Is NVIDIA’s High-Margin Machine Sustainable?

0:00Everybody knows that Nvidia sells the shovels and pickaxes for the AI gold rush. But what happens when the people that are buying these shovels and pickaxes are just selling them to each other? Today, Andrew and I are looking under the hood of Nvidia's massive numbers to figure out who is actually buying these chips and whether the demand for these chips are real or is just some sort of AI hype. We're asking some really big questions. Is Nvidia the most important business of all time? or are we just living in an AI echo chamber? There's a huge misconception that to start a business, you need to invent some revolutionary product.

0:37But the truth is you really don't. Some of the best businesses start as a simple side hustle, like selling a craft you make on the weekends or turning a hobby into extra cash. For a lot of people, the real hurdle isn't the idea. It's the technology. Figuring out how to actually sell online is where a lot of folks just give up. That's exactly why you need Shopify. Shopify is the e-commerce platform responsible for millions of sales worldwide. It handles all facets of your business, your online storefront, your inventory management, and your point of sale. So you don't have to juggle 10 different systems.

1:12One platform is all you need. You also don't need to be a tech expert. Shopify templates and AI tools get you a stunning site up and running fast. No coding needed. And because Shopify handles the setup and checkout, you have more time to focus on actually growing your business. If you're ready to hear the of your first sale today, head over to Shopify.com slash beginners to start your free trial. That's right. Start your free trial at Shopify.com slash beginners. That's Shopify.com slash beginners. Support comes from Wise, the smart way to manage the currencies you need around the globe. Fed up with losing out to hidden fees when you send money abroad with your everyday bank?

1:55Choose the smart way, WISE. You can count on the exchange rate you'd usually find on Google. No unwelcome surprises. Plus, ditch that where's my money feeling. Most transfers arrive in under 20 seconds. Join millions saving billions on hidden fees. Be smart. Get WISE. Download the WISE app today. T's and C's apply. You're tuned in. You're tuned in to the Investing for Beginners podcast. Investing for Beginners podcast. The show for the long-term investor. We cut through the noise to focus on what works. Compounding, discipline, and the conviction to buy wonderful businesses and stick with them.

2:40Your path to financial freedom. Start now. And welcome back to the Investing for Beginners podcast, everybody. My name is Stephen Morris and he is Andrew Saylor. And today we're doing a under the hood check of NVIDIA. I'm just going to throw some of these insane numbers at you real quick, Andrew. 2024, 126 % growth. 2025, 114 % growth. And it just, it's not slowing down, it appears. Their current PE is 34 and their forward looking PE is down slightly to 22. I mean, this is insane, dude. Like, have we ever, have you ever seen a company with growth numbers like this? Usually these are numbers you'll see from a startup, but this is one of the biggest, if not the biggest in profits company in the entire world.

3:39$4 trillion in market cap. They flirt with$5 trillion in market cap. biggest company in the world from a profit standpoint and growing like a startup. Just the numbers literally just, it's like, what world are we living in right now? So what does the PE tell you? Because like I said, it drops about 12 points on the forward PE from its current. I mean, what is that telling you? That they're expecting it to continue? Yeah, basically. So if the latest quarter had this huge explosion in growth for revenues, you're basically expecting that to sustain over the next four quarters. So yeah, a huge gap between the fact that they're already just insanely profitable and then expect whatever step up change in growth they had to continue over the next four quarters.

4:34it's it's absolutely like i'm like words escape me because i just don't know how else to describe it you can use all the all the the big words and it's all accurate so i guess i mean when you say all that like the question that comes to my mind is this a titanic waiting to happen um because i I mean, we've talked about it in the past on this show. You can never look at a company as being too big to fail. Because when you start to do that, oftentimes they end up due. I mean, how can we look at this and even begin to garner?

5:21Is this safe? because I mean, you know, for example, you know, the, the Indy 500, because I live in Indianapolis. Um, I remember when I was a kid, like everybody was so amazed when they got to 200 miles an hour going around the track. And then you fast forward 20 years and now they're intentionally putting governors on the engines of these cars because they could go 300, but it's not safe to go that fast for the drivers, at least not right now. So, I mean, is that something we, is this safe for, for a company to grow like this? So there's so many, I started peeling the onion on this and like, it just, there are just more and more layers.

