have enough years in this game you end up winning most people suck at trading including myself whether it's swing trading day trading like I've been trying to learn to day trade this year some wins some losses but it's not easy right so I I think it's important to highlight that if the if the bottlenecks are trades and I still think the memory stocks are not investments I really do think they're trades because eventually they cannot keep putting up these numbers And the numbers at least on Micron has been amazing. Three quarters in a row and the stock has not been able to make a move higher. Maybe there is something deeper than sentiment because you're right, we haven't had bullish sentiment even though we've had bullish results. But if bullish results can't lead to bullish sentiment, maybe it's the Iran war. We could blame that. But if it's not, then it's something about it's something else about where the streets wanting to rotate their funds right now. It's one of these things like in terms of the trade for Micron, it's like I I was thinking as you guys were answering and I'm glad you didn't bring it to me because I didn't know if it is a trade or an investment. Um, but the more I think about it, you you have to consider how much cash that these companies are generating. And there should be some sort of rule of thumb like if you can generate your market cap in terms of net income over the next 5 years or in free cash flow over the next uh 5 to seven years, at what point do you even say, okay, well, even though it's a bottleneck, this is still an investment because the buybacks are going to be so egregious. They're going to be so large. >> Yeah. And I think this comment also is uh Jen says, "I'm bag holding AOI at 158. It was supposed to be for a quick trade." And I don't even blame you for buying it because you it was down from 250, right? So you were like, "Okay, I'll get up." Now it's at 80. So to your point, Neil, if it goes to 160, it doubles. But like once again, you're timing something that has so many external factors here. And if it never was meant to be at that initial level, you know, it brings up a lot of questions on where funds are going to go in 2027 because once again, when the high beta trades go up, like Bloom could be up 15% tomorrow once sentiment changes and everyone Nebia, same thing. But is that sustainable? Like I think that's the biggest question. For the first six months of this year, it was every other day these names were up two to 3% and they were at much cheaper valuations given the companies were smaller. But now I feel like it's a different question. >> Yeah. But eventually they go up because the business has gone up like AMD is up substantially this year because we are expecting to see an acceleration in growth. So bottlenecks name they went up a lot based on what could happen. Then the market noticed oh actually we just made these companies worth a hell of a lot more money than what they should be worth. So we're going to pull them down until the growth gets realized. like a lot of these bottleneck names are still going to experience a lot of growth ahead because the bottleneck it it didn't get solved over the last two weeks. The the the business everything we still need all of it for next year and for the year after that they're all expanding so they can have more supply so they can sell more. So that's not going away. It's just the valuation didn't make sense maybe in June July got a correction now and now we're we're back at square one. We have an earning season. All of these companies have to now show look maybe we deserve to trade at a little premium or maybe we deserve to trade at at a lower valuation because we we cannot execute. >> But uh but but the the last thing you said is the point like maybe we deserve to trade at a little bit of a premium from here. No one's buying the bottleneck trades for a little bit of growth in terms of their stock price. They're buying it for this explosive like 3x 4x 5x at least a double. I mean do these names deserve to be up 100% from these levels? Does Marll deserve to be a $400 stock? I mean I mean you got a lot of people saying Marll should barely be in the 200s 220s even though they have massive growth but the valuation is kind of crazy. So like do you buy Marll here as an optics bottleneck? You know what I'm saying? Like and Marll is just one example of many names that on the green days are super green but they've not been able to hold their green. And I are we reading too much into the fact that it's been five months and these names have not been able to hold any of their greens since the Leopold liquidation or is there is the street saying we're shifting our strategy for 2027 and we can smell that the market you know sentiment around these bottleneck trades has changed and we're going to healthcare, we're going to financials, we're going to something else. I I I do think that uh you know people's sentiment is being uh decided via the price right now. Like I I saw a comment that was passing by and