Chit Chat Stocks - The State of The Semiconductor Market With Chip Stock Investor (NVDA, TSM, ASML, INTC, AVGO, and More!)

Starting point is 00:00:00 Welcome to Chit Chat Stocks. We have a fantastic episode coming up, but first, let's talk about our presenting sponsor, Public.com. Heads up, folks, interest rates are falling, but you can still lock in a 6% or higher yield with a bond account at Public.com. That's a pretty big deal, because when rates drop, so can the interest you earn on your investment. A bond account allows you to lock in a 6% or higher yield with a diversified portfolio of high yield and investment-grade corporate bonds. So, while other people are watching their returns shrink, you can sit back with regular interest payments. But you might want to act fast because your yield is not locked in until you invest. The good news, it only takes a

Starting point is 00:00:39 couple of minutes to sign up at public.com. Lock in a 6% or higher yield with a bond account only at public.com forward slash chitchat stocks. Brought to you by Public Investing, member of FINRA plus SIPC. As of September 26, 2024, the average annualized yield to worst across the bond account is greater than 6%. Yield to worst is not guaranteed. Not an investment recommendation. All investing involves risk. Visit public.com slash disclosure slash bond dash account for more info. Welcome to Chit Chat Stocks. On this show, host Ryan Henderson and Brett Schaefer analyze businesses and riff on the world of investing. As a quick reminder, Chit Chat Stocks is a CCM Media Group podcast. Anything discussed on Chitchat Stocks by Ryan, Brett, or any other

Starting point is 00:01:28 podcast guest is not formal advice or recommendation. Now, please enjoy this episode. Welcome to the Chitchat Stocks podcast. Today, we are joined by first-time guests, I believe, friends of the show and hosts of the ChipStock Investor YouTube channel and podcast. So if you ever go over to YouTube and you're interested in any ChipStock whatsoever, I guarantee you've come across their page or you should at least go ahead and check it out. It is Nick and Casey Rosalillo. Welcome to the show, guys. Thanks, guys. Thank you. Yeah. Good to be here. sweet we are uh we're going over a lot of things all all things chip stocks and semiconductors but i'm going to give you sort of an impossible question because i know it's been a busy few

Starting point is 00:02:22 years can you give us me and listeners a brief timeline of what has happened over the last few years in the semiconductor market i'm thinking about back in like as we were coming out of COVID, kind of that resurgence, all the talk was we have semiconductor companies have overinvested, we're going to see a down cycle. And it seems like maybe there's been that for some companies. So I guess, has that down cycle materialized at all? I don't know. Give us the thumbnails of the last few years. You said brief overview. So Casey's going to handle that part. yeah we'll keep we'll try to keep this brief because you're right a lot of things have happened in the last few years but one thing that we've focused on is that we really have

Starting point is 00:03:12 been in a down cycle since 2022 especially in enterprise computing and consumer electronics but the auto industry which is kind of what kicked off this chip shortage mantra you could say was really in 2021 and 2022. And that fell off a cliff starting in Q2, 2023, which is almost about a year ago now, but it doesn't really seem like, sorry, Brett. No, I was going to say, I was going to say, so everything outside of AI or almost everything outside of AI, if you exclude that it would have been a normal down cycle is almost what listeners should be understanding. Yeah, exactly. And it, it really doesn't feel like it's a downturn because of all of this hype towards accelerated compute and AI, aka

Starting point is 00:04:00 NVIDIA. All right. Beautiful, beautiful. And I know almost every listener has heard about AI. They've heard about NVIDIA. I mean, it's probably, I think took over from Tesla as the most followed or maybe Palantir in there as well as so many fans of it among individual investors. What do you guys make? We're going to get into it because there's so many things to talk about within this sector what do you make of the surge in ai spending and who are perhaps you know there's there's the obvious beneficiary like nvidia and we can talk about them as well but who are also the less talked about beneficiaries like who there's tens of billions maybe hundreds of billions of dollars going into this who's getting helped out okay i like i like this question brett

Starting point is 00:04:54 But before we answer it, there was one part of the question that we've tried to reframe this for our listeners, because this is really, I think, where the media arm of the investment community has kind of taken this thing and run with it. And that's that this isn't a surge in AI spending. It's not at all what's happening. It's a surge in accelerated computing infrastructure And AI, specifically generative AI, that's what most people are referring to now when they say AI these days, is just one specific computing workload for accelerated computing. Accelerated computing can work for anything. It just speeds up whatever the software program is. It just runs it faster and more efficiently than non-accelerated computing.

