Nvidia [Stock Geekout Ep 4 – 22 May 2021]
In Episode 4 of Stock Geekout, we geek out on semiconductor company Nvidia with resident stock geek Thomas Thio.
Connect With Us:
Like what you heard, you’ll like a few other things we work on. Click here to see what they are.
Learn how to make 8-12% each year in Passive Income
Watch this Free Masterclass on Income Investing with The Fifth Person
Reggie: Today in TFC Stock Geekout, we’re going to explore a company that has managed to corner a part of the semiconductor market. Yes, the semicon space is very huge and complicated with multiple players engaged in highly specialized and sophisticated processes. But these guys have found a stranglehold on the market, owning more than 90% of the GPU market, aka graphic processing unit, which in layman terms is a turbo charge in the computing space, enabling compute to the highest order, which is extremely important in gaming, machine learning, data centres, crypto money, and all the fancy-schmancy latest stuff.
Expand Full Transcript
Joining me today to explore this semiconductor giant will be Thomas Thio, our in-house stock and tech geek. We explored Nvidia, exploring its supply chain, where it sits in a broader semiconductor space. What is the product use for the base case for its near monopoly and its future growth prospects versus its risk factors where in the future, tech may move away from GPU. Your phones, for example, do not use GPU. So what is stopping Apple from eating their lunch and other tech players from copying them?
For your reference sake, this episode is recorded on the 22nd of May, 2021 and released early to our community members. Our discussion today is solely for education and entertainment purposes only. It does not serve as any form of advice or recommendations. Thank you for loving what we do and empowering us financially to do more for you. Let’s geek out.
Welcome back to TFC Stock Geekout and following all the discussion with all these tech companies and following our daily live, talking about semicons. I think Thomas has a deep desire to talk about the semicons space and there are a lot of things going on. This is a very complicated space, all these different companies and there’s one company that is leading the way in a particular field, like in a big pack, in a space where Intel used to dominate this space at the microchip level.
It’s not very easy to find its own footing, but there’s this one company that for some random reason… okay,maybe not that random. They have a lot of Bitcoin miners, they have a lot of rise of gaming, so they have found their footing in this game. So what is this company, Thomas? What do you want to bring us through today?
Thomas: This is Nvidia. Nvidia you might know only if you are a gamer and maybe more recently because of the rise of cryptocurrencies and all that then you hear this name like “what is this NVDA, some weird name and all that?” What do they actually do? They are in the realm of semiconductors, but not the fully technical kind of semicons. They are towards the end of the chain.
They use whatever or most of the semicons already produced to actually make these things called graphical processing units, GPU for short. For the gamers, these are the things that is used in your PCs. You want to play your Battlefield, World of Warcraft or something like that in 4K. You need the best GPU ever.
Reggie: Extremely important.
Thomas: In these same technologies actually powers a lot of the cloud compute that you see today, which also is the same kind of technologies which is being used to go and mine your Bitcoin. So yeah, that’s why I’m pretty interested in doing this.
Reggie: Nice. Yes, and I think for a lot of people that don’t know how these work, I want to try to explain the product. Essentially, in your computer screen, there are all these pixels. There are a lot of small, little pixels and they are emitting light. But to orchestrate all these emission, where to emit, at what time to emit, what’s the intensity, you need a particular thing to orchestrate this whole thing and that is the GPU. Is that what it is?
Thomas: Actually, that is more of the motherboard. The GPU is like a turbo. You have, let’s say your main engine is your CPU (Central Processing Unit), right? Your GPU is like a turbo attached on top of it which is designed for certain kind of workload beyond a certain point, because there’s just a lot of math that needs to be computed in gaming. Because of your graphics, a lot of these pixels and then the physics engine all that, it’s just crazy stuff… as well as for Bitcoin mining, as you might know, and for compute clusters, let’s say there’s a lot of data science jobs that needs to be done. All this is just mathematical calculation at the end of the day.
Your CPU, of course can do this, but very slowly. It’s not optimized to do that. It’s only optimized to actually streamline all the operations of your operating system, your macOS, your Windows and things like that. But when you run some very intensive computation, there’s better ways to optimize the same kinds of underlying technologies. That means ordering them in a certain way, such that it computes it much, much faster, and that’s its only job.
Reggie: Okay, so it’s a turbo charge in the computer and it’s required in all the very high, intense things like computer games, like Bitcoin mining, like machine learning, all those very advanced, new edgy kind of tech requirements. Is that what I’m hearing?
Thomas: Right, right.
Reggie: Okay, cool. So then when you say that they are the last of the line, GPU are last of the line or Nvidia is in the last of the line kind of manufacturing work, can you elaborate a little bit more for us? For people that don’t understand how these semicon space actually works?
Thomas: At the top of the chain, you have things like… or rather companies like TSMC (Taiwan Semiconductor Manufacturing Company) or you have maybe even somewhere a little bit further down would be like… what is that called? GlobalFoundries. So they take all these chemicals, they take the certain minerals and all that, they put it into all these kinds of wafers. You can say that wafer fab and all that, some you hear…fabless. That means they don’t use a fabrication plant. They source all these materials together and you just put them together. There’s different ways to just produce a microchip.
That is at the start of semicons. But then all these microchips, at the end of the day, is going to be put onto a bigger piece of… bigger slab of a microchip but it’s still being called a microchip to all of us. There’s of course a technical term for it, but just generalizing here. So as you get further and further down the chain, you get into something that is… looks more like a real product, rather than something that comes out of the ground. No more chemicals… helium, boron, something. You also don’t look at… whichever. All already come up.
Down the chain, this gets into more and more use cases. They are being utilized by different parts of the world. There’s also a semicon being used in automotive industry. Things like your simple switch… on, off your ignition is also a very simple analog switch. But with your EVs (Electric Vehicles) and all that, there’s more advanced use cases and because there is electricity being used throughout the car, there’s more ways to actually design around this, also using the old analog systems. So that’s also a part of semicon.
Semicon is also producing these kind of microchips for your refrigerator, for your camera, for your phone. Other than all these other use cases, one part of this portion will go towards memory, storage of the memory where you have your hard drive and your SSDs (Solid-State Drives) and all that.
All these are what we commonly know as just storage. This is just like flash storage or USB drives and all that. Another portion would be your compute. For compute, these are also a set of microchips but it’s designed specifically for very fast processing of mathematical operations.
