Semiconductors Market: The Core to the Modern Economy

The semiconductor market is a major part of the modern economy. Since these chips power our daily activities and make up a large part of financial indices, we discuss some important factors to understand when exploring the sector for investment.

Summary – Cents of Security Podcasts Ep. 101

The following is a summary of a live audio recording and may contain errors in spelling or grammar. Although IBKR has edited for clarity no material changes have been made.

Cassidy Clement 

Welcome back to the Cents of Security Podcast. I’m Cassidy Clement, Senior Manager of SEO and Content at Interactive Brokers, and today I’m your host for the podcast. Our guest is Brian Colello, Technology Equity Strategist at Morningstar. 
The semiconductor market is a major part of the modern economy. Since these chips power our daily activities and make up a large part of our financial indices, we’re going to discuss some important factors to understand when exploring the sector for investment. Welcome to the program, Brian. 

Brian Colello 

Thanks, Cassidy. Good to be here. 

Cassidy Clement 

Yeah, it’s awesome to have you on. So why don’t you tell our listeners a little bit about your background since it’s your first episode on the program? 

Brian Colello 

Yeah, gladly, happy to be here. Again, Technology Equity Strategist with Morningstar. I’ve been an equity analyst with the firm going on 17 years now, covering semis—starting with small cap to large cap to now mega cap. When you think about Nvidia at $3 trillion or so—I covered Apple for a few years in between. 
I had some managerial roles in between as well. But now, focusing a bit more on artificial intelligence, which is certainly keeping me busy. My background is in accounting. I’m a CPA, so I had the finance background and had to learn the technology. And I’m always jealous of the engineers that have the engineering background and needed to learn the finance because they could dig into the technical details a little bit better. 
But yeah, I’ve been covering the space for quite a while, and so many things have changed and continue to change—and nothing is moving faster right now than artificial intelligence. 

Cassidy Clement 

Yeah, when it comes to AI and almost any talk of the town, if you will, when it comes to financial commentary, it seems like semiconductors tend to be at the core. 
When we recorded our previous episode, it happened to be on minerals and metals, and we were talking about how that kind of impacts how semiconductors are made—the manufacturing, the pricing valuation sometimes when it goes into companies. 
But while it may seem like a hot topic, there’s definitely a history to it—past 15, 20 years. With your background in the industry, why don’t you tell the listeners a little bit more of, we’ll say, the brief history, the highlights, the emergence of it all? Because it actually goes pretty deep. 

Brian Colello 

It does. You could go back to 1940s and 1950s, and there’s definitely some fascinating stories there with the formation of Fairchild Semi, the defections of some of those engineers that led to Intel. 
Silicon Valley has its name based on the semiconductor manufacturing. These are all based on silicon. And so that’s where Silicon Valley comes from. 
For any listeners that are interested in the space, I think Chris Miller’s book Chip Wars does such a great job of describing both the history and where we stand today from a geopolitical standpoint. It’s not just the technology of semiconductors that is top of mind, but the geopolitics around it—that most of the high-end manufacturing in this world is done in Taiwan, which is not the necessarily safest or most benign of regions, given how close it is to China and some of its ties there. 

But in general, starting—if you go back—starting let’s say with the dot-com bubble, or starting with the ’90s, it was the emergence of the computer with Intel inside on those PCs. 
And maybe even just taking a step back: semiconductors are used in basically any type of electronics that you have. So it’s PCs, smartphones, certainly in the cloud for artificial intelligence. They’re in your car, they’re in your electric toothbrush, your garage door opener—really everything in between. 

So if you go back to the ’90s, the Intel and the PC process was probably really the biggest, most important piece or where it came into everyday consumers’ lives, because it powered those PCs with the PC boom and the dot-com boom. 

From there, if you get into the early 2000s, you have the rise of the flip phone and then the rise of the smartphone—another huge wave. So again, perhaps one of the best pieces of Apple hardware in a phone—I always look at my phone when I’m doing that—but it’s the Apple A-series processor inside of there, that amazing technological advancement that Apple made to go there. 

