MI267: Investing In Chip Companies: Deep Dive Into QCOM, TSM, TXN, AMD, NVDA w/ Nicholas Rossolillo

0:00You're listening to TIP. The demand for chips is skyrocketing. So I'll throw some figures out here. $470 billion was the global purchase of chips, the chips themselves back in 2018. And fast forward today in 2022, global chip sales were just shy of$600 billion. So it's a huge, huge market, just the chips themselves, almost$600 billion. And it's on its way to$1 trillion a year. Some estimates point to it being maybe$1.3 trillion every single year by 2030. So if you're looking for a growth industry for the next seven, eight years, that's a compound annual growth rate of 7 % to 10%, depending on if you want to use the$1 trillion or the$1.3 trillion estimate.

0:46It's a huge market.

0:50On today's episode, I chat with Nick Rizzillo. Nick is the founder of Conkinis Financial. He's an investment portfolio manager and host of the Chip Stock Investor Show on YouTube. I brought on Nick to chat all about the semiconductor industry and to learn more about the best companies in this space as today our economy can be described as a chip economy where microchips have been applied to many technological advancements in different industries. And so Nick gives us an overview of the chip market, which industries are expected to drive the most growth in demand for chips in the future. He explains what the different types of chips are and which companies produce which types of chips.

1:34And he walks us through a number of chip companies like NVIDIA, AMD, Qualcomm, Taiwan Semiconductor, as well as Texas Instruments and shares his thoughts on which companies he believes are undervalued at today's prices and so much more. So without further delay, I really hope you enjoyed today's episode. Millennial Investing by the Investors Podcast Network, where your hosts, Robert Leonard and Rebecca Hotsko, interview successful entrepreneurs, business leaders, and investors to help educate and inspire the millennial generation.

2:16Welcome to the Millennial Investing Podcast. I'm your host, Rebecca Hotsko. And on today's episode, I'm joined by Nick Rosalillo. Welcome to the show, Nick. Hey, Rebecca. Thanks for having me on. Excited to do this. Thank you so much for coming on today. I've been really interested in learning more about the semiconductor industry. And as I was doing my research of who I should bring on, your name kept coming up. You have such a great YouTube channel, Chip Stock Investor, which I've already learned a ton from. And so I'm really interested to know how long have you been investing in chip stocks and how did you get interested in this?

2:52almost right from the get-go i started buying chip stocks when i started buying individual stocks like 17 years ago 05 06 a chip stock got added to my portfolio right off the bat so at that time it was i bought a company called on semiconductor which i think we might talk a little bit about later that was my first pick i had no idea what i was doing In hindsight, I should have bought like NVIDIA or Texas Instruments and just sat on it. But I bought OnSemi, which at the time was very commoditized manufacturer of power management chips. But yeah, I've been buying chip stocks for a while. I've always been interested in electronics.

3:34And so when I started investing, I got interested in semiconductors. Yeah. So I want to dive into a ton of the stocks that you're interested in today. But first, I think it'd be super helpful to just go through an overview of the industry as a whole, because today our economy can kind of be described as a chip economy because microchips have been applied to so many technological advancements and different industries. So can you walk us through a brief overview of the semiconductor industry? How big of a market is it today? And which sectors use the most chips? Yeah, it's massive. And I think we're at a really interesting inflection point in the economy right now.

4:15So obviously, semiconductors are not new. They were invented like 70 years ago. But we're reaching this point in time where I think they're coming out of the office. Historically, you have maybe like an army of office workers using computers, and maybe they have some computing network shoved in a closet somewhere. But we're reaching this interesting point where both the power of the chip, or I should say the computational power of the chip is increasing to a point in intersecting with the affordability of that computing power, where it now makes sense for all sectors of the economy to start using semiconductors.

4:53It makes sense. It makes sense for every sector, every industry. And again, not just office workers using computers, using PCs, actually embedding those chips into daily operations. So it could be an energy or a utility company. It could be a medical company that wants to connect its medical equipment to a network, obviously electric vehicles and the, I guess, smartphone-ification of the car, of the cockpit of the car. So all of these different things happening all at once and the demand for chips is skyrocketing. So I'll throw some figures out here, some pre-pandemic figures,$470 billion was the global purchase of chips, the chips themselves back in 2018.