How is these mathematical operations being done in CPUs, GPUs? It is basically a lot of zeroes and ones doing calculations. You flip the switch, something goes to zero or to one. You flip another switch, another part goes to one or zero. A form of operators in a certain sequence will have certain patterns. The states are important to go and hold and to switch as fast as possible.
The more you can squeeze these kinds of states into a smaller space, it means you can do more computations on the fly. That is why this is very important for games especially, as well as for let’s say Bitcoin mining.
Reggie: Okay. So in that sense, what are some major things that we got to look out for? I think that there are two things that’s happening in the market these days. One is that a lot of the low level kind of chips, which are the things that are… go into your fridge, things that go into your car. Those chips are not very high computational, they’re very low level. They’re relatively high up the chain.
There’s a shortage there. There’s multiple reasons as to why there’s a shortage. Apparently, one of the reasons that it’s being flouted out there is because further down the chain where Nvidia is more at, these chips are priced much higher. So instead of producing the ones where the cars used, which are like very low level, more of the production line are optimized to produce all the way down to where Nvidia is, or where Intel or AMD, where they are, further down the line or whether is it Apple M1, all those kinds of chips.
So that’s one problem of shortage which I hope you can elaborate, and then the other part is about nanometer. When people are investing in microchips or in semicons, you always hear this thing called nanometer. It always sounds like an arm race. “This is five nanometer, three nanometer”, then I think IBM has a two nanometer rubbish… that they are flouting. Can you elaborate on these two things for us?
Thomas: Right, so let me talk about the nanometer part first then I’ll go into the first question. Nanometer refers to how much things you can squeeze within that amount of space. So nanometer is like really small. If you already know what a centimeter and a millimeter… a nanometer is even way smaller than that. We’re talking about microscopic level.
How many computation parts can you actually squeeze into that nanometer is essential because once you get to that level, you just have a lot of scale related to a certain use case. So this leads to the part about your first question. Why is there a shortage… Is it because of price or because of other reasons and all that? It depends on the use case and we also need to look at the different players that’s in charge of producing a lot of these semiconductor chips for these end use cases.
These semiconductor producers like TSMC, they are manufacturing the chips. They are doing everything possible, automotive chips, they are doing headphone chips, they are doing microphone chips. Anything, you can name it.
But all of these, it needs to go through a certain kind of flow, a process. You cannot possibly put 100% allocation into just automotive. Yah lah, EV is a boom and all that kind of stuff. You want to allocate all your… you want to allocate your whole portfolio to EV meh? No what, you wouldn’t want to do that. So in the same case, for business, they would diversify also. They would spread it across.
But of course, they have certain kind of R&D (Research & Development) arms that they can apply onto this process where they produce these semiconductor chips. This is where the nanometer comes in. How many of these production processes for certain end use cases have these nanometer capabilities? They need to configure each of these processes because it’s a certain sequence of steps in order to reach a certain microchip for end user. These needs to be very, very finely tuned. It’s not something that can be changed overnight or even a course of a few months if they want to change a production line. So these things tend to be quite stable.
Now, why the semiconductors being a shortage? It’s because suddenly, in my view, there is actually a high demand. There’s a high demand for certain kinds of chips and there’s just not enough of these being able to produce on the foundries itself, on the manufacturers itself. Why is there a high demand? Everyone’s at home in Covid and all that, they want a computer, right? They faster buy their fridge, store as much vegetables and fruits as possible. Your TV also, you want to upgrade.
Reggie: Yeah. Everybody has multiple devices these days and it’s because at home, like what you are saying, more people are buying more.
Thomas: Mmm mmm. From the demand side, there is a pool. There’s a pool from the consumer side that… “I want more things! Produce more for me.” TSMC is like, “No, I’m like fully stretched here. Holy crap, how am I going to ramp up? Okay, if you really want me to ramp up, I have to raise my prices.”
At the same time, supply side of these manufacturers, they also suddenly say “our contracts were usually quite consistent, you only ask this amount of chemicals, you only ask this amount of minerals. You suddenly ask for so much? I’m going to raise my price.” So it’s two sides… and supply side is like “eh not enough!” There’s only so much things that we can pull from the earth and yeah, that’s it. Limit stretched already.
So it’s on two sides. Basically, you see the prices increasing at the supply side as well as on the demand side. It’s a pull in both directions and basically TSMC… it can be a good thing or bad thing. They are at max capacity already. They are still trying to actually roll out more scalable ways to actually produce these microchips, but there’s a maximum to it. It will think definitely this long to actually address these kinds of new use cases as well as the existing pent up demand. At the same time, your prices are already high. They already raised it, your cost down the chain.
Reggie: Fundamentally, what I’m hearing is that there’s a very complicated supply chain. You don’t really just amp up like… oh, I want to amp up and just amp up like that. Essentially, there’s a lot of high switching costs in this game and we can talk about TSMC another day. I think there’s a lot of discussion in this whole compute, semicon space.
Thomas: So many players.
Reggie: Yeah so many players. So many guys in between, nitty gritty. I think the other day, we had a live with the guys from FIRL (Finance In Real Life). We were also talking about another player in this space. There’s so many things here. If we move down the chain towards where Nvidia lives, where they are in this business of GPU, what do they actually do in their process other than the sales portion… like on the manufacturing line? Help us get a litte bit of colour. What do they do in this space?
Thomas: Sure. So they have a certain R&D that they have ongoing for their certain products which are used in gaming for the data centres or certain kinds of minor mining workload. They know specifically already how to build it. They are just ordering components from the manufacturers that fits well with their design. They’re not actually doing the manufacturing, maybe a little bit of assembling. They’re not doing the manufacturing of the chips. It’s a little bit different.
So when these are already produced or configured together and assembled by Nvidia, these are ready to go GPUs that can be used right in other end use cases. It’s just before it hits the final customer already, like Amazon data centres. Or it could be a retail customer or maybe a distributor, like Challenger or whichever. “Here are your new GPUs. Please tell me how many people you can sell to.”
Reggie: You’ll be backed by the computer manufacturers, slot into comms, and then continue to sell up. So, essentially Nvidia sounds like a guy that is at the very end of the computing supply chain, right before it gets to the customer. Who are their customers? Do they go into direct customer or who do they actually sell to?
Thomas: They have different kinds of segments. Broadly, we can split them into the retail as well as the business. Retail wise, it’s just like you and me. Maybe we game a lot or we mine crypto somewhere. So we buy these GPUs because we don’t really know how to actually configure them. We don’t know how to build them ourselves or something. So we have them, we put them into a computer, we say “okay, this one fits our PC. We just buy, can play game very happy already.”