And then you also need chips in there to connect to the wireless network, like those from Qualcomm to connect to 4G, 5G. You need a chip for Wi-Fi made by somebody like Broadcom. And then there’s a host of chips in between to regulate the voltage, filter the signals, interfaces, store memory—things like that. 

The next big category that I think of is broad market. And there’s really two areas there that I focus on. First is automotive and the car. And if you think of any car commercial today and the new bells and whistles going into cars—that’s all being done with electronics, and that’s powered by semiconductors. So that’s the first big wave in cars. 

The next one is electric vehicles, which again, you have a battery, you have plenty of semiconductors in there to manage that power. They’re used in electric vehicle inverters. They’re used in the onboard charger. You have battery management systems to make sure that the fuel cells are balancing and loading properly and things like that. So there’s a whole wave just going into the car. 

And then I think about industrial and the Internet of Things—IoT. And this is—for the average consumer—you could think of smart gadgets, and they’re probably all over your home. So if you’re using, let’s say, a Ring doorbell or cameras, Wi-Fi, smart meters—all these things that connect to the Wi-Fi in your home, that’s all again done with semiconductors. 
And then in factories and industrial equipment, medical equipment—we’re seeing the same sort of thing. Everything’s getting more processing power, more connectivity, more sensors. 

And then finally, the last piece—the biggest piece and maybe the most important piece going forward—is AI, which is being led by Nvidia. 
And Nvidia is a name that has been on the ground floor of AI development—not just building the semiconductors to enable AI in the cloud, but building the software that allowed all these AI developers to build their large language models. 

And so Nvidia is really embedded in that ecosystem, and that’s why they’re so dominant and have grown to be such a valuable company. And we think that continues going forward. 
All of these ChatGPT queries and any large language model you use—any AI—it just takes a massive amount of computing power, exponentially more than what we were using just a few years ago. So what Nvidia and others are providing behind the scenes is really a massive opportunity for the chip space. 

Cassidy Clement 

Yeah. I think when you start to actually break down the integration of semiconductors throughout various products and actual day-to-day utilization throughout—we’ll say, I think you said around the forties or the fifties. Fifties. So let’s say around the past 60, 70 years, realistically. These companies, as long as they were able to stay in the game, the way that it was severely advantageous for them and not necessarily the product makers was because the products themselves can have obsolescence, but the chips themselves are able to adapt and have to be used in all products going forward. 

 
So it allowed a lot of a… adaptability, but then it also, as you mentioned with Taiwan, it shifted a lot of the manufacturing necessity to other places that are able to yield a higher manufacturing skill with utilization of westernized economies paying for that. The other flip of that is Moore’s Law. When it enters the picture, you’re starting in with a set of a handful of TVs companies—or maybe when remotes started happening—and then you immediately go, as you mentioned, all the way to something like Apple manufacturing their own silicon, which is not something that every company can say that they can do in-house. 

But so much of today’s processes and day-to-day activity, even in the smallest of ways—even the thermostat on the wall that I’m looking at right now—I know there’s a chipboard in there. So how exactly would you define semiconductors’ role within the economy? Do you think it’s the main core—no pun intended—to the modern economy because of how much is relying on it, and that’s why these companies’ performances mean so much within almost all main indices for stock market performance? 

Brian Colello 

Yeah. E-e—exactly. You have a lot there. I actually chuckled when you mentioned the thermostat, ’cause that’s one of my favorite examples as well, ’cause I cover analog semiconductor companies like Texas Instruments and Analog Devices. And the thermostat’s a great example—you set your thermostat to 71 degrees and it’s never exactly 71. So you need the analog chip to sense those slight variations in temperature, but then you need to then convert that to a signal—a zero or one—that a microcontroller might run to say, “Okay, yes, now it’s actually below 71, so turn on the heat,” or whatever it may be. 
Yeah, that’s just one of the countless examples you have there. 