5:38And I start with 2018 because 2019 was actually a down year. Remember the US-China trade war. And fast forward today in 2022, global chip sales were just shy of$600 billion. So it's a huge, huge market. That's again, just the chips themselves, excluding a lot of the software, a lot of the equipment, the raw materials, excluding all of that, just the chips themselves, almost$600 billion. And it's on its way to$1 trillion a year. Some estimates point to it being maybe$1.3 trillion every single year by 2030. So if you're looking for a growth industry for the next seven, eight years, that's a compound annual growth rate of 7 % to 10%, depending on if you want to use the$1 trillion or the$1.3 trillion estimate.

6:25It's a huge market. So that's a broad overview. Let's see, you asked about end markets as well, didn't you? Yeah, for sure. You can get into that. Yeah, you had this chart here that was a really nice visual showing smartphones actually are currently the largest market for semiconductors currently. Smartphones passed up to the PC last decade as the largest consumer of chips. So again, remember office workers, historically, the users of computers, but then everybody became a user of a computer every single day, sometimes every minute of every single day, depending on how addicted we are to our smartphone.

7:00But in the next decade, I think there's a growing anticipation that data centers will become the largest consumer of semiconductors. So data centers being the main computing unit of the cloud. And there's some different things we might get into later as to why that is. But those are the biggest end markets currently, smartphone, PC, data centers, and servers. I want to dig into the growth in or the expected growth in sectors in a second. I first want to ask you about the different types of chips, because I think that'll help us grasp this overview of the industry. What are the main types of chips?

7:40Basically, there's two broad types of chips, analog and digital. So an analog chip, you can almost think of it as just basically a sensor. Maybe if you want to compare it to like how our bodies work, our five senses, an analog chip is kind of like that. It might be a type of chip that reads a radio wave. So that would be like a mobile device. Or it could be a power chip that somehow interacts with or alters the flow of electricity through a computing system. It could be LIDAR or radar or a vision chip on a lot of new cars are getting these to help enable some advanced driver systems, or maybe it's an industrial robot that uses one of these analog chips.

8:23So again, just think of like a sensor and kind of like our five senses help us interact with the world, the physical world around us. That's an analog chip for a computing system. And then the second broad type of chip is a digital chip. And so a digital chip works with ones and zeros. So you could compare these to like the brain, our brain. And so within digital chip, you also have two basic types. You have logic and memory. So memory, it stores information for later use. And then logic, it crunches the numbers. It's figuring stuff out. And so the language that these things use are ones and zeros or bits.

8:59And those are completely meaningless, but you have countless ones and zeros lined up together. And that's kind of like the most basic alphabet for software code. So that's a digital chip. And oftentimes, these are like the most recognizable names, these digital chips. So like your Intel, primarily a digital chip, a designer and manufacturer, NVIDIA, AMD, both design digital chips. Micron is the top designer and manufacturer of memory chips in North America. So some very name brand type of companies fall within this digital chip segment of the industry. And then which of those two chips is the market growth expected to grow the most?

9:39Which has applications to, I guess, which industry? That's an interesting question. So over the last decade, it has most definitely been the logic chips, especially the digital logic chips, more so than memory. And memory is kind of a different topic I think we'll touch on later. But so NVIDIA, now the world's largest semiconductor company, as measured by market cap, massive growth over the last decade from logic. The world has this insatiable demand for these logic chips. And I think the chat GPT viral piece of software here the last few months kind of, I think, illustrates. It's not necessarily a new thing, but for the first time, I think a lot of consumers got introduced to AI.

10:26Behind software like chat GPT is a computing system designed by NVIDIA, or at least the heart of the system is designed by NVIDIA. So massive growth for those. But again, we're at an interesting inflection point in the economy now where for the next decade, analog chips, which have for a couple decades now been kind of a sleepy, kind of a commoditized part of the semiconductor industry, are all of a sudden also expected to grow at a very, very fast pace. And again, a lot of that is being driven by electric vehicles, new automobile technology, and parallel to that, a lot of similar technology used in the modern car also getting applied to industrial robotics.