Reggie: That is the lay person. We really look at it like that. “Wah 这个 (this one) like very good…
Thomas: No choice.
Reggie: …let’s just use this”. So that’s really how it works. Okay. It’s cool.
Thomas: And you really see a game that you like to play and it’s like “wah it needs this certain graphic requirement?”
Reggie: Exactly! Oh my god, damn irritating. Every time there’s this new game that’s coming out and you want to play and it’s like, “ah not good enough, gotta upgrade your hardware.” This is exactly where the upgrade… your hardware comes in.
Thomas: Right. Even your Minecraft, your Fortnite, maybe… okay, I don’t know who plays Cyberpunk now, but maybe those who play Cyberpunk, their graphics and all that are a certain level. Take GTA, for example, the graphics upgrade a lot. You cannot use graphics cards from three years ago. Maybe you’re playing on the lowest setting.
Reggie: Exactly what I was about to say. It’s very sad.
Thomas: No choice!
Reggie: Yes, yes, which is why gamers are so hardcore when it comes to buying the latest hardware product and as gaming grow, as a tailwind, this will potentially grow. We can talk about this later. Okay. Yes, continue.
Thomas: Sure. So these are your gamers. They will be one segment of the customers. Really, these products are just designed for games, but the underlying technology which Nvidia has assembled and researched and developed can also be used for compute workloads.
This will lead to the other segment, which is more for business, say B2B (Business-to-business). For example, AWS (Amazon Web Services), they have a huge data centre. We are talking about the storage part of it, we are talking about the compute part of it.
Reggie: Storage is actually very cheap really today. We can talk about storage another time.
Thomas: So compute workloads, basically your AI, your data science workloads, things that need to be done on the fly, things that need to be computed on the fly, process on the fly, like a lot of workload that needs computation. AI, any kind of workload needs it. Tons and tons of gigabytes. I think this is very low in supply. As much as possible, AWS is just like “whatever you have, give me.” GCP (Google Cloud Platform) also, same thing. Azure, same thing. Nvidia is the number one leading supplier, if not by market share. I think it’s about 90%.
Reggie: I think they almost cordoned the…
Thomas: They are just dominating this space.
Reggie: They are the only real big boy. It’s not like two, three corner fight. They are the big guy and then someone else is trying to come in like that. They’re the real big guy, and definitely we can talk about, storage another time, but I just want to take this time to kinda let people understand a little bit about the kind of computing space that we’re looking at.
Storage is essentially your hard disk, what people call hard disk where you just store your things and that has become a commodity. A lot of people can produce and it’s very cheap today. There’s the intermediary which are like your RAMs (Random Access Memory). Those are like doing shorter term, mid term kind of computing level. They are not crazy as what Nvidia products are in, but they are doing a lot of the middle, which is like your motherboard and your CPU and all those guys are there.
So essentially your Intel, AMD, they’re doing a lot of those things there. Nvidia is right at the very top, which is all your high level computing needs which falls into gaming, into data centre, falls into mining. Is that how it is?
Reggie: Ah not bad, at least I know some things. Very good. Okay. Yes.
Thomas: But next time you want to play a game and you say “oh, you have insufficient RAM requirements.” RAM refers to your GPU. If not enough memory, which is oftentimes in today’s world, it shouldn’t be the case. You will always have enough space. But sometimes for technical folks, they refer to memory as the RAM because it’s actually memory also, random access memory. RAM is…
Reggie: It’s like an intermediary.
Thomas: Correct, correct. That part does mini computation.
Reggie: Yes. Not bad ah, I got work in the tech company before, so I know a little bit, but yes. Cool stuff. So then if Nvidia is so dominant in the space, what are some things that we’ve got to look out for when we are looking at that business to try to understand where are they at? What are the metrics to be in this business where Nvidia is in, which is in GPU?
Thomas: I think you have to look at the different segments which they are trying to actually cater to. Number one is for gaming. It’s not something related exactly to their company. Like Facebook, like active users of the platform and all that. Because Facebook is very much a platform business, right? It’s their own business.
But gaming as a trend… so what are the metrics that matters most are things like number of people on Discord. Discord is where people actually do their livestream of the games.
Thomas: People are actually playing… yeah. The support of these kinds of games and you also know which games are trending the most. Based on that, you actually know the trend of where the games will also hit next. Definitely more higher, power graphics and what kind of games also. Is it your multiplayer game? Is it like MMORPG type of thing, or is it more of like a first person shooter? Or is it those indie games you just jump, jump, hop around? Not so much of the GPU…
Reggie: The only game I watch on Discord…I don’t know if I should say, is Tetris.
Thomas: Whoa, you watch it because of the game?
Reggie: No because they damn cute. It’s like two big, very grown men down there, play Tetris NES. Very old, traditional… so it’s very funny and they hype it. So it’s quite cute, very indie game and I like the community in that way and I can understand Tetris for sure. But I get that most of the guys that are playing the games or streaming on Discord or do the Twitch, they are doing the very high power games, exactly like what you said where Activision is in, EA is in, Epic Games… they’re all in that space. But I never thought that I should look at Discord as a metric to understand the future of where Nvidia will go to. That’s interesting.
Thomas: The type of game matters. We can talk about it a little bit later. Just an example, mobile gaming as compared to PC gaming. This is a big shift in itself already. You won’t need your GPU on your mobile phone but you need the equivalent. So what is Nvidia actually doing about that? We can go into that later.
But just some numbers for perspective, we have 2018 numbers come back to 2020. So 2018 was 26 million active Discord users. In 2020, it’s 140 million. That’s how many people are on Discord just watching livestreams. But it’s also the interest in a certain game. people, maybe they livestream also. I hope they are also playing the game at some point.
It’s a proxy. It’s not a definite number, but we can tell which kind of games people are most interested in. Other things also, YouTube, the number of gaming hours that is watched. It’s 50 million as compared to 100 million in 2020. That’s double.
Thomas: For e-sports, the global audience of e-sports, like you watch your Dota competition, or your League of Legends competition and all that. 2018 is 361 million, 2020 is 436 million. Not as big as a growth, but this is the global audience, usually in a team sport. But yeah, numbers is just increasing by a lot. Steam is this gaming market place that’s in charge of…
Reggie: I use Steam.