it said if Nvidia falls below 200, the party's over. And I'm like what what does that even mean? Like if we we we just saw uh Nvidia continue to fall on Amazon Google Microsoft all claim that capex is going to significantly increase in 2027 and that the stock price is going to dictate how much sales that Nvidia is going to make. Now, stocks are not like uh baseball cards or something like this that go up based on how much nostalgia people have for Nvidia. They go up based on how much cash flow that Nvidia can generate because if we don't buy it, they will give out uh returns to shareholders via buybacks and dividends and special dividends. Fundamentally speaking, these companies should be trading on a multiple. Now that multiple we can decide on uh you know how much we want to pay for for the future of of how much cash they're they're going to generate but who is actually bringing down their targets for earnings on Nvidia and if no one is and we're more bullish on Nvidia's earnings over time then my belief is how much of this selloff is actually being more towards jaded things like midterm volatility uh people getting worried about uh what the Fed is going to do, more more macro decisions, but not actually changing how much sales or or free cash flow that Nvidia is going to generate over time. There's certain bottlenecks that we can decide whether or not, you know, Micron's going to do well or or like even Neil said, being a little bit more choosy around AOI or um you know, very very high beta names. But >> um at the end of the day, like you said, >> go ahead. I was you're saying a fair point and this is why I think the Nvidia conversation is a bit more tilted to this being an investment. Nvidia has IP and the ability to have 75% gross margins that most companies can't replicate. Now every company is trying to compete. I'm trying to say Nvidia has a moat whereas the memory name >> Google and Amazon are selling off too and they've got moes. >> Well well no but that's my point. I don't think Nvidia is as much of a trade as the bottlenecks because Nvidia controls effectively the entire ecosystem of bottlenecks. Um, and Nvidia, they have a moat that can't be replicated easily even though they've got competitors. Micron's moat is just like timing. If Micron built 10x more fabs three years ago, they wouldn't they wouldn't have this pricing power. Even Sam Alman on the Cheeky Pine today said, "We think Micron can," he literally said, "We think the memory names can build more capacity, but they're so scared of the previous cycles where they overbuilt and then their stocks got destroyed that they're being a lot more cautious." Nvidia, it's not about capacity. Nvidia, >> that's he can say this, but this is exactly what TSMC does. And I don't think anybody's blaming TSMC of not overbuilding. Like it's it's about managing your business. Like TSMC knows there is a lot of demand. It's not why they're going to build like 10 fabs and then suddenly nine out of the 10 >> No, but but but TSMC is different. There's a lot of memory makers. There's not that many TSMC companm makers at all. >> That's fair, but there's more than one, right? Like you do have >> three for the whole Yeah, there's three that control most of the pricing power. I mean, if they wanted to build more, they could. I mean, I don't I don't think he's saying anything wrong. Like, yeah, Micron and S and and and and Samsung and SKIX are doing what they need to do for the interest of their shareholders and for years they've gotten screwed from big tech. So, they should be doing this, but they know that at the end of the day, if they overbuilt, they're done. They're cooked. There's no there's no thesis anymore. There's no pricing. >> No, I agree. So, the the and the long-term agreements are super bullish. But that goes back to the same conversation. if their stocks are still stuck and it's been five months now then is there a deep is it just a ron war as Tanner said which I think there's a world in which that's the case or is it the market knows these memory names are going to top out at some point in regards to the ability for them to grow earnings and that's being discounted a year a year in advance >> it's a I guess it's a mix of of all of it but let me just go back to the Marll example Marll is still up 144% year to date applied up to electronics is up 127% year to date. So these names have more than doubled year to date. So it's completely fine if for the last four or five months they are doing absolutely nothing. Nvidia did absolutely nothing for 7 months until it went up by 15% uh 20% and then got back down. So stocks go up a lot in the short run and then they do nothing and people freak out. Oh, it's completely over. But no, it's just the business just needs to catch up to the stock. >> We are lucky that the stock run ran ahead of the business, but logically the business needs to be ahead and then the stock is just >> I think that's missing the point of what most investors are freaking out about. Most people did not buy before