11:07So manufacturing, food production, healthcare, and so on. Okay, that was very helpful because I found that when you first dive into researching chip companies, there is so much to learn about how they work and which chip applies to what industry, which company makes what chip. And so I'm glad that you kind of broke that down for us today. And I want to dive into the expected growth of those different kind of sub industries a bit more because, yeah, I found that chart super helpful to just see the projected growth. And so data centers is expected to actually grow the most by 2030 and beat even smartphone demand, which I thought was super cool.

11:47And then automotive is expected to grow a ton as well. And then personal computing a little bit. And what was the other one there? Automotive, consumer electronics, decent growth over the next decade. I mean, a market doubling in size over the next eight years is pretty impressive, but one of the sleepier parts of the market. So yeah, PC, consumer electronics, and internet service infrastructure, the slowest moving parts of this industry. Right. So then that data centers one, you touched on that. That is driven then by AI demand, essentially? It's become twofold. So over the last decade, it's been just general purpose cloud computing, or oftentimes you'll hear it referred to as HPC, high performance computing.

12:39So these massive data centers that operate huge applications. So like Meta, the parent for Facebook and Instagram, operates massive data centers. I think this is actually a little known fact, Meta is not just a social media app. They actually do some really deep, impressive engineering behind the scenes because they own their own data centers. And to operate something that can support 3 billion plus users around the world, and people using it constantly, every second of every day, somewhere on the globe, you have to have some massive infrastructure. So over the last decade, it was just the general purpose cloud.

13:18But now all of a sudden, it's AI, which oftentimes is just getting embedded into the cloud somewhere. So yeah, that data center market, I think it's going to actually surpass some estimates still think the smartphone market will be the largest consumer of chips by the end of this decade. I really firmly believe that data centers and servers, maybe we could break that down further into public cloud like Amazon AWS or Microsoft Azure, and then also companies building their own private cloud network as well, their own data center-based network for personal use. Those two things combined, I think, will outpace smartphones in the next, by the end of this decade.

13:59Some estimates actually pointing that memory chip demand will actually get sold into data centers more frequently than they do smartphones as early as next year, as early as 2024. And then logic chips, digital logic chips will catch up and in total surpass smartphones. Wow. And so, okay, so the main chips for data centers are memory and logic, you just said? Correct. Yes. So data centers are primarily logic chip driven. So you have all this information flooding into these remote locations. Maybe there's one somewhere in central BC, central Washington, Central Oregon, in the Northwest, you have all this massive information flooding into these things from the internet, or maybe from a company's private network.

14:53So again, it's like the central nervous system. If you want to compare a computing system to the human body, there is so much information flooding these things. And that massive amount of data is only getting larger as more devices get connected, more people get connected to these complex software systems and AI algorithms and whatnot, a massive amount of demand now coming from things like AI, and that's all going to happen in a data center somewhere. Let's take a quick break and hear from today's sponsors. Even Einstein had blind spots. That's why modern science is built on the idea of peer review.

15:32Investing may be more art than science, but that doesn't make peer feedback any less valuable. The tricky thing is finding qualified people who are interested and willing to help vet your investment ideas. That's why we built the Intrinsic Value Community. It's a place to connect, share ideas, learn, and get feedback. Nobody ever wishes they'd spent more time buried in spreadsheets, but connecting and building relationships with others who may be smarter on a topic than you, but who are also schooled in value investing, that's valuable. We make spots in this exclusive community available in cohorts every few months.

16:03And last time around, our 30 available spots filled up pretty quickly. If you're interested in our next cohort, which will be even smaller, you can join the waitlist at theinvestorspodcast.com slash intrinsic value community. That's theinvestorspodcast.com slash intrinsic value community. Support for the show comes from public.com. You're thoughtful about where your money goes. You've got your core holdings, some recurring crypto buys, maybe even a few strategic option plays on the side. The point is you're engaged with your investments and public gets that. That's why they built an investing platform for those who take it seriously.