Thomas: …The distributing of Dota and all that. Half-life, last time… Counter Strike, all the oldies. Team Fortress, 17 million in 2018, in 2020 is 25 million. So all these definitely… you know lah, gaming is going upwards. But what matters most is that individually for these platforms or for these YouTube videos, for these e-sports, what’s the game that’s gathering the most attention? Which is the ones that are being played the most because ultimately these shapes how Nvidia’s core products are going to be used.
Reggie: But this is only the retail side of the business, which is selling to the end user of the gaming space. What is the percentage of their business going here and what is the percentage of their business going into the business side of data centres and all that?
Thomas: It’s about… 50% is in gaming, but around 50% are also in data centres. The rest is not very significant, but we see a lot more growth coming from data centres actually.
Reggie: Why is that so, actually?
Thomas: The demand for all these workloads by companies. Let’s say the customer of AWS, Azure or GCP, they want to do all their data science workloads on the cloud and this is something that GCP… all these cloud vendors, I’ll just categorise them. They also have trouble building up capacity. All this is just a continuous thing. Any time you have supply of these GPUs, I buy. I don’t care how much, I buy.
In fact, there’s just not enough of these GPU resources throughout the whole world. It’s only maybe given as a free trial in Google for some time and all that. You think like it’s a lot. Actually it’s not, you need a lot more than that. Another thing is that the gaming part of it is also going to the cloud. So instead of needing all these being processed onto your computer, you actually don’t need that high quality of a GPU anymore. You just need a decent display and a fast internet connection, which is also where the rest of these telco industries are going towards. So the shift towards these data centres, it’s a lot more other than just the workloads, but it’s also because of the gaming part of things. There’s a gaming added component to their existing workloads.
Reggie: Nice, which is why whenever people talk about Nvidia, there’s always this discussion about the game future. Is that kind of where it is? Because there’s a very big overlap, even at a secondary level where the data centres are expanding, they’re also trying to serve this high computing need, gaming space and the crypto space. Essentially, that’s where we are seeing, right?
Reggie: Okay. That’s very, very cool and what are their numbers then? How have they grown over these few years? It sounds like they are the only… almost the only guy in this space and it sounds like their products are highly sought after. Like you open shop, no more already. How are they doing from a financial standpoint?
Thomas: It’s pretty insane. Revenue wise, we have numbers from 2021 and then 2020, we’re comparing based off now versus last year. Now, it’s $16.7 billion vs $10.9 billion. It’s a 53% growth.
Reggie: Over one year?
Thomas: Over one year.
Reggie: What the hell… okay, that’s a crazy, crazy growth.
Thomas: The largest contributor to this was a shift towards data centres. It became the single most largest driver of the gross margin improvement. Single most.
Reggie: Nice… Meaning in the past, most of their clients come from gaming and then over this one year, because the data centres were buying so much from them, it made their margins better and they essentially grew up by 50%. That’s kind of what I’m hearing.
Thomas: For [indiscernible] context, the revenue by segment, gaming was $5.5 billion, then it became $7.8 billion. It’s a 41% increase. Data centres, from $3 billion, it grew to $6.7 billion. The growth is 125% and it’s still growing on that kind of annual growth rates. So it’s a bit ridiculous. The rest of the other segments is the other miscellaneous stuff. It’s still contributing some things like pro visualization.
People that still use very advanced 3D, AutoCAD, that kind of stuff. They also need certain ways to compute, and of course automotive. But automotive has actually been declining. Although we think automotive is EV, but when you think of it. Nvidia is the end of the chain of all the semiconductors and they’re only focusing on these processing units. So why do you need one on the auto side of things? Not yet, maybe the AI haven’t come to the auto yet.
Reggie: We are not at that level yet, but in the future, I think with the whole Blackberry QNX and the whole EV growth and multi feature requirement, sync to all these weird things that you want to sync to, sensory between multiple cars and hardware on the road… maybe by then, we’ll need this kind of very high level processing. But essentially, is that a legacy business of Nvidia? Maybe that’s where they started and then now they are in this corner where they are at the GPU level and they have found their footing here. So one of their legacy business was auto and they are cutting it. Is that kind of what it is?
Thomas: Actually not really so much of legacy, more like along the way they did some R&D. They found that there are some applications in auto. Because they have a few, I would say joint ventures or partnerships with a few of the leading auto manufacturers… Mercedes-Benz, Volvo, SAIC. SAIC is a Chinese car manufacturer for all these EVs.
So basically they also have the notion, other than electrifying the vehicles, also having the AI component. But it was a little bit hyped up when they initially went on this JV (Joint Venture) thing. It was like “oh EV will have [indiscernible] from the start. But it’s only actually Tesla has a really strong and robust AI that can handle automated driving. The rest is still pretty semi-auto and all that. I would say there’s a lot more that can be hitting Tesla’s level today. But back then when they started it, they were thinking like “oh, AI is going to be at the front of the car. It’s got to be there, let’s put our GPUs there.”
So now we see this decrease because that’s not really happening. The electrification is still ongoing. You still have to make the existing fleet lighter. You still have to make the existing fleet running on more efficient batteries first before you even start to put a GPU which guzzles so much energy in the first place.
Reggie: Yeah, for sure. With that as the premise, we are not discounting the future growth of this segment. It’s just that at this moment of time, Nvidia is not focusing on it also because the market may not be ready. Is that what I’m hearing?
Thomas: Yeah, totally.
Reggie: Okay, that’s cool.
Thomas: It’s like on the way there… “let’s park it there. When the time comes, maybe the technology changes or anything, you can either adapt to it or we can acquire someone who can.”
Reggie: Okay, that’s cool. Nice. So what about their cost margins?
Thomas: Right. So cost has also been increasing, but not as fast. It’s $6.2 billion compared to $4.1 billion. So that’s a 50% increase.
Reggie: Wow, okay.
Thomas: This also needs to have a lot of shift towards the data centres type of workloads. A lot of R&D also are going into say, these kind of Bitcoin mining use cases as well as really just optimizing for their future needs.
Reggie: When you talk about costs going up, is it because the material costs are going up? What you said like… because there’s an increased need globally across all the different sectors so the suppliers increased their price, that’s why it’s coming up on their end? Or are they investing their money into optimizing supply chain, creating all these other things and all those things then factor into their cost? Do you have an idea of what’s happening here?
Thomas: Actually, I didn’t do an analysis into the actual costs for Nvidia, but from what I read inside their filing is that actually a bulk of this was R&D related.
Reggie: Okay. R&D is always a big cost, and it’s always the kind of cost where if you’re not in the space, it will be a very big question mark for you. You don’t understand what’s going on, but as long as you’re in tech, I think you can easily factor 5-10% into R&D so that you can continue to grow and grow and grow. Okay. That’s cool.