they 4x or 3x for the year. Everyone bought closer to the top, >> you know. So that's why everyone's asking can they get back to the top and if they don't then they're screwed, right? that like Amazon will get back to being a $280 stock one day. I think we can agree. >> Will a will these other bottlenecks actually get back to those June highs? I >> I'm still invested in Micron, so my belief is yes, they will break new highs. But I think that this idea that every day needs to be a slow, gradual green day. And you know, if they're expecting growth into the future, then that's just going to be linear. There, this has happened multiple times where they've sat in that sort of 350 450 range for a long time. Now, uh, you know, now everyone wants to claim like there's not been an opportunity to buy in at those prices now that we've 3xed and now we're sitting around this $1,000 to $900 range or $850 range and it's hovering here. And it's like there's been plenty of opportunities to to buy in. Maybe everyone did buy up at the $1,200 range and now we're sitting much lower than that right now. But if you do believe that the they're signing deals all the way out to 2035, they should be able to generate enough cash to buy back, like I said, huge percentages of their outstanding shares. And if that happens, the market cap doesn't even need to go up for the share price to go up. And and investors will be handsomely rewarded. >> I agree. I just I don't think anyone's asking for slow green days. I think people are >> everyone's asking for for the market. We we went into this year we went into this year going, "Okay, if we look back in in history, midterm election years have been brutal every time." And then everyone is ripping their hair out. Looks like I've been ripping my hair out. Someone pointed out that I'm going bold. I'm looking at this. I've unreal. But um I think like this is exactly what you would expect from a midterm year. And now that it's happening, people are going, "Why is this happening? this must be something around the AI trade. And it's like maybe that's not true. Maybe the amount of spending that's happening in memory and for the chip layer, for for optics, for all of the component layers are going as well or better than expected across all Wall Street expectations, right? Who who forecasted out memory long-term agreements going to 2035? Tell me who forecasted longer than that. And then they're disappointed by these contracts. not one person like so >> yeah I mean you might you might it might be that simple right the point of this discussion is not to have a definitive you know black or white I think that that's why we do this analysis the point is it might be as simple as the midterms you get through the midterms and you know we we 2018 was ugly midterms 22 was ugly midterms 2019 and 2023 were massive up years and if we do have a massive green year bloom and sandisk and ai and like the most rocket lab high beta names are going to go up in that environment And so usually buying the dip on that makes sense. Uh I think the broader question here is if the large caps are getting destroyed and the semis are not benefiting from them getting destroyed which is what happened for the majority of the year. What does that look like in like do we think there's a world where Micron can get to 2 three trillion and Amazon and Google are barely holding above three trillion. >> Does that make okay so then Amazon and Google have to do something which means this dip has to get bought on them for the dip to get bought on Micron. And if we're in that stagnation phase, the semis continue to stagnate as well. Even though all the money is going from Amazon, Google to the semis, which is the irony of this. The the GPU companies will end up pricing based on the profit that they can make after paying for for the expensive chips and the expensive memory. So they are GPU prices are going up. You know, SpaceX is now charging like $50 million per megawatt, but Enthropic is making like a hundred million per megawatt. And then the customers who are buying from Enthropic are making like 234 $500 per megawatt. So I I don't think that there's going to be one specific company. Maybe that's memory right now just because the huge margin, but that won't last forever. Um margins will come down and margins can come down healthfully while they still have a business. I I think people are pricing too much of either memory becomes the biggest part of the chain or they go to zero. uh I I don't think enough people are maybe suggesting that their gross margin falls to 50% or something like this and the suppliers are very happy to or the chip names are very happy to purchase at those levels and they still have a business uh rather than the the tight constraints that we have right now. But because we looked back and they they went to negative gross margin, that's what everyone expects them to do. Just a a super cycle of what's happened previously. >> Yeah. Yeah. And look, I mean, if these amazing amazing companies like Amazon