6:10This is like a massive onion. And I thought I was just going to be able to just scratch the surface here. Um, so I don't know, like, I'm hoping it's instructive because the fact that it's a semiconductor company, the fact it's in GPUs, the way it's in the ecosystem of data centers and AI. This all has to be part of that conversation. And so I hope it can be instructive and like, yeah, I get it if people don't want to understand about semiconductors or all the tech nitty gritty behind it. But if you want the answer to how to think about NVIDIA, you can't just not ignore all the parts that have brought NVIDIA to where they are today.

6:53Yeah, I mean, I don't know. It still just feels insane to me that, like you said, they're growing at a startup speed, and yet there's still this massive, massive company. And like you said, they're hands deep when it comes to all the GPU stuff, all the data center stuff. are there any disruptors that you've found on the horizon for nvidia so nvidia and amd historically have always kind of gone back and forth with each other both of them have been the leaders in gpus which is a type of semiconductor that took over from cpus intel was the last nvidia they've had a decades-long run where everybody was buying CPUs, and we all relied on that chip.

7:45So whether you were buying a laptop or a desktop, Intel was the king. So when NVIDIA took that over, and the fact that we need parallel computing, that's what's made GPU just so utterly invaluable, and then turbocharged with AI and the fact that the way that the large language models work, they all need these GPUs. So disruption-wise, it's really hard to see anything on the horizon. So if you look from a moat perspective on NVIDIA, I don't think anybody's arguing that their moat is so far ahead. And we could talk about all the reasons why, and we could get into the tech why. But they're just the best solution, period.

8:33And what's crazy is they're coming out, So some of this 80 % growth for the quarter, and then the higher expected growth is NVIDIA is on a one-year product cycle. So they're releasing something every year for these data centers so that they can do AI better. The next generation, this Vera Rubin, makes the cost of AI tokens 10 times cheaper. So we're not talking about like, hey, clearance, 25 % off. We're talking about orders of magnitude improvement on the cost of generating AI tokens. These are all benchmark numbers, marketing numbers. Obviously, everything's so complex and detailed and all of that.

9:24But they've gotten to this level that they've gotten to because their technology is so strong, because of this ecosystem called CUDA that, like an iPhone, the iPhone gets better every time we update it. So they've done the same kind of idea with their GPUs. So it's just really just this type of value that they've provided to technology that maybe pissed some people off, but they deserve all the profits they've gotten. And then the other factor to this, which I think, for me, I didn't really understand, but now I understand, is basically because they design semiconductor chips, they're kind of like a software company.

10:13It's not a one-to-one thing, but the margins are super, super high when all you do is design the semiconductor and then you give it to TSMC and they do the hard manufacturing. So even to get to like a hundred billion in profit for a year, they haven't needed like crazy amounts of revenue. So that's where like some of this just crazy profit growth has come from and been enabled by the fact that their margins are so high. So like McKesson is a company I own. They do distribution in pharma and medical they do like an insane volume of revenues like 300 billion if you look like a cvs or a walmart 500 billion 300 billion like huge chunks of the economy run through these big companies nvidia doesn't need that scale to get to 100 billion in profits because their margins are so high.

11:15So that's kind of why everything in my gut that makes me human says what NVIDIA is doing shouldn't be possible because of everything you're saying, right? Like the Indie 500 going to 200 miles per hour. There are limits. We all understand there are limits to economics. There's limits to physics. There's limits to the world. But everything that has come up to now to put NVIDIA where it is is because they're in semiconductors, because they are the king of semiconductors, because semiconductor design is high margin, and because of everything that AI has just been absolutely insatiable for. Do you know how they got to just being the king of the GPU?

12:00They've always developed GPUs for gaming, and that's kind of been their thing that their company was known for for a long time, for graphics processing and gaming. And so I saw a video the other day talking about the backstory of Jensen Wong, their founder. And basically, he's had this philosophy of, I'm going to build the tech even if people don't need it. And so they just kept improving on it and kept improving on it and kept improving on it. And then the greatest thing ever happened to them was AI. So just this like sheer determination to just continue to improve, even though the ROI wasn't there is literally how NVIDIA is where they are today, which is kind of interesting.