Thomas: Costing wise for end of the value chain kind of producer, you’re also using a lot of people’s parts and sometimes when you use these parts, you need to pay royalties. So the more that you actually sell, the more also you have to fork out in order to pay these people, the royalties. For example, QCOM. Qualcomm…
Reggie: Qualcomm, Broadcom, all the rubbish. They have a lot of license everywhere… licensing, patent issues that are coming along with, so yeah. You have to pay them essentially, because like what you said, they are end of the chain. Okay. That’s good to know.
But then does that mean that Nvidia itself, do they actually have all these technological advantages then? Because if they have all these people midstream that is… have all these patents and they’re blocking. Essentially, they own all these IP. What about Nvidia? Do they own a lot of IP also and how are they protecting themselves in this technology space?
Thomas: Sure, they have tons of patents. It doesn’t mean that they use someone else’s patents inside their own patents. It means that anyone else can also produce because there is just one way to go and configure their own product. But they’ve done it so well. They’ve invested so much into R&D from last time until now. That is their moat actually rather than something that works against them.
So it’s super, super complex. Everyone just thinks “oh, just chip only. Just slot in.” There’s a lot more work to it. There’s a lot more things that you need to optimize and how you even… say for the first time, they go to TSMC and say, “hey, I need you to go and design this kind of chip right in this way, this way, this way… no other way. I only want this way.” TSMC is like “what the heck? Why?” “No, I know. Don’t ask question. I just want to build it this way.”
Imagine that was the first conversation that Nvidia had to TSMC and TSMC is like “okay, I build lah. You pay me this much. Okay done.” And then based off that, they were the first ones to go into it this way.
AMD, which is another competitor of Nvidia, doesn’t come close in terms of their compute, doesn’t come close in terms of their efficiency also. This is… way, way high advantage which they have amongst competitors in terms of product. So that is a moat. Once they say “okay, we already ordered TSMC. Now we have an ongoing relationship. Help me to configure it in such a way that we can also produce a slightly more efficient power usage.” That’s all.
But this becomes part of their design inside Nvidia, it becomes part of their patents also. For TSMC, they can hold the IP of how to make it. But at the end of the day, the end product and all that is going to be held by Nvidia. TSMC just produce. “I know how to produce the chip this way. I have no idea how to use it and it’s not my business also. So I rather focus on the process, the manufacturing, their own portion of the assembly line and all that. Nvidia, you settle your own.”
Reggie: Nice. That’s good to know. And by the way, for everybody that don’t know, TSMC is 台积电. It’s a Chinese Taiwanese company. We can talk about that another time. Are there any other things that we should look out for in their financials?
Thomas: I think debt, for those that just don’t like to look at owing money and all the kind of stuff. It has been growing, so something to be a little bit alarmed. It’s $7 billion compared to $2 billion last time. So definitely, as their revenues have increased, they have also increased their debt load. But not to worry, because they can pay this off at any time. They have tons of cash also. But it has increased more than their free cash flow for the current period… something you might want to be alarmed about, but with the current growth rates, may or may not be an issue, depending on your preference.
Reggie: That’s good to know. That’s it for financials. We get a broad idea of what do they do and what are our financials looking like. What about the team behind it? It’s always about the team, especially when we’re in this fast growth space.
Thomas: The CEO, Jensen Huang, he is also the initial founder, started Nvidia in 1993 until now and he’s been helming it ever since. So this guy, he’s just been there since the start. Engineering background, but also knows how to manage people and knows how to reward them, incentivize. Also knows how to run a whole engineering work and just really getting the right people to build Nvidia’s products, as well as lead the R&D teams.
Reggie: Do you realize all the big CEOs, major… most of them in the semicon space are Chinese people? Nvidia, AMD, TSMC… I think they’re all Chinese people if I don’t remember wrongly. You see where we’re heading towards. A lot of the manufacturing lines are in Asia. So I think that’s probably something there, and a lot of engineers are Asian. I don’t know. Just random thought, random recognition.
Thomas: Ahem, budding engineers.
Reggie: Yeah, continue to strive up. You can be the next CEO. Yes.
Thomas: But Jensen Huang is Taiwanese.
Reggie: Yeah, exactly. Why TSMC? -wink wink- Great trust of their own people. It’s very normal. Japanese companies, they go abroad, they hire Japanese for top roles. It’s very, very normal. There’s a lot of heritage and nationalism in all these kind of hiring process.
Okay, so founder-led company. That’s important. In my view, I like founder-led companies because founder have a bigger drive beyond money, beyond compensation. They are the ones that shaped this company from scratch. That’s kind of my personal bias for founder-led companies.
Thomas: 28 years in the company, can you imagine that?
Reggie: There must be a big reason why he continues to do that. That’s good to know.
Thomas: A Nvidia fellow… basically he is also one of the co-founders, but taking charge of the very heavy engineering and the tech side of things in Nvidia. The background of this person, his name is Chris Malachowsky.
Reggie: Sounds like a Russian… I don’t know. Sound like Western European or Eastern European. Definitely a cool guy.
Thomas: He has a really very strong background. He had engineering and tech leadership positions at HP (Hewlett-Packard) back then when it was cool.
Reggie: Back then when it was cool!
Thomas: There was a period, HP was selling like hotcakes.
Reggie: Exactly, and that’s the reality of this business in the microchip… or even just in the hardware computing space. It’s an arms race. Everybody is competing and for a period of time, this guy will be doing better and then someone, somewhere will hire some crazy geek and then they would be able to develop something even better and then they’ll outpace this other person. This is quite a cycle in this space. So I totally get the whole idea. HP used to be a thing, and now it’s like “HP? 什么来的? (What is that?)”
Thomas: Sometimes tech is not aligned with the management side of things. The manager may just decide “oh, I wanted to go in a different direction” and that’s it. You may have a fantastic product but say, the marketing was not good, the execution was not good and all that. So really, definitely the management matters.
Okay, this guy, he worked at Sun Microsystems. Sun Microsystems was also where Bill Gates worked last time. There’s a lot of other tech founders with some background at Sun Microsystems. So I’m not sure what they learnt there or what they built together there, but it seems to be like that place where all the zai kia (smart kids) all come out.
Reggie: That was the place. Now it’s Google and Alibaba. When you graduate from Google and Alibaba and then your investors will be like “okay can. I invest in you.” Back then, during the Bill Gates era, that will be Sun Microsystems. That’s cool.