Starting point is 00:05:49 CPUs, primarily made by Intel, that's still some way, somehow over 90% of enterprise computing and cloud CPU market share, Intel and AMD. And so accelerated computing is kind of just taking over that because it wasn't the generative AI bit that was the breakthrough per se. open AI's chat GPT kind of opened everybody's eyes as to what the future possibilities were for generative AI. But the real breakthrough actually happened before that. And it was simply NVIDIA just demonstrating that their GPUs could accelerate workloads and make them more efficient. I'll stop there because this kind of is starting to sound like we're in an, or I'm an

Starting point is 00:06:36 NVIDIA employee and that's not the case. More than happy to talk about flaws with the company later on but i think that's an important way to reframe this it's not an ai spending boom okay maybe i'll follow it up with casey and you can answer this one the accelerated compute is not necessarily new maybe not even new chips or new say workloads it's replacing intel and that's where they're losing that market share and say they went i think the number is from 80 billion dollars to 50 billion dollars and now that might not all be from this but you're saying that there's this rip and replace from the existing, you said 90% market share. Do you think, and here's my question, do you think that will continue even if the AI boom turns into a bust? Is this a long-term switch from CPUs to

Starting point is 00:07:23 GPUs? It certainly seems that way, doesn't it? Yeah, I think it's for a while, I guess is what I would say. Yeah. And it's probably more complex than just ripping out and replacing the old infrastructure with new, it's kind of being built side by side, because there's still a lot of useful life for the old CPU powered servers. So it's kind of, you know, this new thing is being built alongside the infrastructure that exists. But yeah, long term, it kind of seems like this will be the start of a long term rip out and replace cycle. And then also, in addition to that, it's not just ripping out the old stuff and replacing it with the new, but there are still massive amounts of computing workloads massive amounts of traditional data non-computing based

Starting point is 00:08:14 electronic based data migrating to enterprise computing or and or the cloud in some form or fashion so um and that kind of goes back to your question brett about who else benefits besides just nvidia just these gpu makers and that we talk a lot about the equipment makers that make these chips like we call it the fab five asml applied materials tokyo electron and kla core they make the manufacturing equipment that enables these these amazing computing systems that nvidia is making especially okay that's that's super helpful context we're going to play a bit of a true or false game here as well to kind of go through some of those fab five is Is there anyone we're missing, though, I guess, in terms of – it sounds like you just talked about both some of the fabricators or the fabrication facilities, the chip design companies like NVIDIA, the software companies.

Starting point is 00:09:19 I think Applied Materials is mostly software, if I'm not mistaken, or is that – am I getting that one wrong? They're mostly the machines themselves that make the chips, but they do have a large – about a quarter of their business, let's say, is software and service-related. And it's actually growing faster than the equipment of all of those companies. They have probably the most diverse set of equipment and software. Okay. Is there anyone that any company over the last few years that surprised you that did really well out of this,

Starting point is 00:09:52 that you would have never thought would be a beneficiary of this accelerated compute spending? Super microcomputer. That was the thing. Exactly. oh no we're not touching that I was going to say that was the surprise and that's

Starting point is 00:10:09 interesting company I'll say that it has been a surprise but not a good surprise I know exactly what she's thinking about we just talked about this we just talked about monolithic power systems that Nick passed on

Starting point is 00:10:26 yeah I don't blame him I don't hold it against him any longer but he chose on semiconductor versus monolithic power you know the dumb thing about it is like i've followed monolithic power for years and have talked about the unique business model and yeah it's just it's a unique business model because typically like power management chips like like they specialize in is an idm it's a company that both designs and manufactures the chips but in the 90s when they when the business

Starting point is 00:11:00 was was founded still a founder-led business 30 years later um the ceo decided to make it fabless they they they outsource all of their manufacturing which is unique for a power chip because those are generally pretty low value parts and pieces in an electronic system um you know we're talking maybe pennies per part at the at the low end very cheap but they went fabulous and so i think we did a video calling them like the NVIDIA of power chip stocks and took some flack for it being clickbait. But it's true. In a sea of like integrated manufacturers and designers of power chips, they stand apart. And that is a big factor as to why they have so greatly outperformed most of their peers, if not every single one of their peers. I think it's probably a pretty

Starting point is 00:11:55 safe statement okay what markets do they serve uh for anyone that's interested in yeah i am seeing a little check while you guys are watching here even this is excluding dividends something's i mean assuming the total return is over a hundred bag but it's eight thousand nine hundred percent as of uh as of the last trade so yeah hey been a lot good long-term performance yeah the last few years have been great but yeah what what do they serve what's the difference between them because as a as someone that doesn't know the industry very well outside of you know kind of the big names, what makes a power systems company different than what NVIDIA is serving? You want to grab that one?

Starting point is 00:12:31 No, go ahead. So they serve every market. If there's a device in your house that uses electricity, there's a power chip in it. If there's a cable that plugs into the wall or a cable that plugs into another electronic device, like a usbc cable there's actually a chip in there that manages the flow of electricity because every single part of an electronic system needs a very very specific amount of power if it gets too little power the device doesn't function the way it was designed if it gets too much power well then your device gets fried uh all the all the components get too much electricity

Starting point is 00:13:17 and they they get zapped the same as if you like stick your finger into the wall and you get zapped you're not supposed to have that much electricity flowing through you same with any electronic device and that's where the power chips come into play they regulate that in simple in simple terms and so when you're thinking about oh sorry no no go ahead casey just in in light of like accelerated compute their monolithic power is is shrinking down and making their power chips more efficient to leave space for the gpus and hbm3e high bandwidth memory and then they're also supplying power to not only in the server but to the server for these giant racks okay makes sense and it was monolithic power systems right that's yeah for anyone interested tickers npwr so