12:44So transitioning from the GPUs, which was kind of their springboard to what you said about AI. Now they've grown into like this data center revolution of where they're helping support and grow those as well. Like, I've seen companies grow, and it's not like you just throw money at a company and you're like, here, go grow. It requires a massive amount of infrastructure, logistics, manpower, personnel, management, all these things. and yet somehow and like you were saying, it seems impossible because it blows my mind that they went from you know like you said the hard components of gaming and chips and all that stuff to now they're helping revolutionize these data centers that is insane an insane level of growth, not from just a stock standpoint, but just personnel out.

13:52Yeah. And that's what's scary about this AI stuff is how scalable all of it is. I was looking at some numbers last night and this is from Gemini. Again, ironic that I'm using AI to research AI, but some of the best software companies in the world took, you know, five years, six years, seven years to get to a billion dollars annual run rate of revenue. anthropic did it in two years and anthropic has gone from a billion dollars in revenue this is according to a bloomberg report very recently to about 60 billion 65 billion so huh like what you know i mean like so that's just the thing like um that this technology has been so insane and there's a there's a term for it but um basically the marginal cost to add a user is not super high and so that's why it breaks the laws of normal compounding or normal business commerce like normally usually you have to hire more people to to uh generate more revenue generate more you know build a sales team uh build your inventory build factories but ai model like ChatGPT or Clod, Anthropics Clod, it's already out there.

15:21And then they're just, you hook up your API and it's like they're a digital toll road kind of. So it is weird. And then again, NVIDIA being like the design aspect of these chips. So here's something that you won't hear anybody say, and I'm just going to say it as an andrewism. but there is some weird things going on with price right now. If you look at memory stocks, remember we did that, that, um, end of year breakdown or, or we did like a 12 month screen to see who are the highest performing stocks. And it was like Micron SK Hynix, all these memory tied, uh, semiconductor stocks. And so, um, what's, What's happening right now with those is there's an oligopoly of Samsung and Micron and SK Hynix.

16:17Outside of those three, you can't get memory chips from anybody else. So that's why the prices of iPhones are going up. The prices of laptops are going up. People can't get gaming PCs and they're all upset about it because all the data centers are buying up all this memory. And so my Andrewism just thought is maybe some of this explosive growth, because that's what's happening in memory is it's not that they're necessarily pumping out 10 times more chips than they were before, but everybody's clamoring for this, that it's like a feeding frenzy and people are just bidding to be at the front of the line.

16:57And so prices demand at its finest. Yes, and I wonder if some of that's happening with NVIDIA right now. Interesting. How dependent is NVIDIA on companies like TSMC? Because, correct me, and I don't know this for a fact, but correct me if I'm wrong, doesn't NVIDIA rely on TSMC for their silicone work? Yes. So, I mean, something happens with China and Taiwan or something like that. I don't know, some sort of disruption happens to their supply chain. How dependent is NVIDIA on companies like that? So everybody is dependent on TSMC. Even China is, because everybody in China with their new iPhone, all their different smartphones that they have.

17:49So all tech runs through TSMC. It's truly insane. are there any other key components like tsmc nvidia relies on that you know of um like by extension asml so when we say like the whole world relies on tsmc what what people are referring to is like cutting edge the the bleeding edge of the most advanced tech it all runs through there and then as tech matures, then you have more fabs that can kind of pick up some of the pieces there. But yeah, definitely TSMC, ASML. I would assume they're software providers, so probably Cadence, but I don't know for a fact what their mix is on that. Gotcha. I was just curious because, I mean, we've talked about before that that could be a potential danger when looking at a company.

18:47I can't remember back then the example you used. It was some sort of cloud company, I think. Like their sole, they were the sole provider of something for Apple. Or Apple was their sole, like 80 % of their customer base or something like that. And so I remember you pointing that out as like a very potential red flag that you got to be aware of. So I was wondering if we saw anything like that with NVIDIA where it's kind of the opposite disruption. If something happens to this company NVIDIA is kind of screwed. Yeah. I mean NVIDIA almost plays that part of Apple in this scenario. They said in their 10k that every single major cloud provider buys from NVIDIA.

19:41The more I've gotten to learn about Bitcoin and start to dip my toes with it, the more I realized some of my preconceived notions were incorrect. For example, I don't have to be all Bitcoin or all stocks. I can learn very instructive lessons about assets, currencies, and investments from the very long-term history of the world, and I don't have to be a speculative trader or radical enthusiast or even somebody who frets about the volatility when I have the right tools to set up a prudent allocation for my finances. That's why I use Cash App and love the recurring feature, which automatically takes a percentage of my paycheck and places it into Bitcoin, which I can spend at any time on the app or continue to let it sit and potentially grow over time.