16:41On Public, you can put together a multi-asset portfolio for the long haul. Stocks, bonds, options, crypto, it's all there. Plus, industry-leading yields on your cash with no fees or minimums. Switch to the platform built for those who take investing seriously. Go to public.com slash T-I-V-P and earn an uncapped 1 % bonus when you transfer your portfolio. That's public.com slash T-I-V-P. Paid for by public investing. Full disclosures in podcast description. Just like everybody else, there was a time when I was a beginner investor, and I have to say, investing is particularly filled with jargon that can just make it so difficult for new investors to understand what is going on.

17:21But it's never too late to get smarter about stock investing from the ground up. At The Investors Podcast Network, we've made a habit of studying the world's best investors. And now I'm distilling those learnings into a simple course for you or for anyone in your life who you might want to share the gift of knowledge with. With my How to Get Started with Stocks course, you can master the principles of excellent lifelong investing with valuable insights for both beginners and pros. The course covers 10 different sections, beginning with the basics of what a stock actually is and how the stock markets work, to strategies to optimize your retirement savings, how to pick great companies for the long-term, what to look for in ETFs, and how to monitor your investments, plus so much more.

18:00To begin getting smarter about investing, just visit theinvestorspodcast.com slash get started with stocks. That's theinvestorspodcast.com slash get started with stocks. And for a limited time, you can use code stocks15 for a 15 % discount at checkout. All right, back to the show. Okay, so now I want to get into your top picks. And we talked about this before the episode, it might be helpful to break it down kind of by sub industries. So let's start with cloud AI. Who do you think has the or which companies do you think offer the best long term prospects in this sub industry? One of them I've already mentioned, NVIDIA, which in recent years, I think has become a lot of people's favorite semiconductor stock, hopefully, because the business is awesome and not just because the stock prices rocketed higher.

18:54But I think that is the top pick and the market has really caught on to it. So even though by revenue, actual sales, NVIDIA is not the world's largest semiconductor company. It is, however, the largest semiconductor company by market cap. And again, it's just being driven by this expectation that the data center and how much information needs to be crunched by these data centers is going to soar over the next five to 10 years. And NVIDIA has built itself as not just a platform company. It's another big tech company. We can think of Google or Amazon as a platform business that other businesses build themselves on top of.

19:36NVIDIA is that type of company now. It's that important in the global economy. So as AI takes off, NVIDIA has the market cornered. If you're looking to build an AI system, you almost, almost, keyword I think, almost can't design an AI system without NVIDIA being somewhere in the mix at some point. So I think that's the top pick. Expensive stock, ridiculously expensive stock. And it's just because of this expectation that they're going to gobble up the most of this market in the next 5 to 10 years. So I have a trio of them, though, that make up my top pick here in data center and AI. AMD, again, household name in years past, kind of the underdog to Intel, but they've really passed up Intel technologically in recent years.

20:22And then a third one that almost no one has heard of, Marvell Technology Group. So this company makes a type of processor called a DPU, a data processing unit. And they made a string of acquisitions the last four or five years to bolster their portfolio of different chip designs and IP that they can license out. And so they're actually a key company when a company is trying to design high performance computing for the data center. Marvell is a key ingredient in that work now as well. Okay. I want to jump back to NVIDIA for a second because I'm wondering, you said they kind of have the market cornered.

21:02Why is that? What is their competitive advantage that no other chip company can do or maybe are they trying to do and have failed? Actually, so we're recording this a day after NVIDIA's last earnings report. And on the earnings call yesterday, an analyst asked CEO Jensen Wong that question precisely. And he pointed out, They do, in fact, have competition. They've always had competition. NVIDIA's secret ingredient is the GPU, the graphics processing unit, which historically is for high-end video game graphics on a PC or laptop. They figured out, though, years ago that the technology that their chips utilize can also be used as a computing accelerator.