Starting point is 00:14:11 yeah i'm sure you guys have covered that on the channel quite a bit we're gonna play that true or false game i think this will be a fun one you mentioned the uh the equipment providers and one of the hottest names within that is asml with the euv stuff and i've heard time and time again people saying look no this is impossible to replicate uh china's trying to do it as a whole government project and they can't seem to get it solved yet. True or false? And I know this is just a podcast, so we're not going to hold any takes here. It's just a way to explore different opinions. No one can replicate ASML's technology. What do you guys think? What way do you lean? No one to me is a very bold statement. Never say never. But at this point, they're the

Starting point is 00:15:04 clear leader in lithography. Yeah. And it's possible. I agree. I agree with Casey. It's possible. However, you can make a really strong case at this point in time that even if it were possible to replicate what they're doing, you know, find a workaround to all of their extensive patents that they have on, especially EUV and some of the DUV machines as well, that kind of the previous gen lithography machines, even if you found workarounds for all of that, does it make financial sense to do it? And the answer is probably no. It doesn't make any financial sense because it took ASML decades and countless tens of billions of dollars of R&D to get to this point. And so what exactly would be the benefit of replicating that,

Starting point is 00:15:56 getting a machine ready, and then going out to a semiconductor manufacturer and saying, hey, look, I have a machine just like ASML. And they're going to say, well, is it better than ASML? Because I also get great customer service from ASML. I've been working with them for decades. And so I know exactly how to use their machines in the manufacturing process. So I think that's, yeah, I agree. No one is a bold statement to make. No one is going to catch up to asml is a bold statement because china obviously does have some incentive to try to home grow their own version of this stuff given that they've been restricted from accessing it but it would be very difficult outside of that market to bring the machine to a manufacturer

Starting point is 00:16:45 okay yeah right right and it's almost like you know from a science perspective maybe they can get there, but you're saying there's almost a higher switching cost because if I'm the manager of the plant or say a hypothetical manager of a plant, and I say, look, I want to switch out these machines and then it doesn't work. I mean, you might've screwed over the whole company. Is that kind of like how important these can be to these plants? That pretty much sums it up. Yeah. All right. Sounds like a potentially wide boat business. And I guess I'm surprising that it trades at 30 40 times earnings and has been such a great performer but ryan did you have another follow-up there too no sure let's move to this second true or false one here and this is

Starting point is 00:17:30 i don't know i mean i guess maybe a uh polarizing question because people tend to stand on like very firmly on both sides so true or false nvidia is a wide moat business am i taking this one yeah you go ahead. It's a tough question. It's one that, yeah, it's probably the hardest question we put in there. Okay. I'm going to, if I can. Nick has a really good analogy. So yeah, let's talk. Let's do that. I think it needs to be reframed here. Kind of like, you know, that the spending boom in AI needed to be reframed. It's not AI, it's accelerated computing. We need to stop talking about moats especially in technology seriously like this is something that has driven me nuts for many years um and you know this is just me i'm a person with strong opinions and i get

Starting point is 00:18:26 you know really well entrenched you can cut all this out if you want yeah it's anathema the buffett uh the buffett followers are uh you know they're writhing in pain right now but hey if you get the there's so many comments out there are so many followers of people that talk about say these technology companies and i i feel like the comeback every time is well does it have a moat like a very sassy comment so i feel like for you guys that's probably driven you insane over the last few years it's very cliche um moats are relevant primarily in industries that are mature, that don't move very fast anymore. You know, capitalism has done its work on the industry. It's squeezed out most of the

Starting point is 00:19:14 efficiencies already. And then you start looking in industries like that for the big companies or small companies, for that matter, that have some sort of defensible moat within that industry ecosystem. If you're talking about tech, which let's maybe just zoom out and let's say tech is not computing technology let's just say it's anything that's just simply fast moving and fast changing lots of development happening capitalism is still like doing its work to like find new efficiencies new outlets for growth why why do you want to moat for in a business like that okay so here's the analogy casey said i had an analogy because i have we've been thinking about this a lot as of late and mind wandering while reading a boring book recently. And I thought,

Starting point is 00:20:02 okay, instead of like a castle and moat analogy, this is fast moving technology. What we really actually need to talk about here is like a race. Who has the fastest car? When you're talking about a technology industry, you don't want to moat. What do you want to like your car to like, just go park it in the pit and be like we have a fantastic pit team we're parking it right here you're gonna lose the race okay i like that i like that yeah who has the fastest car and who has the best management team to make sure that car can stay in the lead and so mind was wandering about this and i was thinking actually what are the statistics for formula one everyone watches formula one now right uh thanks to or at least we all watch drive to survive on yeah that's that

Starting point is 00:20:53 put me in that camp i've watched more of that than perhaps the actual races me too me too at least americans talking here like yeah so i was curious like if you get your car in pole position what's the percentage that you win the race it varies from decade to decade and generation to generation depending on rule changes and the car technology, but it's basically anywhere from 30 to over 50% of the time you win. If you're in pole position, you, during the qualifiers, you have the fastest car. And so you get to be at the front of the starting grid.