20:21If Bitcoin is part of your long-term strategy, consistency matters. Cash App lets you set up automatic Bitcoin purchases with AutoInvest so you can build exposure over time instead of trying to time every move. You can also use roundups or pay them Bitcoin to make Bitcoin part of your normal money flow. Automatic Bitcoin purchases on Cash App have zero fees and zero spread, which matters if you're buying regularly. Download Cash App today. Visit our link in bio. Bitcoin services by Block Inc. See the Bitcoin disclosures at cash.app. September is World Alzheimer's Month, but most people never check their brain health until something's feeling off or wrong way down the road.

21:03I wanted to stop waiting and look at my own data ahead of time. I highly prioritize long-term cognitive health. I mean, you can feel everything going right in your body, but if you've already set yourself down the road mentally that you don't even realize you're on, it can be difficult or impossible to recover later on. And I wanted to know whether it's just a bad mental foggy day or if it's a sign of something for the future. Your focus and mental health leave a data trail in your body and function tracks it. Did you know that these core biomarkers are tied to brain health? Omega-3 index fuels brain cell membranes and is tied to focus and cognitive performance.

21:36Function actually helped me find out I was deficient in this. Homocysteine, when high, it can be linked to brain fog and cognitive decline risk. Plus, Function members can add on brain-related add-on tests like Alzheimer's detection tests, blood markers that can flag Alzheimer's risks years before symptoms. We're talking about life-altering signs that, if discovered early, can put your life on a completely different course than you would have otherwise been on. Check your brain and health the way I do. Function provides 160-plus lab tests for$1 a day and member pricing on advanced imaging. Join at functionhealth.com slash beginners and use code BEGINNERS25 for a$25 credit.

22:12What's it take to be a mentor with big brothers, big sisters? It takes 30 minutes and a cup of coffee. A video game. Best two out of three. It takes sharing thumbs up emojis, clapping emojis, and loaded nachos at halftime. That's how you make an impact on a young kid's life. So go on, share a few hours a month, because mentorship may sound huge, but the truth is, it takes little to be big. To learn more, volunteer, or donate, go to BeBigNow.org. Let's say that again. Every single major cloud provider like these hyperscalers, they all buy from NVIDIA. So it's like either buy from NVIDIA or get left behind is basically the idea.

22:59That's nuts. I mean, that is just, it's crazy. So kind of going back historically, you know, because maybe some of our listeners aren't. Can you just kind of break down like the history a little bit? Because I know you've done a massive amount of research into like this type of computing, you know, the silicone chip, things like that. Can you just kind of break that down so our listeners can kind of understand why this chip is so. or their their gpu is so revolutionary and insane you want to go back to like 1873 we could do that can you uh kind of if you can't do it but okay so um the thing to it's just weird uh semiconductors have always been very cyclical and um cyclical yes partly because people get really excited and then they overbuild and then there's a bust period.

24:03Lots of boom bust, boom bust, boom bust throughout the history of semiconductors. But there's also real world reasons why there's boom bust cycles. So to build a fab, which is what TSMC runs and owns, can take years. So the cycles are very long-term in nature. They have been historically. And so you've seen big technological advancements get tied with semiconductor cycles. But there's other reasons too. So for a long time, because we've owned TSMC since like 2022, I think. Back then when I first bought it, I remember almost like three-fourths of TSMC's chips were all going to smartphones. phones um iphone was it was like apple was the one directing the chip industry in a lot of ways and all the other chip makers were all trying to make apple happy apple was the king um nowadays data centers are are the king and nvidia is the king but um whether you're looking at memory or whether you're looking at cutting edge semiconductors it has always been boom and bust And so you can't separate what's happening with these AI buildouts from semiconductors.

25:30They're linked. And there's a great book that people are more and more talking about, but it is called 1873. I haven't read the whole thing, which is very on-brand for me, but I've read enough of it where I'm going to talk about it. and uh we have seen like crazy build-outs like this before from a perspective of like percent of capex um because i'm kind of jumping ahead here but basically amazon uh amazon microsoft alphabet and meta is also part of this and then spacex also part of this i don't know what the number is it seems like it's always changing but like 700, 800, 900 billion, a trillion, whatever it is in CapEx.