Thomas: 40 patents to this guy’s name.
Reggie: Wow. Individual patent. That’s very, very cool. Really a tech geek, and it’s extremely important in this business. It’s like you run a retail business, you need someone in merchandising that very… understands this space. When you’re in this hardware business, you need someone that’s really an engineer and fully understands the space. If not, it’s very hard to have an edge. So that’s very good to know that we have someone like that in the team.
Thomas: Then we have Colette Kress. She’s the EVP (Executive Vice President) and Chief Financial Officer of Nvidia. She joined since 2013, and her background was spanning a few major tech companies. 25 years in total. Some background: previously before Nvidia, she was the senior vice-president and CFO at Cisco under the business technology and operations finance organization. For Cisco, it’s the Cisco Systems, not your Cisco police. It’s Cisco Systems.
Reggie: The very big tele communication producer.
Thomas: Yes. Before Zoom was cool, there was Cisco.
Reggie: Not your Cisco police, not your Cisco security.
Thomas: You know why? The Cisco police, they were forced to change the name to Certis Cisco because of this.
Reggie: Oh really?
Thomas: Yes. There was a mix up with the Cisco technologies and Cisco police, so they changed to Certis Cisco after that.
Reggie: Okay. They’re not as cool anyway.
Thomas: She also had background in Microsoft, so 13 years there and then also 4 years as a CFO for the server and tools division inside Microsoft and also doing a bunch of other stuff. Way before that, she was at Texas Instruments. TI was also another place where a lot of tech leaders also came up from. So really good management and really good practices.
Reggie: Yeah. That’s very good. Almost all of them have very deep knowledge in this space and they all come from previous companies where they are all… actually, a lot of these companies, whether it is Texas instruments, Sun Microsystems, Cisco, they are still very big players in their game. They’re maybe not as big as back then, because now there are like more interesting and bigger ones coming up and/or are in the space already, but they’re still very dominant in the space of semicons and all these kinds of stuff. So it’s good to know that the team is filled with people with such experience. Is there any other notable figures in the space in their management?
Thomas: Yeah. There’s Jay Puri. He’s another EVP but in charge of worldwide field operations. Worldwide field operations is basically your global sales and regional marketing.
Reggie: Important. Make money.
Thomas: Other than Nvidia tech company, you have someone that’s really marketing all these product. He joined in 2005. That’s some time ago. It’s not someone that just stay here “oh, I market already, then I zao (go).” He’s been there since… for quite some time already. 16 years prior, he was 22 years at Sun Microsystems. 22 years! It’s maybe some of the viewer’s lifetime already. That’s his career at Sun Microsystems before he joined Nvidia.
Reggie: Crazy, crazy.
Thomas: He did sales marketing there, general management also. Then he held several roles before that at marketing, management consulting, product development, positions at HP, Booz Allen and also Textile Instruments.
Reggie: Nice, I can see where they’re hiring.
Reggie: All the alumni, but it’s important. I think a lot of people that are not in this space, they don’t understand the importance of… even for a sales person to be in this ecosystem. Because in this ecosystem where a lot of your sales processes are extremely centralized, they are very big distributor chains in this space. It’s not a direct to customer, one to one sales kind of thing. There are a lot of distributors in this ecosystem where they sell to their local merchandisers, local distributors, or there’s a lot of big buyers in this space. So if you are not in this space, you don’t have the relationship, you have not cultivated this whole process. It’s something to look out for.
I think it’s great that they have someone that has done all these years of sales in the space of semicons. Of course, it’s not always the same, like from memory to processing to GPU to all these different things that they sell, but as long as you’re in the space, your connections can be pivoted along. So I think that’s extremely important. That’s good to know.
Thomas: Right. I mean, the list goes on, man. This is just a running list of world-class talent.
Reggie: Cool, cool. Interesting. Good to know that a company is hiring from all the alumni. It’s important. Given such a team and where they are at, how do you see the business going forward? Do they have very strong moats? Because they are dominant in this space, but can they continue to be dominant? I think that’s also a very important question.
Thomas: I think they’re quite secure until the mid term. Long-term, I’m still a little bit unsure. Here are some reasons why. For data centre, it’s quite secure. It’s 97% market share of the data centre GPU computing and all these resources is being used by the Fortune 100 companies. Fortune 100 companies are not going to go away anytime soon and 97% market share is also very high for data centres. They’re also not going to phase out so much of their computing source so quickly. As much as possible, if it’s reliable, it stays there, and then they will only upgrade it when the time calls for it.
But there are cheaper ways to actually expand their scale rather than just buying more LAN, and all that kind of stuff. It’s by actually working with the same supplier and making sure that the R&D… to optimize it, then just swap out. It’s the point where the market follows what Nvidia proposes right now. So they do have that kind of influence and because the products are so very high quality.
You don’t see AMD being heard much, being used in data centres. Unless it’s really some chapalang (random) but you wouldn’t want to use it. Only the three big names are in the cloud today. It’s basically Azure, AWS, GCP. Anything other than that, you think already… okay lah, second tier. Cheaper, but I’m not so sure about the quality or consistency or data security, all that kind of stuff .
Reggie: Okay, that’s interesting. That’s good to know. How are they planning to scale their business then? Is it just about… because what I’m hearing here is they are already capping their supply. That means they’re maxing out their supply already. So are they going to be investing more in supply chain to be able to produce more? Or are they going to be raising their prices or is there some light as to how are they going to continue to scale their business?
Thomas: Right. I think it’s to do with more of how they are addressing the shifts in the trends. Earlier on, we talked about gaming moving to the cloud. That’s one, and another is that the AI is being needed to be run on the device itself rather than in the cloud and this one is a little bit inverse. Gaming is going to the cloud, AI in some use cases is going to the device level itself.
The term is edge computing, at the edge. So you compute at the edge, that’s basically the hyped up name that Gartner uses. With this edge computing and gaming going to the cloud, that’s like the inverse of what their current GPUs for gaming or for the automotive side is actually doing. But for now, we still see that gaming is being in high demand. People are still playing PC games and all that, but we don’t know for how long.
We already know for mobile games, that’s becoming more and more of a trend already. It’s becoming a common thing. People on the way back from home, even people at home and all, they don’t switch on their PC to play games. They play mobile games instead. These mobile games also either need the AI on the device itself or the graphics and the AI can be on the cloud itself.