Starting point is 00:21:30 If you get that starting grid position and there's 19 cars behind you, instead of having a 5% chance of winning the race, you know, 20 cars divided by 100% divided by 20 cars, 5% chance. It's not a simple statistic. In this case, if you're at the front of the line with the fastest car, you're going to win possibly more than 50% of the time. Do you really want a moat? Do you want to look for a company with a moat in that type of situation? Or do you want to look for the company with the fastest car and the best the best management team managing that driver all right so would you describe nvidia as being in pole position with the fastest car and the fastest driver

Starting point is 00:22:14 yeah i think so i mean well there's a reason they're one of the biggest companies in the world right yeah exactly okay that's actually wow yeah that makes perfect sense yeah okay so for anyone that's not following the analogies jensen wong would be the uh the driver uh leather jacket man and the car is their current technology and maybe or maybe even their like r&d team or whoever is you know keeping them in the lead with the software and the pole position is their existing you know the stuff the chips that have been i don't know far surpassed everyone else but the best out there for all this accelerated compute um so let me run yeah the other thing i was going to mention is you got 30 to 50 chance in this if you're the pole position you got 30 to 50 chance

Starting point is 00:23:07 of winning this race it's not like you're just winning one race the the upside is it's not you're not just whatever doubling your money or something like that the the winnings are potentially much higher because of the industry and the the applications so the other question i was going to have though and i'm trying to rephrase kind of the wide moat question here if you're a customer or a potential customer for nvidia's gpus what is the trade-off like like what's the alternative if if you don't want nvidia's gpus you want something cheaper is that a good idea i guess from your perspective is it just pay up for nvidia's gpus because i'm hearing there's the supply crunch and everyone's prices are soaring. It's becoming very costly to

Starting point is 00:23:52 integrate some of this stuff. So I guess, what are your thoughts? Is there any alternative? Yeah, there's for sure. There's alternatives. Yeah. Do you want to go first? Let me think about this. I think, yes. The answer is yes. There is alternatives to an NVIDIA Blackwell system, for instance, that'll be their new, the newest chip set that they'll, they'll be sending out. And some, some businesses, it makes sense, financial sense to go with another company, like maybe AMD or. Broadcom. Broadcom, you know, and, or even, I think we'll talk about this at some point about

Starting point is 00:24:36 making their own chips, using their own designs. And it may just make financial sense. However, some of these hyperscalers may want the best in the business at crunching data, at least in some applications. Yeah, and I think maybe I would just add to that because that's an important point. the hyperscalers, Microsoft, Azure, Amazon, AWS, Google Cloud, Meta, the big data center operators. That's who's primarily investing in this right now. The vast majority of NVIDIA's data center revenue is coming from that industry. And everyone talks about like cost being a concern. yeah it's always a concern these are these companies are cost conscious as well but when

Starting point is 00:25:34 you have you know 100 billion dollars in cash and investment sitting on your balance sheet and your google um are you like gonna play like a sob story and like i don't have the money to buy these things no your primary concern is i have deep pockets but so does microsoft so does amazon so does meta and you know they're kind of frenemies all these hyperscalers like they they're friends but they also compete with each other do you really want to be left behind technologically again remember this is not a moat type of business this is about who has the fastest car uh so do you want to settle for the second fastest car the third fastest car not if you can help it. So primarily at this point, this is a new market, more or less, accelerated computing,

Starting point is 00:26:27 generative AI, again, just one workload for accelerated computing is a new market. And you don't really want to just settle for second place if you can help it. So you invest the money. Later on, as the technology matures, then you can start having conversations with smaller companies that are most definitely very cost conscious after the technology has been ironed out. There's no more wrinkles. Most of the defects are gone. There's a very solid use case for how a small business can use it. Okay. Now we can say what's the best,

Starting point is 00:27:05 most cost efficient accelerated computing system for me, for chit chat money or for a chip stock investor, you know, small tiny little media businesses the hyperscalers it's a very different conversation and maybe you're not buying the blackwell system but you're buying the last the last iteration of this the h100 you know it's still better than the competitors makes sense and i'm trying to imagine myself as one of these uh hyperscalers google amazon uh i guess one of the famous customers for aws is Netflix. And if Amazon was saying, look, it's not affordable for us. We're just going to stop investing in these new chips. They cost too much. They're raising prices. Well, what you're saying

Starting point is 00:27:55 is anger their customer base. And they really don't want to do that right now, especially with these contracts that can be so, so large. And one more interesting but very important item that you mentioned here, Brett. So primarily up until this point, most of these accelerated computing investments have actually been going into the back end part of the data center, not the front end service that a company like Netflix accesses when they rent cloud computing from AWS. And actually, maybe I'll just remove AWS from the equation here for just a moment. Let's just compare Google and Meta to each other because they both make most of their money off of ads, digital ads. Most of what they're purchasing from NVIDIA is just going into the back end of their data center to help make the existing digital ads empire they're sitting on run more efficiently because it means more profit in subsequent years.

Starting point is 00:29:02 so that's like this is round one of the accelerated computing boom and that is where the vast majority of these chips have come from some of these like cutesy little services that have come out yeah maybe a few of them were like trained using nvidia accelerated compute systems like open ai's stuff even google gemini they've loved to trumpet the fact that their own custom chips did a lot of the training, fine. They're not super useful, really. They're not ready for mainstream commercialization yet, but they're actually using the best of the best themselves first. They're not going to share that. It's for their own operations. I think that is a very, very misunderstood part of this. These things are not being purchased and sold via the cloud.

