Motley Fool Money - Semiconductors 101

Episode Date: October 8, 2022

Your remote control, the fastest supercomputers in the world, and artificial intelligence applications have something in common: the semiconductors needed to make them work. Jason Moser and John Rotonti dig into one of the most important industries in the world and discuss: - The semiconductor value chain - Geopolitical risks (and opportunities) and the industry's cyclicality - Implications of the CHIPS Act - One critical chipmaker that’s cheaper than the S&P 500  Companies mentioned: NVDA, CDNS, SNPS, ASML,  LRCX,  AMAT, TOELY, ASML, TER, INTC, TSM, SSNLF, AMD, QCOM,  KLAC Host: Jason Moser Guest: John Rotonti Producer: Ricky Mulvey Engineers: Rick Engdahl, Dan Boyd Learn more about your ad choices. Visit megaphone.fm/adchoices

Okay, caller one wins court side seats to tonight's game. What? I won floor seats? You did. I've been calling for 13 months. Wait, Chris? Yes, I finally did it. What are you going to wear? Men's Warehouse.

They've got today's looks for any occasion, and I need to look like a celebrity. Don't want to stick out. Exactly. They've got Chill Flex by Kenneth Cole, Joseph Abood, and a tailor at every store for the perfect fit. Congrats. You can stop calling now. Not a chance. Hit any look for every occasion at Men's Warehouse. Love the way you look.

10 to 15 years ago, the industry was largely driven by PC cycles and iPhone cycles, to be honest with you. But now, like we said, they go into everything. I mean, is demand for the cloud going to go away? I don't think so. I'm Chris Hill, and that's Motley Fool's senior analyst John Ratante. He joined fellow analyst Jason Moser for an introduction to semiconductors, computers, computers, cars, and whatever you're using right now to listen to this podcast. podcasts need them to function. So they're kind of an important building block for our economy. Today, John and Jason discuss companies in the semiconductor value chain, the industry's risks,

and one important chipmaker that's cheaper than the S&P 500. Let's just go ahead and start from the very beginning. What are semiconductors and why are they so important to the economy now? Thanks, Jason. One of my favorite things to talk about. So semiconductors are the brains of all electronic devices. They partially or semi-conduct currents, and they are largely made of silicon, which is made from sand, like beach sand. They vary in sizes, but most of them are, you can think about the size of your thumbnail or a postage stamp, and these very small devices are packed with, in some cases, tens of billions of trillions of trillions of trams.

transistors. That's a lot. Yeah. The size of, exactly right, especially when you're packing them onto something the size of a postage stamp. And these transistors amplify currents and actually turn electrical signals on and off. And so you've got these electronic brains that power devices, and they are in everything from the remote control on your TV to the fastest supercomputers in the world, to healthcare devices, to our military's weapon systems. And so they are a matter of life and death, literally when it comes to health care, when it comes to national security. And they are the devices that power our modern digital economy and digital world. So we would not have cloud computing without semiconductors. We would not have artificial intelligence or machine learning. We would not have

video gaming. We will not have the Metaverse if that comes to play. All of these things are EVs, 5G, all of these modern technologies, all of these innovations are powered by semiconductors. So to sum up, I would say they are the critical infrastructure and modern building block of the digital economy. Yeah, and it really does feel like we're only headed more and more in that direction. So, their role really should only grow in importance. And so I think, when we look at this industry, I find the actual value chain in semiconductors be very interesting. I think there's a lot that goes on behind the scenes that maybe some folks don't realize. In the market that comes to mind, it reminds me a little bit of the payments of history, right?