You also mentioned Moore’s Law—maybe taking a quick step back into the history—that has really been the driver of this industry over the past six years and why it’s been as important as it is: that you’re getting basically double the performance, double the amount of transistors every two years. 
And Intel led the way for 50, 60 years—everything until the last few years, quite frankly. And Intel had a very strong competitive advantage in that chip manufacturing. They were in the lead, which allowed them to have premium pricing so they could invest more and stay on that cutting edge. 

And then really, you get to the point of—now you’re moving atoms and electrons just at a… just such a miniature scale that it’s really hard to execute, and Intel’s execution floundered. And so they went from having a competitive advantage to having a competitive disadvantage to Taiwan. 

And so when you get into geopolitics and things like that, it’s that sort of—let’s call it slip-up—from Intel, of just not being able to keep pace with what Taiwan was doing, that is potentially causing some of the issues right now in the economy. And certainly geopolitical. 

I think the U.S. woke up and realized, hey—two things. One, semiconductors are extremely important. So to your point, are they the core of the economy? Like, they are tangential for sure. They’re in everything that you… whatever you think the core of the economy is, there are chips inside of those products. That’s the piece. 

But second, the realization that the vast majority—and certainly all of the leading edge—are done in Taiwan, or even… Samsung was, quite frankly, second place in Korea. And Intel is still the main chip manufacturer in the U.S., but not on the leading edge. They’re trying to catch up. We think they’ll get close—they’re still not there yet. Taiwan’s going to keep rolling along. 

And then finally, you mentioned… there’s such high fixed costs with manufacturing too, that over this history, you had a lot of companies. The saying was, “Real men build their own fabs.” 

 
And that saying goes back, I think, to the 1980s or so. But fewer and fewer companies do it. And now Intel’s really the only digital chipmaker that does that. And that might even be fading away over the next couple of years, ’cause it’s just so costly to build. 
So there is a concentration of talent, a concentration of capital, of investment, a concentration of equipment in Taiwan. And it’s not easy to move that back over from Taiwan to the U.S. 

I think that Taiwan ecosystem would still be extremely hard to replicate. And they’re trying—there are incentives for Taiwan Semi to build in Arizona, and they’ve committed to building new places. And again, when you get into geopolitics and how important it is to the economy, they’re discussing big numbers to build in the U.S. 

But that is not—that’s not a switch you can make in three, six, even twelve months. We’re talking maybe a three- to five-year process. 
And again, if you think about tariffs on semiconductors coming potentially—you could put on whatever you want—but nothing is going to happen much faster than three to five years, no matter what the tariff is or is not. And that would probably be—I guess if I were in the White House—that would maybe be my quibble. 
Anything involving tariffs and semis over the next three months is really just a price increase, because you can’t—the industry just can’t shift that quickly. Again, concentration of talent, big fixed cost equipment. 

Yeah, I think it is core. And then, back to your question of—is it core to the economy? If you don’t think semiconductors are, certainly people think automotive is, or technology, or AI—and chips are at the center of that. So they certainly have their piece, and it’s rising. 

And again, I’ve been covering the space for 17 years and it’s more popular now and more in the national consciousness than certainly when I was covering it. 

I think one of my realizations when I was trying to learn this industry is, “Oh, this is all pretty cool—technology and equipment.” And there was this new fancy item coming out that was called the iPhone. 
And I realized that I was covering all of the companies that sell all of the components that go inside to make that work. 

At a cocktail party talking about it all and everyone showing their iPhones for the first time—I could point to the chips inside and how it all works. And that was really interesting. 
And I think I actually probably educated some people, and I don’t think I have to educate that many people on that sort of discussion anymore. I think everyone knows the complexity inside these things. 

Cassidy Clement 

Yeah, it’s also very interesting from the perspective of the supply chain changes—or the anticipated supply chain changes. 
The education and the skillset that comes with that is something that also, outside of Intel, the United States as a country… usually isn’t… usually prepared for. 
So that’s why that three to five years, at least from discussions I’ve had with other people in the space—that three to five years seems way more realistic than the one-year turnaround. 

Since you mentioned AI and all of these supply chain issues, but there’s still a high desire for semiconductors, obviously—what are some trends or drivers that you’ve noticed in the space, since you’re so close to it and covering it often? 