Starting point is 00:29:50 the nvidia chips the ai systems whatever you want to call it that's not what's being sold to customers on the cloud not yet yeah you mentioned under followed uh that just reminded me of i think two years ago when meta was talking about their big investment cycle i guess i that i just remember they were talking about hey we got to buy all these chips for ai compute and yeah that's not for at the time no one was really talking about the generated ai stuff it was all for their ad server so it does make sense that leads to the next true or false game and you already hinted at this a bit but i want to hear your opinion because there could be as you mentioned two divergent paths that a company could go on uh they could say look we

Starting point is 00:30:34 don't we want to eventually get to a place where we replace nvidia and that could be a huge implications for the industry as you mentioned these cloud providers are the biggest out there So true or false, you think the hyperscalers are smart to invest in their own chip design programs. and the latest stock pitches on joinyellowbrick.com slash chitchat. We genuinely use Yellow Brick every single week here for stock research on small and micro cap ideas and think you can as well. Try it for yourself. Simply go to joinyellowbrick.com slash chitchat and search a company or ticker you're interested in. You are bound to find a great report on just about any company.

Starting point is 00:31:38 That's joinyellowbrick.com slash chitchat. heads up folks interest rates are falling but you can still lock in a six percent or higher yield with a diversified portfolio of high yield and investment grade corporate bonds on public.com you might want to act fast because your yield isn't locked in until the time of purchase lock in a six percent or higher yield with a bond account only at public.com forward slash chitchat stocks brought to you by public investing member finra plus sipc as of september 26 2024 the average annualized yield to worst across the bond account is greater than 6%. Yield to worst is not guaranteed, not an investment recommendation. All investing involves risk. Visit public.com

Starting point is 00:32:21 slash disclosure slash bond dash account for more info. Maybe, even probably. I think I would lean true more than false on this one myself, because if they can produce a good enough chip for all of their applications that they need why wouldn't they why wouldn't that be why wouldn't they use that cost-effective method and i think that's that speaks to companies like broadcom and marvell who are helping design these chips they're they're very large market share so is that like consulting for broadcom and marvell are they like there's going in and helping these companies are they Like, are Amazon or whoever's designing their own chips asking Broadcom to do it? That is a great question.

Starting point is 00:33:18 Let's see. Briefly. How do we address this briefly? It's not consulting. So Broadcom and Marvell own very large patent portfolios. And that's what makes them a great custom chip design house. If you know you're a hyperscaler like Google, Google's TPUs are probably, there's probably a mix of like mostly Broadcom and maybe some Marvell design work on these TPUs. And then it goes on to get manufactured by Taiwan Semi Manufacturing.

Starting point is 00:33:52 So it's kind of weird, like, you know, this is again where I think the media arm of the investing world has taken this idea and just gone wild with it. uh they're not actually google's chips they're not actually whoever microsoft's chips meta's chips it's ip that is purchased and licensed from broadcom and marvell and they use those patent portfolios to design some custom chip for google it's kind of weird because you know they they do ultimately own it they licensed the tech the patents and and whatnot but it's not google that designed the chips really not not really okay no that's i think that makes sense yeah it's like uh similar to arm i know there is that they had that big ipo i don't know if that's the exact same so maybe basing on that if if this continues where and again you said that

Starting point is 00:34:54 they kind of run with that narrative from the IR and media space where they try to hype up their own internal units. I don't know why. Well, I guess it just makes it seem like they're getting ahead technologically. But if they keep doing this, Broadcom can be a huge beneficiary, say, over the next decade from something like this. Absolutely. Just like Arm Holding has been a tremendous beneficiary of apple designing its own chips for the iphone and in more recent years for the macbooks um yes as long as these companies are interested in designing their own chips broadcom and marvel will be fantastic beneficiaries okay so this it feels like i've got two follow-ups here which actually lead into our remaining true or false questions so i'll go ahead and take this

Starting point is 00:35:45 one. It seems like all roads lead to TSMC one way or another. So true or false, TSMC is uninvestable due to geopolitical risks. We get that question a lot. And this one kind of leans towards taking a political stand or a geopolitical stand. And I think neither of us complain or excuse me, neither of us claim to be a geopolitical expert in this. And I don't want to discuss politics per se, um, at all, but it really, it really just depends on who you are as an investor, because it depends first of all, on how risk averse you are. And if you're willing to look at the bigger picture, um, we've got, uh, we have a lot of community members that have invested in tsmc over the past year and it's it's obviously done very very well and they are

Starting point is 00:36:46 the clear leader in manufacturing these chips i'll let you you can you can uh you can you can add in what you like can i rephrase it a little bit because yes it totally makes sense it's yeah it's something that's not really a part of like analyzing the business it's kind of its own thing you have to analyze. So let me rephrase it to you. Do you see this as the biggest risk to an investment in TSMC? No. No? Okay. I think we both say no. Okay. Let's envision a possible future that hopefully never happens or nothing even remotely close to it happens where China does invade Taiwan and like the island gets leveled, meaning TSMC goes offline, is no longer manufacturing chips. That would be probably worst case scenario, right? Imagine a world where that