In either case, Nvidia doesn’t address that well yet. If it’s in the cloud, probably, but it’s quite indirectly because it goes through your cloud vendors, your cloud providers. It’s not Nvidia themselves. So it’s not something that they can exactly bank on, butthey know that they are the key supplier for these data centres. Fine. But what if? .
Reggie: Because the big three game developers, EA, Activision and Take-Two… other than the Chinese market, these are the big three game developers. They are all going mobile. They all have some sort of division that is focusing on mobile and their mobile game division putting more capital in, developing more franchises in the mobile game space. So if you want to talk about information, they probably got the most information out there in terms of where is consumption coming from, what is the future of gaming. Their activity will also shape the game space.
Exactly like what you say, maybe Nvidia is doing very well now because there’s still a lot of high computing needs in the PC game space or the console game space, which requires a lot of these kinds of high computing requirements.
But if let’s say gaming, because it’s such a big business in Nvidia, if it shifts more and more over time towards mobile, lesser and lesser people using big consoles then they definitely have to keep up with this and it will affect their core business.
Thomas: Right, and you need a GPU of a certain size also because where it’s supposed to fit in these each computing cases or even on your mobile phone? Can you imagine on your phone and then you just slot a huge… no way, that’s not going to work. It’s got to be small. It’s got to be really, really small and something that is efficient also so it doesn’t heat up so fast and your hand doesn’t get burnt and your battery doesn’t get used up so quickly.
So all this is the same kind of needs for auto for your edge devices or even for IOT. IOT is internet of things, your sensors and all that. Your AI is going to go onto the sensor itself. How does Nvidia actually address these then? Biggest way… or they can address it in one shot is acquisition. You might have heard of Nvidia trying to acquire ARM. So what…
Reggie: ARM is big.
Thomas: ARM is basically like a processor. If you compare that to Intel, you always see Intel Inside because that’s your processor on your motherboard, your PC powering everything in some sense. It’s just your brain. But it does more than that. It does a lot of the computations and for ARM specifically, its their architecture. They did it so well that even Apple, the new version of the Macs that comes out… yes, the M1 chips is based on that architecture and they’re just moving off Intel altogether.
Apple’s not going to pay royalties to Intel. Apple’s just going to say “we’re just gonna use the ARM architecture. We’re going to build it ourselves for less costs.” Now, if Nvidia really gets the acquisition without all these antitrust lawsuits going on, even if they managed to pass it, this is a very big boost for them. Because one is that free business. Apple, whatever device you sell, I’m also going to get a cut of it.
The second part of it is that, for ARM’s architecture, if you think about it, when it’s being used inside the M1 chip, it’s being used in very small devices. It’s not just in your Mac or your MacBooks, it’s also in your iPhones and in that factor already, you can actually power games very well with such a good efficiency to energy ratio and the heat dispersement is just phenomenal.
There’s a lot of YouTube videos on it, you can go watch. But everyone is shocked, like what the heck? This thing doesn’t even have a fan and it can power my game. I am so shocked.
Reggie: ARM is amazing. Yeah.
Thomas: It’s crazy stuff. It’s really a big technological breakthrough. So if Nvidia is able to get their computing portion over into ARM’s architecture, that is phenomenal. Then it solves their problems with just one shot already. But, of course this takes R&D time. This also means that the acquisition needs to come through first.
Reggie: For sure. I think a lot of people don’t understand, may not be able to understand ARM right from the get go, but we’ve talked about it throughout this whole discussion about licensing, IP, licensing, IP. You have this structure that you’ve created and then you license it and then other people, if they want to use it, they have to pay you royalties.
ARM just does that. Every day, they are thinking about how to make it better and then they will license this whole process, how to get things done, where to put what? And they just do that and they are actually at the very top of this game of ideating and creating the best computer chips out there.
They don’t own the manufacturing line. They don’t produce. They just think of how to do it and then they license to whoever that want to produce, and almost everybody in this space use some sort of ARM license, definitely one. So for Nvidia, if they can actually own ARM, which is currently owned by SoftBank, or is it Alibaba? Currently owned by Soft Bank.
It’s actually a UK company by the way, owned by SoftBank now. If they can buy, Nvidia can own this, then this can potentially become a thing but exactly what Thomas said, it is not a definite kind of thing. They still have to develop and they have to come together to see market fit into the future. Also, because now there’s a lot of antitrust problems. Nvidia is a very big player and if they were to own ARM, then they essentially dominate a lot of the ideation and creative market in this semicon space. So I think there’s a lot of discussion there. Whether can go through is another thing.
Thomas: The antitrust one is actually being fought by the industry players. It’s not so much about coming from government. So we also see the validity of these claims. A big one will be Qualcomm. Because once this accquisition goes through, the R&D part is already settled, and then they managed to merge these ARM architecture as well as the GPU together. There’s no need for Qualcomm. That’s just eating into the business, big time.
Also, the M1 chip is also a very big boost to data centres. So you imagine that, now the data centres just guzzle tons and tons of energy because of the cooling requirements and also because of the electricity just to run the thing. With M1, this cost is just reduced dramatically. Minimum at least 20%.
Reggie: That’s the power and I love how we are talking about Nvidia, we talk about all these other guys also. Because you cannot look at semicon companies, or you cannot look at Nvidia in isolation. You must really see all these other things and I’m sure over time we can keep exploring all the companies that we’re talking about and then give people an even more solid idea about what is going on in this space.
But at the base itself, Nvidia is definitely leading where they are with GPU, which is something that is extremely used in modern day computing, whether it’s from gaming or whether is it in the data centre process, very high level stuff. They are also innovating on other things and potentially trying to acquire so that they can continue to extend their moat. I think that’s the basic idea where we are at.
So to sum up, what are some of the headwinds? What are the tailwinds in this business? What is the opportunity for Nvidia?
Thomas: I think a large portion is coming from the cloud. That is going to stay. The tech trend for companies and all the workloads are going to be processed in the cloud. Even storage of data, some sensitive data, it’s going to be stored in the cloud. The cloud vendors have certain SLAs (Service-Level Agreements) which is ready, and it’s quite secured. It’s just whether your manager or whoever is approving, understands these things.
If they don’t understand it, they will say no. Everything is going to sit on premise. It’s going to sit on your hard drive or something, then too bad. But that is going to be the minority of cases already going forward, at least in the next five years.
Reggie: Sounds like the military here. Intranet, 什么年代了 (what era already)?
Thomas: Private companies…
Reggie: I get it. Just disturbing. This thing cannot circulate outside.