Brian Colello 

Yeah, exactly. And I think it all starts with AI at this point. That’s the big one. I think AI—and we could start with OpenAI and ChatGPT—I think it’s just such an interesting technology, not because of just what it is—oh, it was a really cool model when it first came out and very powerful—but I think even the average consumer could see the path of where this is going. 
That this chatbot is only going to get better and better in terms of giving us answers, giving us searches, building more for us. You could see the offshoots from there that are going to specialize. 

So I think it caught everybody’s attention—consumer and enterprises—and that’s led to this massive boom in spending. That’s the first one. 
And we get a lot of, let’s say, questions or skepticism of how big will AI be and what the returns are—as if people think that just the chatbot, “Oh, it’s only a chatbot, it’s no big deal.” 

Think about how important just even the chatbot is. That’s extremely impressive on its own, but that’s still just a piece of what this does. 
When we think about the typical enterprise and already how it’s changing software coding and customer service and software and data and how it’s analyzed and how it’s being built and the support you get with any new tool like that—that’s really impressive work. 
And again, we’re just on the cusp of that. 

If you think about—we have another ecosystem of startups that are going to be built. When we’ve discussed 4G and 5G, the biggest winner of that technology is Uber. And it has nothing to do with building phones or building telecom networks or anything like that, but that product wouldn’t exist on a 3G network. 
We just wouldn’t be able to hail car rides the same way. So they are the winner. And there will be winners based on all this AI technology that we don’t even know yet—but that’s coming. I think people are expecting it to be here today, and it’s not. But it’ll be here soon. 

Again, I think we’re going to see a lot of software improvement. 
You could even think about this chat we’re having right now—we’re only a couple years away from having a real-time transcript and images coming up, or I mention a name and the ticker pops up. There’s just so much improvement in communications as well, but all areas of software—we’re seeing that. 

And then really the moonshot projects are probably the really most exciting piece—so like robotics, autonomous driving, drug discovery. There’s so much more processing power needed to power those sorts of advancements that it’s really exciting in those industries. 
Because now they’re not even all the way there yet—of having as much processing power and AI models for something like drug discovery—but they’re going to get there. 

And I think AI is interesting because you have—there are a lot of companies involved directly and indirectly. We discuss Nvidia, and again—they are in charge because they have the best chips for AI, and they also have the software on which it’s all built. And even if AMD and others come close, they still have the best networking and connectivity. Because really, we discuss GPUs for AI as if Nvidia’s using one chip. That’s not really the case. They really glue 10,000 of these chips together to train a model. And these training runs cost easily millions of dollars, maybe even hundreds of millions of dollars. So it’s that connectivity that matters. 

For every Nvidia GPU board, you have the memory on it—it’s manufactured by a company like Taiwan Semi. You have the chip equipment makers that make the equipment to enable it all. So it’s really propping up the entire industry right now. So again, AI is the big one. And then I mentioned my other two favorites that are more in my coverage: just automotive and IoT. And really, it’s just smarter devices across the board. So that is a slower, steadier growth rate. It can face headwinds depending on where we are in the macro economy. Certainly, it’s hard to grow really well if car sales are plummeting—for example, like they did during COVID, or I think there’s concern that they might again with tariffs coming around in the automotive space.  

But in general, you’re still seeing more bells and whistles. There’s so much room for more safety. And I think we’ve all known for the past decade that the autonomous vehicle is coming—that self-driving car. We see it with Waymo. It’s out there. But more and more cars in everyday life—all those advancements getting to an autonomous vehicle—they’re going to come into the regular car as well. So your basic sedan, your basic SUV—it’s going to be smarter and smarter five years from now, ten years from now. And again, that’s all semis. And your house is going to get smarter, your factory is going to get smarter, and you’re just seeing that advancement. So software and semis are really driving this change in a lot of industries. 