Starting point is 00:37:48 happens and what happens to the global economy at that point. Are you really going to be freaked out But when you wake up that morning and realize China has just leveled Taiwan, TSMC is offline, oh, my God, what about my investment in TSMC? Is that what you're worrying about right now? Yeah, lower on the priority list. Yeah, that's a good way to put perspective in there. I would say, like, probably not even in the top 10 worries when you wake up that morning and that's what's happened. Like, literally, the world runs on semiconductor technology. this is not as simple as like some vanity issue where like well gee i guess we're going to have

Starting point is 00:38:27 to go back a few generations of iphone for a few years like the world would come to a grinding halt so much of the economy relies on semiconductors it would trigger a massive global crisis and i don't think owning a few shares of tsmc would be very high at all on your priority list and if you think that for anybody out there that thinks that i think you probably need to have a real hard think on that right like that's your bet yeah i get i get what you mean there yeah that would affect all your holdings most likely yeah exactly uh okay so let me while we're on tsmc let's talk about the business a little bit what we talked about nvidia kind of being in the pole position and the alternatives here. If you're a customer once again, and you're

Starting point is 00:39:21 a customer for TSMC, what are the alternatives for you? Why would you go anywhere else? Is there another, I guess, cost-effective solution? Not yet. This is kind of where I would say we could talk about Intel. You could let Nick loose on Intel because Intel is trying, wants to be at least certainly in north america and expanding into other places like europe and i mean they put their their plans on hold there but so they're trying to be a competitor to tsmc but that is we're a long way from that and don't forget samsung and say yes i i never talk about samsung but yeah go ahead yeah i mean very capable uh obviously that the problem with samsung is the same as or at least was up until just the last few days the same as with intel that you know

Starting point is 00:40:19 they're they design their own chips themselves as well as manufacture them so that could cause you to you know wonder okay can i really trust samsung with my designs tsmc is a customer service business it's a manufacturing business but but like first and foremost kind of a customer service first type of manufacturing operation and you can't really say the same for samsung or intel at least not not yet samsung probably not going to change its business model obviously they're doing and they're doing fine they're doing just fine not not changing the business model yeah intel's a different story though okay yeah we're going to get to intel And maybe a bridge question to that is the follow-up on maybe TSMC, and you mentioned the geographic diversification.

Starting point is 00:41:16 How is that going? Because I've heard so many conflicting news reports like, oh, they can't build it in Phoenix. Oh, it's going great. They just got the subsidies. You guys are probably tracking these new factories fairly closely. How is it going? Any surprises? Should investors be bullish or bearish on this type of stuff?

Starting point is 00:41:35 Hmm. It's going well. It's, these are mega projects, especially the fabs that they're building in Phoenix. Tens of billions of dollars don't just get spent overnight and the fab gets built without any hiccups whatsoever. It's a mega project. Uh, it just takes time and issues come up, but they're working through those issues. And, you know, Japan is another place. Uh, they have that, um this is a joint venture yeah jasm jasm yeah yeah uh that doesn't get talked about very much at all at least not you know in the states um again you know for americans you're talking yeah what is it sorry the it's the they're basically doing the same thing trying to build a fabrication facility over there yeah and and they they're complete you know they're and they're

Starting point is 00:42:32 getting ready for the next phases of the project so jasm is tsmc's joint i think it's a joint venture um in japan with there maybe it's just a subsidy maybe it's a wholly owned subsidiary casey's fact checking me yeah that's why you guys are a team anyway it's the same diversification efforts it's going well and and they're expanding in europe as well. These things are going well, but it just takes time. You don't wake up in the wake of the pandemic and say, hey, you know what would be great to have? Global supply chain resilience. Let's kind of remake everything that we've built over the last 40 years and make it a thing. Today, it doesn't work like that. It just takes time. And I think TSMC is on board with this too,

Starting point is 00:43:29 But, you know, it's also a business with shareholders. So they're going to protect themselves and their shareholders along the way, too. But they're on board with the plan of this geographic diversification. It'll happen. It just takes time. All right. And speaking of taking time, Intel, maybe I should give context to when we're recording this. It's September 23rd. This weekend, I believe there was reports that Broadcom actually proposed. It wasn't like maybe someone just talked about it, but a potential acquisition of Intel stocks down, I think 70% from their dot com bubble highs. So clearly the business is struggling. Am I true or false for Intel is Intel will be a individually publicly traded company within three years. What do you guys think? I feel like this is a solid true for me. It will be publicly traded. Yes. I think true, but I think parts and pieces of it will be part of another publicly traded company. That's what I think too. I think it'll be multiple publicly traded companies.

Starting point is 00:44:43 Yeah. So what does that mean? Is that the foundry gets separated? You guys think that gets forced on them? I think that was always what needed to happen. I think we started talking about this several years ago as needing to happen. And now everybody's like, maybe this should have happened. Yeah.