If you look at it on the surface, you think it's kind of just simple. Oh, well, companies make these semiconductors and boom. It goes into your device and consumers use them. But then when you look at the value chain behind the scenes in all of the different participants in this market, it's worth knowing because there are a lot of different ways to capitalize on this as investors. So let's talk about the main touch points of this value chain. The semiconductor name, most people are probably familiar with names like Qualcomm or Intel, right? And those are semiconductors companies, but there are companies even before that, right, that

are involved in making the actual physical material, right? I mean, you've got everything from foundries and the equipment companies and then designers. So walk us through that value chain, if you could. Yeah, Jason. The value chain is fascinating because it is so complex and technologically advanced and capital intensive that if you look at the value chain, which we'll go through in a second, at every step of the semiconductor manufacturing value chain, these global oligopolis or duopolis, or in some cases, even earned monopolies have formed. And so basically the first stage of making a semiconductor, designing a semiconductor,

is designing it using software. And so you take a company like Nvidia or Apple that wants to design their next semiconductor, they're going to use software very likely from one or two companies, because these are the two companies that do it at scale globally, and it's synopsis and cadence design systems. If you think about an architectural blueprint for your house or a building, this is the architectural blueprint for a semiconductor. Just much more complicated, because like we said, there are tens of billions of transistors. They need to map and blueprint onto the chip. And so these two companies, they are electronic design, automation, software companies.

really their computational design software companies. This is highly, highly complicated. This software is using a combination of matrix, algebra, multivariate calculus, AI, advanced geometry, and more. This is not a software that you take an online course in Python or something. You learn how to use it. This is software run by specialists, masters, and PhDs in this kind of software. So that's the first stage, is to blueprint it out using this. software. And then what happens after that is that software is printed onto a mask. And then that mask, which is just like a stencil, basically, or a model, that mask is then used to print the blueprint onto a silicon wafer, which is usually 12 inches in diameter. So it's a circular, thin piece of silicon that's 12 inches in diameter. And to do that,

you use extreme ultraviolet light and a series of lasers and mirrors to print that blueprint onto the silicon wafer. And there's only one company in the world that does that. Extreme ultraviolet lithography. We'll get into that in a second, but it's ASML, a Dutch company. Then from there, basically, there's three main steps. You deposit a bunch of chemicals onto these chips. and those deposition, there's basically three primary companies that do the deposition or deposit, and that's lamb research, applied materials, and Tokyo Electron.

And then you etch away or carve out little holes where the transistors go, and that's called etch. And there's basically three companies that do that. And it's the same three. It's Lamb Research, Applied Materials, and Tokyo Electron. So you had two main software companies. You've got three main semi-capital equipment companies that do the etching and the deposition. You've got one company doing the light sourcing, that's ASML. And then you have a couple companies that test it after the semiconductor is made,

and the two big ones there are Terradine and Advent Test, which is a private company. And all of this takes place in a foundry. and the big foundries, the largest foundry is Taiwan Semiconductor. Samsung is another leading player. There is a major player for lagging edge semiconductors' foundry manufacturing, and that's global foundries. And then Intel is a fourth player trying to get into third-party manufacturing. So that's the semiconductor manufacturing value chain in a synonymous.

Well, I mean, very well done, by the way, and I appreciate it. I think you really just, you showed this industry that just seems to be so simple on the surface as consumers, we buy these devices, but the behind the scenes is just so much more involved. And it's always great to know that from the investing perspective. I mean, knowing that value chain, I think is key for whatever industry you're focused on. Semiconductor, certainly you can see is a very important.

involved value chain as well. And it seems like there's a theme there in that this is just very highly technically skilled work. I mean, the barriers to entry just on understanding how to do this and getting the talent to be able to do this and getting the equipment and the software to be able to do this, it just seems like those barriers to entry are very, very high. I would agree. Yeah. So you focus your coverage specifically on foundries and semi-cap equipment companies. And so I wanted to dig into those a little bit more because that's where you, That's where you pay most of your attention. Who are the major competitors?