21:47So a CPU, which in times past was dominated, a type of chip dominated by Intel, has its limits as far as what it can handle when it comes to these really, really complex pieces of software where there's massive amounts of data. The CPU can only process much of that data so fast. But the way the GPU is designed is it can kind of chop up this massive amount of data into smaller pieces and then process it in multiple threads all at once. And so it's being used as a computing accelerator. You still need a CPU in a data center. But let's say if you want to run an application like, let's say, ChatGPT, and it's been trained on this incredibly huge set of information on the internet, let's say, general information found on the internet.

22:36You can't just use CPUs. You have to use a GPU. And NVIDIA has been working on these things now for decades. And they're way, way far ahead, both AMD. AMD has GPUs, and they're also far ahead Intel technologically on these GPUs. Okay. And so what happened to Intel? Because I think I heard on one of your YouTube videos that they just kind of became complacent. But could you explain what happened with them and why aren't they farther ahead? I promise I won't rant about Intel like I have on my YouTube channel. Intel's problems really go back. I think we can go back a couple of decades if we really want to trace Intel's problems today to their root source.

23:21So back in the early 2000s, a small company called Apple was working on this quaint little device called an iPhone. And they approached Intel and asked for help in designing a chip for it. And Intel took a pass. They said, you know, basically, that's a neat little device, but that's not what we do. And so Intel completely missed the boat on mobile computing technology, logic chips, digital logic chips. They completely missed the smartphone market. And so they really kind of pigeonholed themselves basically in CPUs, which is one of the original types of logic chips. And they got pigeonholed in PCs and laptops and some data center CPUs.

24:02But again, as we just talked about, as these data centers get more complex and the software that needs to be run by them gets more complex, NVIDIA is actually ahead as all of these data centers. Data center operators say, hey, we need new equipment to run an AI algorithm. They're not going to Intel for the stuff they're going to NVIDIA. So NVIDIA kind of helped create this market as well. So Intel has missed multiple shifts in computing technology. There are some other issues going on as well, where I think they could have turned things around earlier, but they were just lazy. They got complacent.

24:39And after the handwriting was on the wall, the industry had shifted towards mobile, and then later on cloud, they weren't quick enough to get up to speed with the times. Okay, that is very interesting. I guess it just speaks to management. and I guess why good management is so important as always. I want to jump back to NVIDIA now for a second because, okay, their competitive advantage, you mentioned they are just far ahead in the GPU technology. And so there are a couple competitors though catching up. How strong is that moat, would you say? For NVIDIA, I think at this point, it's pretty strong.

25:16An interesting thing about NVIDIA is it's not just a semiconductor company anymore. They're also a software company, but maybe not the type of software that we might think of. I think we've all gotten hooked on software as a service, SaaS, over the last decade. AI requires, and high-performance computing and AI require oftentimes a highly customized piece of software. It might be dedicated specifically to a business or maybe a specific segment within a business. There is no one-size-fits-all. And so NVIDIA has done a fantastic job over the last couple of decades now, developing this software that makes its chips, its GPUs usable for the user, for the individual or the company that's buying them.

26:04So they have this very wide portfolio of software now. Oftentimes, that software is bundled together and given away for free with the GPU itself. But increasingly, the last few years, they have been branching off into more advanced software. that they can actually charge for. So again, in the last earnings call just yesterday, they actually gave the most solid quantitative number to date on their software business. And they said it's in the hundreds of millions of dollars. So that's super nonspecific. I don't know if that means$200 million or if it's closer to$1 billion in software revenue, but it's becoming a very significant part of their company.

26:43And when you combine two things together like that. Very well-entrenched hardware that's very far ahead of its peers technologically with a software business that makes it very easy for customers to implement that hardware and put it to use as quickly as possible in their day-to-day operations. You now all of a sudden have a company that is going to be very difficult to uproot from the very fabric of the economy. So I think NVIDIA has a very, very strong mode at this point. And that's reflected, obviously, in the sky-high valuation. One-year forward earnings expectations of well over 40. That's a very, very high price, even for a semiconductor company growing as fast as they are.