Cassidy Clement 

I remember back in my college days, having to write a term paper on quantum computing. 
And I remember going through the idea of harnessing that power. Because that is, a lot of people—at least theorists in the area—that’s the main, say, hiccup for now: is the processing power. Trying to find it in a sustainable way, trying to make sure that the power can actually be utilized now, not frying itself, accurately being used for something a little bit more than generating an image of your cat with sunglasses—things along those lines, yes. 

But when people initially think of these companies—you’d mentioned AMD, Nvidia, Intel—we could start to bleed into other areas that semiconductor is exposed to: Samsung, Apple. What are the ways that people could invest in this industry? We all know about stocks, but are there other ways for people to get exposure within their portfolio? 

Brian Colello  

Yeah, stocks are the obvious one. And again, you mentioned some of the large caps. There’s a lot of high-quality—let’s call it—in the mid-cap space as well, particularly if you’re thinking about broad-based companies—maybe not as exposed to AI, but with good exposure to the car and industrial and things like that. So there are a lot of names there. 

Obviously, you can invest in options off of that as well. For us, as equity analysts, I focus day-to-day on the stock piece of it. There are certainly ETFs out there that track baskets of stocks. I would encourage anyone that is looking at that to look at sector weighting and cap weighting and things like that, because Nvidia is such a big piece of the semiconductor industry now. Some are—some have caps, some don’t. That probably will matter to performance one way or the other. 

And then there’s corporate debt out there. So in the fixed income space, these companies do issue debt in order to pay dividends and make deals and things like that. 
My day-to-day is looking at the stocks, and that keeps me busy enough. 

Cassidy Clement 

Yeah, I’m sure it’s never a calm day in the chips or semiconductor space. At least that’s what I found in all the news I follow. 

Brian Colello 

No, it’s— 

Cassidy Clement 

So when you’re adding—in that vein, actually—when you’re adding or considering adding any of these semiconductor-directly-exposed or indirectly-exposed products into your financial strategy, what are some things that listeners should think about? Are there certain pros and cons that are associated, or is it mainly focusing on the manufacturing and the geopolitical risk that we talked about in the beginning? 

Brian Colello 

Yeah, that’s a good question. So geopolitics always matter. And semis is top of mind—that’s for any investor in any industry, right? If you’re investing in bank stocks, you have to worry about regulations, or healthcare, or energy. And so now semiconductors fall into that, whereas that wasn’t an issue. 

So that’s new for me and new for us in that industry—of who will be allowed to sell chips wherever. Right now, Nvidia and AMD aren’t allowed to sell anything decent in China. I wouldn’t buy those names hoping for a turnaround there, but they should grow nicely in the U.S. But certainly, they’re going to take a hit in 2025 from that. So geopolitics is one—that is now top of mind and matters. 

I’d probably say the biggest one for the industry is: cyclicality matters. Thinking about whether you’re buying during an upturn or a downturn. At Morningstar, we spend a lot of time thinking about long-term competitive advantage. We take a long-term view with how we value companies. 

So we usually recommend buying these names on downturns—when things are ugliest—because, exactly to your point, chips are such an important part of the economy. They will bounce back. Any downturn in these businesses is not because chips have been replaced by something else. You tend to have big inventory pulls one way or the other, and you get the bullwhip effect on the way down—where a little shortfall in demand in a few different places leads to a bigger shortfall in orders for these companies. 

So yeah, it’s always important to know where you’re buying—where you are in the cycle. 

That’s a common question we get. Some investors will say the industry is less cyclical than in the past. Maybe that’s true, because I suppose the cell phone era was violently cyclical maybe 20 years ago. But we just went through a brutal downturn in 2024—coming off of post-COVID. 

Any time actual demand falls short of forecast demand, you’re going to get a down cycle.   And quite frankly, nobody knew what to order or how much to order during COVID, and so you had some pretty violent swings there. 

Nvidia is a unique case. Again, we’re so early in AI that we’re not at a cycle yet. There will be cycles at some point for Nvidia and for AI spending, but we’re still just in—we’re at the upturn of the first cycle. We’re not anywhere near maturity yet. So that’s something that’s important for investors to keep in mind. 