We talk about them a little bit. Let's dig into the major competitors in that part of the value chain, and the foundries and the semi-cap equipment companies. And we'll talk a little bit about their advantages and their merits as investment ideas. You know, starting with Foundries, we mentioned Taiwan Semi and Samsung. Those are the two leading, advanced leading edge semiconductor third-party manufacturing foundries. There's also global foundries, Intel, maybe a few smaller players, but you can't talk about

Foundries without talking about Taiwan Semi. And the reason is because I think it's one of the, if not the most indispensable company in the world. It has 50% market share of all of the outsourced chips manufactured in the world. That's roughly three times higher than the next largest player, which is Samsung. They have, so 50% market share of all chips manufactured, but they have 85% to 90% market share of the world's most advanced chips. And because of that, it generates 90% of global contract foundry profits.

So it really is this earned monopoly or duopoly with Samsung right now. And it got this position by investing heavily ahead of everyone else, heavily ahead of demand, and being an early adopter of extreme ultraviolet or EUV lithography machines by ASML. Taiwan Semi owns more ASML machines than anybody else on the planet. Taiwan Semi has more experience using these extremely complex machines than anyone else on the planet. Because Taiwan Semi has 90% market share of the world's most advanced chips, it has a much, much larger library of recipes or process knowledge or process technology for manufacturing these chips. Each chip has its own recipe, and each node builds off the recipe of the prior node.

So with semiconductor manufacturing, Jason, scale and market share beget more scale and more market share and faster and better process learnings. Process knowledge is extremely capital intensive. I think Taiwan Semi is going to spend somewhere on the order of 40 billion U.S. dollars in CAPEX this year or next year. 40 billion in one year. So it's going to be very hard for competitors to catch up. And finally, Taiwan Semi has built up an ecosystem and works extremely closely. with all of the major players in the semiconductor manufacturing supply chain that we already talked about. If you go to Taiwan and you go to their facilities, well, guess what? Land Research is located right there. Applied materials is located right there. ASML has teams right there. So they built up these ecosystems around the business. I think it's very clear the pros, right? I think it's very clear

the competitive advantages that these businesses possess. And I think the follow-up to that I'd really have for you is just, you know, what do you consider the threats? for businesses like these. So we could, you'd talk about, are there really any threats for Taiwan Semi? I think so. And that ties into the valuation right now, Jason. So Taiwan Semi is at, I looked yesterday. I haven't looked this morning. I know stocks are down. I think it's around 72, down from 145 stock price. So it's down 50% from its 52 week high. It has 2.5% dividend yield. Jason, this company that I just explained to you that is integral to the world is trading at a 4 PE of 12. Why? So I think, and let me just put that into perspective, the S&P 500, the market is trading at about 15 or 16. So it's had a three or four turn discount to the market. I think the

reason, and you ask the risk, I think the market is concerned that it's located in Taiwan and that China claims to control Taiwan. And there's been talk and rhetoric recently that China maybe possibly could invade Taiwan. So that's the risk, and that's why I think there's this overhang on the stock, Jason. Yeah, the geopolitical risk. I mean, something always to keep in mind, obviously, out of our control, right? Nothing we can really control, but it's always something to acknowledge. You see obviously what's going on with Russia and Ukraine.

I mean, those are things that happen. You can't very well predict them, but then it also, it goes to show you. Even market leaders like that, there's always going to be, a risk that you need to identify, and that certainly seems to be a very reasonable with Taiwan. Even if you take China and geopolitical risk out of the equation, there are the risks. Intel is spending billions of dollars to try to catch up. Billions. So, you know, competitive risk is there. Technological obsolescence to an extent is there. And then, you know, semiconductors have the industry has historically been cyclical. I think it's less cyclical now than

it's been in the past, and we can maybe talk about that now or another time. But there is some cyclicality as well. So there's a risk that you could be buying a stock at the top of a cycle. I don't think that's a risk right now. But over a long period of time, there's a risk you could be buying at the top, and then you'd have to maybe wait a couple years for the stock price to catch up. But I think those are the big risks. Now, I'm glad you brought up that cyclicality risk, because I think that's something that definitely is worth touching on. I mean, you know, I run a couple of services here at the Fool that focus on technology, immersive technology, 5G, stuff like that.