Starting point is 00:45:04 The foundry needs to be a separate business for what the United States government deemed necessary that was always the best possible outcome uh but there was obviously a lot of interest in keeping the intel empire going and it's in its you know current form resurrecting the old glory days from the 90s 80s and 90s yeah it's always needed to happen and there's a reason for that. There's a lot of economic reasons as to why, especially with leading edge chip manufacturing and leading edge design, there's a reason why the semiconductor industry has kind of split into two. You have this army of companies that do the engineering work, and then you have fewer businesses like TSMC, Global Foundries, some others that only do the

Starting point is 00:45:57 manufacturing. They're a third party fab. There's a reason for that. It's an economic reason. It's because it's expensive to design chips that are leading edge. It's also expensive to manufacture them. You're going to try to do both. It's just really, really hard. And especially when Intel has been mismanaged for as long as it has been. Finchat.io is the complete stock research platform for fundamental investors. They have all the standard financial data on more than 100,000 stocks globally. But beyond that,

Starting point is 00:46:31 they have company-specific segment and KPI data on 1,800 stocks. Want to see NVIDIA's data center revenue? FinChat's got it. Like to track Spotify's premium subscribers? They've got that too. And they just added ETFs to the platform as well. The breadth of FinChat's data

Starting point is 00:46:48 is truly one of a kind. We use FinChat every day and I've personally been using the AI Copilot more and more to summarize earnings calls and conference transcripts. To get 15% off any paid plan, go to finchat.io slash chitchat. That is finchat.io slash chitchat to get 15% off any paid plan today. The link is in the show notes. Okay. So I think that's all the true or false questions we have, but let me ask some more open-ended ones. We just talked about one of the maybe more poorly run semiconductor firms. Let's go to the flip side. What, in your opinion,

Starting point is 00:47:29 are some of the best run semiconductor firms? Okay. So I think we believe to find the best managed ones, just open a semiconductor ETF holdings page, like let's say iShares semiconductor etf socks soxx and just go down the list by market cap and whoever has like the highest market cap is probably a really well does really well managed business and i think the reason we believe that is that you know at this point we've talked about this quite a bit But this is a pretty capital intensive industry. This is a manufacturing industry. It's industrial. It probably will eventually need to be reclassified as such. It needs to go in the industrial sector. So it's hypersensitive to returns on invested capital. There has never been any toleration for running a company excessively long with with low or no profit margins like the software industry has, you know, did at least anyways in the 2010s. And so the market is only going to reward those businesses that generate positive return on invested capital because this is an industrial manufacturing industry.

Starting point is 00:48:53 all right and if it is you know an industrial industry as you mentioned it's historically been cyclical uh although i think what's funny is in there was some hints or there was a debate in i forget when when it was 2021 2022 that the the industry wasn't cyclical but i i feel like usually that's right around the time that the market ends up going into the downturn. But either way, it's kind of an interesting one where it's cyclical, but it's also long-term. The demand over the last, if you take five decades, has grown exponentially. So how do you think about that from allocating your portfolio, making investment decisions, looking at a cyclical industry and evaluating the stocks down. Yeah. It's really important to mentally think about it

Starting point is 00:49:49 in a much longer timeframe than you normally would. Play the long game and really try to cut through all the hype in the media frenzy that happens, especially in a cyclical industry like we're in right now, and just get used to those wild swings and get ready to buy when there's a dip For instance, if you have conviction in a certain company for the long term, buying the dip or the down cycle makes sense because that company you believe is going to weather the storm. Do you think it's better – I've kind of thought about this with semiconductor companies where it might be harder to time the cycle, but you have – if you go back and you look at 20 years of financials, even through the cycles we just talked about this, the bottom of the next down cycle could be higher than whatever, 2015 in terms of demand for any company. do you think it's better to just go into it thinking all right no matter what i'm buying if you let's say you believe in a uh some company in the semiconductor ecosystem i'm buying this and i have to hold it for at least five years that's my like do you think it's good to build

Starting point is 00:51:05 those conditions for yourself so that you you can weather the down cycle especially in this industry Yes, because the ones that underperform are going to have a horrific time during the down cycle. They may run, you know, they may operate at a loss for an extended period of time. So I think you have to by default, good practice for long term investing anyways, right? Like, am I comfortable owning the stock in the case of a recession? I think it's just a good habit to think about that. But especially with like semiconductors, yes, that's a good thing to think. Okay, I want to own this for five plus years.

Starting point is 00:51:48 Am I comfortable owning this for five plus years? Because there will be a down cycle at some point in some form or fashion. Is the business built to survive that type of environment? Because if it is, it'll come back out the other side, a stronger company with probably more sales and better profit margins than the previous cycle, like you just said, Ryan. Okay. Makes sense. Now, the oxygen in the semiconductor room is getting all taken out by AI at the moment. We've talked about that.

Starting point is 00:52:20 Clearly, you mentioned it as a way for, say, a layman like us to understand it is, hey, there's this accelerated compute. There's the new spend. There's the replacement spend. This is going to be a long-term thing. outside of AI, what should investors be excited about in the semiconductor sector as a whole? I think for me, one thing that I'm excited about is to see what accelerated computing and AI can do for advancing biosciences. I think that's a really interesting field. I was in the healthcare field for a number of years and looking at some of the advancements that can be made with

Starting point is 00:53:06 AI training is just absolutely incredible. Just thinking about what NVIDIA is doing with biopharma or genomics or diagnostics, there's a lot of potential there. So for me, I'm excited to see what happens with the healthcare and biotech field. Let's see. Power. Power consumption. I actually, and maybe to preface this, Brett, you mentioned 2021, 2022, there was talk about, maybe the semiconductor industry won't have a down cycle this time. And I bit on that. I was a believer in that. And I'm happy to admit that I was wrong. However, there is more to that story because enterprise computing, non-cloud especially, which, you know, kind of has morphed into accelerated computing, enterprise computing. So think businesses that have like a server on-premises somewhere, not cloud, and consumer electronics got blasted. But then like NVIDIA comes along, doesn't skip a beat.