The third is that, again, this is still, for the most part, a high fixed-cost business. And so gross margins are still very important—even if you outsource everything to Taiwan Semi. Gross margins are a symbol of your pricing power, your ability to fend off competition, to win key designs and maybe stay in those designs. 

We like companies that have high switching costs—where you’re in a device, and you’re in that product for the life of it. And so a car—if you win that design—you might be in that car for 10 years. If you sell into an iPhone, you’re in that iPhone model for maybe two years, three years, and then Apple’s redesigning it. 3So those wins get very price-competitive, and you’re not in it forever. 

It’s much safer, much more lucrative to be in a satellite, a car, a piece of medical equipment—even like an electric toothbrush. You could sell that, and that’s not going to be redesigned too often. So that’s a big benefit. 

But if a company does have high manufacturing, it has high fixed costs. And so revenue is really all that matters, because if you have that shortfall in revenue, they can’t shrink the business or downsize the business or cut costs enough to offset it So you get a significant drop in earnings on the way down—and vice versa. 

And again, why we like these names in downturns—when things look really ugly—you could get massive operating leverage on the way back when sales recover. 

And we’ve seen over time—when you have these downturns, like the Great Financial Crisis, like COVID—semis are one of the places where you see green shoots faster than others. Because all of this production—if you want to start ramping up cars, you ramp up production of your smart gadgets or appliances or whatever—you need to order those semiconductors for production. So the semiconductor orders come first, before all these things start to hit the market. And so that’s usually one of the first signs that we’re bouncing back as an economy. 

Cassidy Clement 

For sure. When you’re starting to look at these types of companies that are making a piece that is integrated into many other things—I know that sounds a little abstract—but just to think about it that way: in simple terms, the market cycle, keeping an eye on their competitive advantage, and how crowded it is in the space, and what their financial metrics are looking like—those are things to keep in mind for a lot of investments. 

But for this specifically—this area with semis—as you said, there are different ways to look at a potential investment. Am I investing in the concept, the direct product, or am I looking at what makes up the guts of this? Is this something that can be extrapolated to other items? 

It’s a little bit more of a deeper thought—a complex school of thought—when you’re looking at something. But if you’re able to do it that way, as you said, you can look at these companies a little bit more for not only their products, but the services that their products provide to many other companies and items—throughout, especially, the United States economy, but now very much so globally. 

Brian Colello 

I would say that the end markets matter probably more than the products themselves. We’ll see press releases of a new analog chip or microcontroller comes out—but look, if the car market is struggling, that chip isn’t going to sell if the carmakers aren’t building cars right now. 
So the timing of that—or the importance of that chip—doesn’t matter. 

I suppose that’s the sort of thing that’s obvious to me and maybe not to the average investor, but—yeah. And vice versa—if the phone market is down, if the PC market’s down, if cars are down—that’s going to matter. 

Again, think about Nvidia—like, Nvidia might have the next greatest product in the world coming out—and they probably do. But if Microsoft’s going to cut spending, and Amazon and a couple others—it’s not going to matter all that much. The end-market demand and what’s being built is coming. And that may or may not tie to the overall economy. 

So for something like cars—yeah, maybe there’s a slowdown, and all those content gains are going to be not as growthy as what we would’ve thought if there are fewer cars being sold. Or maybe because of CE, there are fewer PCs being sold, or something like that. But in something like AI—it seems like that spending’s going to happen either way. And interest rates at 3% or 5% aren’t going to matter very much because it’s such a new technology. 

So yeah, absolutely—the end market and the technology matters as well. I’ll say sometimes, other times a little less. But it is tricky, and it does make it exciting. 

Cassidy Clement 

Yeah, there’s definitely a lot to consider. But thanks so much for joining us today, Brian. 

Brian Colello 

Yeah, happy to be here. Thanks again. 

Cassidy Clement 

Sure. So as always, listeners can learn more about an array of financial topics for free at interactivebrokers.com/campus. Follow us on your favorite podcast network, and feel free to leave us a rating or review. Thanks for listening, everyone.