and Google are going to give discounts going into 27, which theoretically could get bought after the midterms if that's when we bought them. Um, market sentiment can change very quickly. That's one thing you're 1,000% correct on. Sentiment can flip so quick. I mean, sentiment Bloom was at 190 after earnings and it was at 2. I know it was down 5% after, but they got to 230 in like two days after earnings. Like, sentiment flipped on Bloom by $40 within seconds. Now, the choppiness is probably annoying to a lot of people who bought into the strength for Bloom, but like the point is when sentiment switches, it switches aggressively. And every time we've had a hint of sentiment switching over the past two weeks, it's been denied, whether it's because of Meta or SpaceX or yesterday, Google, today Amazon. And I think everyone's waiting for when that sentiment actually flips to like be in a trend. And maybe we just really need to get through a bunch of clearing events like midterms in the Iran war before we get there. Like for example, sentiment right now in the chat. Who thinks AMD goes up after earnings? Like seeing the Intel 20% drop? Who would be comfortable buying AMD here at >> e even I've never had, you know, the most bullish commentary about AMD. I've had worse commentary about Intel. So >> would you buy would you buy it at 530 before earnings? >> No, I would I I'm an Nvidia guy through and through. But um >> I was the reason the reason I'm asking is the reason I'm asking Tanner is because it's down from 680. So it's already taken a very big beating and you're still saying no. And I know you don't love AMD, but the point is you know the sentiment there is obviously a bit more negative even after it's taken a massive dip. >> Yeah, but just because the stock went to 650 or whatever and now it's back down to 530 and I own AMD. I've taken profits at higher prices, but still I own it. It doesn't mean that oh, suddenly it's it's cheap. Like I I'll say something that might be crazy, but when you overpay for a stock, don't be surprised when it goes down. It's uh I know it's it's new for for some, but if you overpay for a name, doesn't mean that the name needs to go by back up. If a name goes to 250 >> and now it's at 100, doesn't mean the stock needs to go to 250. means that maybe the stock should have never been at 250. >> But that goes back to my point around sentiment. I'm not saying AMD is cheap. Obviously, Nvidia Broadcom are incredibly cheaper compared to AMD. But if it went that high and it never deserved to get there, that was euphoric, which means who knows if it deserves to be here. So, the euphoric sentiment got crashed. Obviously, Intel 147 down to 97, then down to 79 after hours, meaning a lot of that price action on the way up wasn't real. It was just people gambling at that point thinking it would go higher. It wasn't supported by anything fundamental. And people hate to hear that if you're long Intel or long AMD or anything, but the point is if the numbers are the numbers and it's showing that it wasn't real and then the sentiment on top of that sentiment gets even worse, that means there's there's not a floor to these prices. What I'm trying to say, they need to find where buyers actually are. And maybe that's lower on a bunch of the bottlenecks. >> Yeah. But um Intel is I I just I just looked a bit at the transcript. They say that free cash of 27 there's probably not going to they're not going to generate any free cash flow. It's a bit different with AMD. AMD does generate free cash flow. AMD is also growing much faster etc etc. So >> plus they also gave us some midter midterm targets for AMD. So you can say oh it's trading at X times forward earnings for I don't can't remember what fiscal year they were using. So you can then start to play around with oh is it now getting cheaper or or not. So, so it it's I guess it it really depends on what businesses we're we're talking about cuz AMD is is going to experience an acceleration in growth. So, they're going to grow into the valuation that is yes higher than a Broadcom that is higher than an Nvidia an Intel. Yes, an Intel went up quite a lot purely because of one political stuff investments. It can still jump quite a lot purely because Foundry might announce a huge customer maybe at the end of this year, maybe starting next year because capex is going to go up and uh Libboutan did say that the only way that they are going to increase capex is if they are if they have concrete demand and it seems like if you're reading between the lines, the reason why they're increasing it is probably because there is concrete demand for for Intel Foundry. So maybe eventually this as well stays around 70 80 bucks for quite a while until the business actually catches up to to reality and then yes and then once the business did catch up then it can continue to go up and maybe go back to to all-time highs and things like that. But businesses that do not execute, that do not grow into valuations, then yes, they will never see all-time highs