So I've got my share of chip companies in there, your Qualcomms and your A&Ds and your invidias of the world. It does feel to me like the cyclicality risk isn't what it used to be. And I guess that really goes back to what you were talking about at the top of the show in that this is the lifeblood of virtually everything we do now, this technology. So it's not like we, it's not, it's in everything that we do. It's exactly right. So it's something that it feels like the cyclicality now is more or less just based on where

we are sort of in the innovation cycle. But regardless, because this technology is so widespread all around the world, that cyclicality window just seems to be shrinking a bit. I think so, Jason. You know, 10 to 15 years ago, the industry was largely driven by PC cycles and iPhone cycles, to be honest with you. But now, like we said, they go into everything. I mean, is demand for the cloud going to go away?

I don't think so. Is demand for EVs and 5G going to go away? Is demand for AI and machine learning going to go away? And so the use cases, the end markets, the total addressable markets for semiconductors is just so much larger than it was. The other thing is that we've had a couple of major periods of consolidation across the industry. So where, you know, some of these, at some points of time, there used to be 20, 30 competitors. now you've got three, four, or five competitors. And, you know, because of that, I think the remaining competitors are much more rational in their pricing.

They don't, you know, that avoids the boom, bust pricing cycles of the past. And then if you look at the fundamentals, you know, the last down cycle we had, I think, was like 2018, like real down cycle. We're kind of in one now a little bit. But if you look at the trough margins now, in 2018, so the lowest the margins get at a down cycle, those are now higher than peak margins in previous cycles, Jason. Wow. And so the fundamentals of the industry have dramatically improved in my opinion. That is, that's astounding.

You know, another thing that's been in the headlines a lot lately, we've been talking a lot about it on the investing team, trying to get just a better idea of where we think this ultimately, how this plays out. the Chips Act, right? I mean, this supply chain crunch that we've been going through. I mean, it's something that we see in virtually every Arnings Call because this technology is in everything. You know, so it seems like every company on the face of the earth has exposure to this. Sure. The Chipsack, I wonder, do you have an opinion there? I mean, this, we're talking about 50 plus billion dollars, ultimately that's going to be devoted to spurred domestic semiconductor manufacturing. And it feels like on the surface, that's a smart idea, right?

Diversify that supply chain a little bit. Yet I can't help but wonder if maybe this isn't money that's just kind of disappear here and there and not really have as material in effect. I wonder if you have any strong feelings one way or the other on the chips act. I don't have strong feelings on how effective it will be yet, just because it's so new. But I'd like to think, Jason, that it's a good start. If we want to build a homegrown semiconductor domestic supply chain, 50 billion is not enough, right? I mean, we talked about Taiwan semi is going to invest 40 billion on its own this year. Just one year. Just one year, one company in one year, Jason. And if you look at the U.S. used to manufacture a lot more

of the world's ships, it used to be 30 or 40 percent of the world semiconductors were manufactured in the U.S. Now it's 12 percent. And so, you know, that's a national security risk. It really came into play during the pandemic when we had all the supply chain bottlenecks. You know, we could really feel it tangibly. You couldn't buy a, you know, auto manufacturers, for example, couldn't make new cars. So they couldn't because they couldn't get the semiconductors because nowadays cars are just chips on wheels. And so what happened was everyone started buying used cars and that drove up used car pricing. You know, I think it's a good first start. I think it's going to take a decade or more if we're going to be successful. And it's going to take hundreds of billions of dollars. But we have

have to start somewhere. And so I'm hopeful, Jason. Yeah. I think that's a good way to it. I'm hopeful. To me, it absolutely feels like it's just a start, because like you said, it requires so much investment. Yeah. One more quick thing, Jason. You know, when it comes to manufacturing, the U.S. is at a deficit. Most of the chips are manufactured in the East. That's just a fact. Two-thirds to 70 percent of the world's chips flow through the island of Taiwan. So, they're made in the East. But we should also point out that the IP, Jason, exists in the West. So those two software companies, Cadence and Synopsis, those are U.S. companies.