26:18We've seen similar levels to this in the past, believe it or not, in CapEx as a percent of GDP, but it was all the way back to 1873. And that's when they built out the railroad system. And so this massive build-out is what has made this boom in semiconductors so extreme. But the fallout to any cyclical, super cyclical boom is it doesn't last. And that's where some of our gut instinct starts to kick in as we understand it's not going to last. So the question everybody asks is how much longer is this going to go? Is this going to go like 1873 where we're looking at, I don't know, three, five years of just crazy hype before it crashes?

27:10Or are we looking at something that's about to stop? Because things have really accelerated in the last, even I'd say, couple months. the latest wave of earnings for Amazon, Google, Microsoft. That's when we saw their cloud backlogs go from, each of them jumped to a couple hundred billion dollars in cloud backlog, each of them, in one quarter. It's like, what? So the backlogs have gone parabolic. The CapEx was already a little bit parabolic. and then obviously you have the Anthropic and OpenAI IPOs which are everybody's still waiting on so we can't just think as much as it's tempting to think like oh you know NVIDIA semiconductors this is the new reality right I think it would be naive to say this time is different like we're going to need these semiconductors forever because if you look at the context of history we have had crazy build outs like this before and it's not going to last forever it's very very unlikely to last forever would you be impressed if i told you that when you started talking about 1873 my guess was was steel in railroad that yeah i'm impressed yeah i didn't get a chance to throw it in there i didn't want to interrupt you but like i really wanted to though because like i I was certain that's what you were talking about.

28:42That was my first guess. So railroad steel, I think they go hand in hand, but I don't know if that's cheating picking too. But I mean, in the opening, I compared it to, you know, the picks and shovels of the gold rush. And so I guess the next logical question I have then, are the companies buying these picks and shovels? Are they finding gold? Are they, you know, is the ROI there for them to sustain this for several years? Or is, you know, are the doom and gloom bubblers correct in saying that this is just a simple AI bubble that's going to burst just like the dot-com bubble did in just a year or two?

29:34yeah this is what everybody in the media has been talking about non-stop for the last year right and seems to get just more and more more and more people who are coming up to talk about it I don't know like that's the thing so I'm watching to see what happens with open AI and Anthropic

30:00there's a lot to the AI piece that's very nuanced.

30:07Large language models, I think a lot of people have gotten pretty sick of them. AI content, AI slop, we've all seen it. We all can recognize it. But also, which is... A lot of people already know this, but the things that Anthropic is doing and the way that Claude can code and the way you can kick off these agents to go build things for you. and literally just through a text message have an app get built and an agent can go for 15 hours or however long it is. You really have a lot of potential. So you have a lot of potential that companies are investing into and then we have a lot of real revenues that OpenAI and Anthropic are generating.

30:57But I think the big question is with Anthropic at around 65-ish billion run rate. OpenAI was somewhere around like 35 billion. So about 100 billion is flowing through these new AI models based on what people are saying. A lot of other people are saying that in order to make the CapEx profitable, that all the big hyperscalers are doing, OpenAI and Anthropic need to get to the next 10x. So that part I'm not so sure about, but it is fascinating to consider, which sucks if you're a software engineer. A lot of software engineers have gotten laid off, but if your AI agent can take the place of a software engineer and however many software engineers you have at your company, that's all money that used to go to an engineer that can now go towards AI.

31:59and theoretically produce 3x, 5x, 10x more than what that software engineer was producing. So that's a pretty big TAM. I don't know how many software engineers there are, but for a long time, I've kind of thought of AI just as like this gadget, you know? Like, oh, it's just people who use ChatGBT or whatever. But there's a lot of things going on in the coding space. And then also like Visa MasterCard, if you've listened to their earnings calls, They're talking about agentic commerce, and that's a pretty cool idea as well, which will need a lot of AI. So there's a lot of fringe cases of AI, but it seems like people, enterprises, companies are investing based on where the puck is going rather than where the puck is now.

32:48And it sucks that so many people are getting laid off and then that money is being reallocated towards AI. but is that how the next winners of this technology are going to be decided like you have the infrastructure and then a lot of the money goes to the AI models and then does the next value unlock go to the companies who are using AI in the right ways I don't know nobody knows but I think one extreme to say that this is all a mirage I think is is So to take the unpopular middle road here, one extreme saying that AI is hype bubble going to crash, I think is overblown. But also saying that AI is the best thing and we're in a completely new reality is also overblown.