Starting point is 00:54:21 And then the auto industry kind of kept on going until halfway through 2023 before the down cycle came for them. And the reason being is there's a lot going on in tandem with semiconductors right now. And this insatiable demand for more accelerated computing, more efficient computing. How do we continue doing the same amount of computing with less power consumption? And there's a tremendous amount of opportunity there because it's an operating expense. Energy consumption is always OPEX for any business. And so you always have an incentive to drive that down towards zero, if possible. And so, therefore, with that in mind, if that's an OPEX that is always an opportunity to drive as close to zero as you can possibly physically get it, there's always opportunity there for power chips to help drive the cost to zero.

Starting point is 00:55:26 And so the poster child for power chips was the automotive industry, electric vehicles, and that's fine. Amazing technology, really cool stuff happening, but that's just like one slice of that pie, one piece of the puzzle. There is a tremendous amount of stuff happening in power chip development right now and a very, very long multi-decade, maybe forever, endeavor in getting work done with nothing. The electrification of everything, potentially. Yeah, and then making that more efficient. Yeah, it seems like, hey, we could keep doing that forever. We didn't want to maybe mention electrification of cars, you know, the transition to electric vehicles. I always thought that was an interesting tailwind for some of these companies because you kind of see like, well, yeah, I mean, I don't know if it's obvious, but I would definitely yes, that the vast majority of sales are going to be electric in 15 to 20 years.

Starting point is 00:56:28 If the progress that has been made keeps getting made, who benefits from that? I know Texas Instruments has been one that gets talked about a lot, but are there any other companies there? And I know they've struggled. So maybe, you know, was this overhyped for some of them? I think our bets in vehicle electrification or electrification of all kinds of things are Qualcomm and OnSemi. And we just had an addition to our portfolio of monolithic power systems. Those companies we see as real beneficiaries of the electrification of cars. Someone's going to criticize what Casey just said too, and don't do it. Qualcomm is part of the electrification process, even though they don't make the power chips, because all this tech is getting crammed into the car to make the modern car as close to this as possible.

Starting point is 00:57:28 because this is what we all look at every day and what we want to look at. Let's make the car just like that. It's just a good sales point, but that consumes energy, consumes power. Any process whatsoever is going to place a power demand on a system, be it an internal combustion engine or a battery.

Starting point is 00:57:53 It doesn't matter. It will consume power. And so a company like Qualcomm that makes very efficient arm-based chips, mobile-based chips that enable connectivity that consume less power over time is obviously going to be a benefit. Because the automakers all know they need to cram the tech into the car. So how do you do it more and more power efficiently every single year? Don't criticize Casey. I can hear it. I can hear it now.

Starting point is 00:58:23 I could just imagine. Yeah, well, as we all know, YouTube commenters are so nice all the time. No, they're great. They're great. But they can get a little bit sassy. Maybe you guys have mentioned Broadcom and Qualcomm a lot. Have they been permanently underrated? Because I feel like when I read stuff, as someone who's not in this industry deeply,

Starting point is 00:58:43 it's like, oh, NVIDIA, Texas Instruments, DSMC, ASML. They seem like they've done so well. I mean, have these companies been just left for, not left for dead, but people forget about them and they've just done incredibly well. Yeah, they're underrated for sure. Broadcom especially, but you know, there's some special situations with that though too, because it's, it's basically, it's often compared to like a private equity firm, more or less. But even Qualcomm, yeah, they've done great.

Starting point is 00:59:16 qualcomm is like everyone thinks about chips in phones like in their mobile connectivity but they do so much more than that and if you look at especially the last three quarters of their automotive segment has grown exponentially in a downturn in a downturn in a down cycle for automotive so yeah i think so i think the answer is yes both of those yeah it was funny with me and Ryan, I think we've gotten explained what Broadcom and Qualcomm are multiple times. It's like the Federal Reserve where you get it explained to you and I'm like, I think I kind of know

Starting point is 00:59:51 what they do now, but there's still a bit of a mystery to me. That's all my questions. Ryan, did you have any follow-ups before we close things out? I think that's it. I guess for anyone, any of our listeners that have never met you guys, listened to you guys before,

Starting point is 01:00:06 where is the best place to see more of your work? On YouTube. at our ChipStockInvestor channel. YouTube forward slash at ChipStockInvestor. Perfect. All right. Well, that is going to do it.

Starting point is 01:00:24 I will bring us out here. Thank you all for tuning in to this episode of Chit Chat Stocks. We want to remind listeners that Brett and I are not financial advisors. Anything we say or anything guests say on this podcast is not financial advice

Starting point is 01:00:38 or a formal recommendation. We may buy, sell, or hold any secure any securities discussed on this podcast so please do your own work do your own due diligence thank you all again for tuning in and we'll see you next time

Chit Chat Stocks - The State of The Semiconductor Market With Chip Stock Investor (NVDA, TSM, ASML, INTC, AVGO, and More!)

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.