Oh, yeah. Those three large semi-cap equipment companies, lamb research, apply materials, Tokyo electrons in the East, but let's throw in Terradine and let's throw in K-LAC. Four huge semi-cap equipment companies, all located in the West. and then ASML may be the most important of the semi-cap equipment companies. That's a European Dutch company. So a lot of the IP, the majority of the IP, is in the West. So there is this dichotomy of the IP versus the manufacturing.

We're just trying to bring back some of that manufacturing. Yeah, yeah, and that makes a lot of sense. Given that, let's wrap up to show today. You and I were reading an article the other day, an article from CNBC. It was just an interesting sort of view. It kind of struck me as sort of the past versus the future, right? This was an article that dug into Intel and Nvidia. Intel is saying that Moore's law is still alive.

And well, Nvidia says no longer. Really quickly, what's Moore's Law and what did you think about this article? Jason, Moore's Law states the number of transistors on a chip doubles roughly every two years. And if we assume the price of the chip stays the same, then the cost of that chip falls in half every two years. So basically, Moore's Law says that the chip performance doubles every two years. If we look back in the 1970s, Intel was a chip leader, the chips that was putting out in the 1970s had between 2,000 and 6,000 transistors on them. Today, the Apple M1 chip, which was pretty revolutionary, Jason, has 16 billion transistors. a chip the Graviton 2 that goes into Amazon Web Services that runs their cloud and their data centers.

It has 30 billion transistors on it. And so the question is, Jason, how many of these transistors can we continue to pack onto something the size of your thumbnail? That's right. I mean, we run into a physics problem at some point, right? I mean, the edge of physics, right? The edge of physics. And so what we do is we do two things. One is we make 3D chips and we start building transitions on top of each other.

It's like a skyscraper. And then the other thing is we use extreme ultraviolet light and these machines by ASML. The reason is the reason you have to use this extreme ultraviolet light is because you need extremely small wavelengths of light in order to trace the pattern of the transistors. You have to bring the transistors closer and closer together. And the only way to do that is to use extremely small. extremely small wavelengths of light. And so, you know, who's going to be, you know, I think in that article, Intel was talking about getting to 100 billion transistors on a chip.

And we're at like 50 billion today. So that's, once again, that's a doubling, right? Doubling from 50 to 100. Can we get there? I hope so, because I want to see what the world looks like with 100 billion transistors on a chip. Invidia is saying, you know, we're up against the limits of Moore's law. to get around that, you're going to have to make highly, highly specialized chips used for highly, highly specialized devices and purposes. And that's what Nvidia. You know, Nvidia doesn't

make commodity chips, Jason. Invidia makes chips that only it can make. And so who's going to be right? I really don't know. But I tell you what, I, you know, Nvidia is an amazing company, and Intel is a company that was once amazing and is trying to turn itself around. And I think if anyone could do it, it's Pat Gelsinger, their new CEO. So it will be fun and interesting to watch. It will indeed. It will indeed. This is an amazing space, tremendous opportunities for investors, and really appreciate you taking the time to go through it and explain it all to us.

John. It's great, great talking with you today. You want to make sure before we wrap up, you're a good tweeter. Thank you, sir. You got a great educational investing Twitter feed there. Where can people find you on Twitter? I am on Twitter at J Row Grow. So at JRO, GRO, GRO.

And thank you for having me on the show, Jason. It's always fun. As always, people on the program may have interest in the stocks they talk about, and the Motley Fool may have formal recommendations for or against. So don't buy ourselves stocks based solely on what you hear. I'm Chris Hill. Thanks for listening. We'll see you tomorrow.