Thomas: Other tailwind would be more auto, more edge computing use cases. This one is really towards the tail end, but it’s definitely something that is already coming. You see the wave coming. Now, it’s just can we match up the technological progress to reach there faster? It’s not so much of a matter of if, it’s when already.
Of course, this has to do with Covid also. It’s affecting some R&D efforts on both Nvidia’s side as well as the compute sitting on these edge devices. But definitely something in the long-term that we can look forward to.
Headwinds… some of the possible barriers or things that may slow down Nvidia. Nothing is stopping Apple from producing their own chips as well.
Reggie: They already do M1.
Thomas: So what’s stopping Apple from also supplying these data centres themself or building their own data centre? Definitely, Apple has their own data centre. It’s just that they don’t open it up to the public. But hey, maybe one day, like your Apple Cloud becomes literally Apple Cloud versus all your other cloud vendors already, then you have a fourth player.
It’s very possible because now GCP, Azure and AWS, they are the incumbents. They are the building things off old technologies. You can even see Nvidia, it’s very high tech, very computing. It’s going to considered old technology soon, once these things actually merges. Very fast. The technological piece is just phenomenal. Apple can just start out from fresh. They just start with a totally new design for data centre, totally new design of how to erect them, how to do the efficiency, storage, energy and all that kind of stuff… floating data centre. No problem.
Reggie: Don’t forget that Apple has a lot, a lot, a lot of cash. For a very long time, for years, investors have been asking Apple “why you’re not spending?” They have so much money sitting around, they are not giving back to their investors, but they’re also not doing mass acquisition. They are not really growing in any places, double down like the other FAANG (Facebook, Apple, Amazon, Netflix, Google) guys, but now you see a potential and they are playing around with some of these things. So that’s a very interesting headwind that potentially can happen. It’s good to know.
Nothing’s also stopping tech driven auto company like Tesla. Nothing’s stopping them also from producing their own AI’s compute ability on their cars. So there’s actually this roster which Nvidia works with in terms of the automobile manufacturing. Tesla’s not inside because Tesla’s doing their own thing. Mercedes-Benz, Volvo, SAIC, they don’t have this capability. They don’t know how to use semicon, so of course they’re going to be tapping on Nvidia. This is a boon for them.
But if you see Tesla as the leader, they have already solved automated driving, the first one to do it. They are the first ones to actually configure the energy to efficiency ratio very well for cars, the first. They might be the first or so for the automated driving. Like edge… if you consider that as a first edge computing use case, and they can probably even streamline down to the other kinds of use cases. Tesla might not be just a car company anymore. It might be something else. Yeah. We can talk about this another time.
Yeah, yeah. Sure a lot of people want to hear your thoughts about Tesla.
Thomas: Yes. It’s not a crazy thought, because last time SpaceX, people thought “eh can launch and then land back by itself? Siao ah (Crazy or what)?” Now it’s happening on a monthly, bi-monthly basis? This becomes the norm already.
Reggie: I can bring a good friend that is in the rocket space to talk about SpaceX another time. Because he said that the biggest problem in SpaceX is not the rocket, but the fuel. The fuel… Is like airplanes, right? Airplanes, no matter how efficient you build an airplane, the varying cost is fuel, not the planes anymore because it’s already very efficiently built. That is a very interesting angle from the people in the space. We can talk about people in the space. All the space people always got this pun. We can talk about that another time. Yes. Good stuff. yeah.
Thomas: So Nvidia’s headwinds are as such.
Reggie: Are we concerned about their competitors? Are we concerned about their competitors? Because they are not living on an island.
Thomas: Not at all. Very, very far. First one, you do a review between AMD and Nvidia, let’s say for gaming, or Nvidia versus AMD in cloud computing. The reviews are bad for AMD. It doesn’t do deep learning very well. Deep learning is a type of artificial intelligence workload. It doesn’t do it very well. It’s either slow or it’s just not efficient. It’s not maximizing its true resource potential, so engineers are looking into it, like “what the heck is AMD doing? I don’t want to look at this anymore. Just use Nvidia.” It’s like that, the switch is so fast.
It also doesn’t do 3D rendering well. For your games, for your visualization workloads and all that, it doesn’t do it well already. This is the base thing that you need to get right. I’m not sure why, but something is happening inside AMD, that is… it’s not seen in the Nvidia. That’s why Nvidia is succeeding. Yeah.
Reggie: What about Intel?
Thomas: Intel is more of a processor thing. Intel’s direct competitor would be ARM. They deal more of the computation, the CPU, not the GPU. They are sort of an indirect competitor now because… let’s say we are extrapolating, Nvidia does acquire Intel, but Intel is going to be blown out of the water. It’s just a matter of time. It’s not… again, it’s not an if.
Why is… because again, old architecture. They have very strong hold over PC. Sure, there’s going to be people that still use PC and that kind of stuff. But where is the trend of the entire technology space going?
It’s not just PCs anymore. We’re going more than that. So for them, it’s a totally different market share. It’s a totally different product also, but we see their core being threatened by ARM itself whether or not Nvidia’s acquisition goes through. Or… Apple progresses the R&D of ARM architecture faster, it’s also going to invalidate Intel. So as an Intel investor, I will be considering quite strong strongly, the core product and where Intel is heading.
Reggie: Cool. That’s good to know. Anyone else?
Thomas: Yes. Qualcomm. QCOM, it’s just a royalty business. Just patent, patent, patent and you pay me license fee, thanks.
Reggie: Yeah. They go into a lot of trust issues recently.
Thomas: Most of these are actually from the 3G and 4G patents, but 5G is coming. They are still on 3G and 4G. The 5G portion is going to come from other vendors and Apple knows it, Samsung knows it. They say “okay, I can continue to pay royalties, but I’m not going to keep negotiating with you when you play hardball with me. I’m just going to let you phase out. What are you gonna do? You don’t advance any of my technologies. You are not even giving me what I want. Ok lor, you charge me whatever then I treat you like that…” Again, this is already a relationship thing, right?
Thomas: You cannot just sit there and do nothing and then just enjoy all the profits all the time and then you really have no contribution value to the entire industry. Apple, Samsung, Xiaomi, they contribute at least three quarters of the revenue to Qualcomm and it’s just from the 3G and 4G patents.
Once 5G kicks in, and this is again nothing to do with Nvidia, but just for this company, that one quarter that’s left is going to be part of their compute. It’s going to be part of that efficiency architecture of what ARM can actually provide to Nvidia, but whether or not Nvidia’s acquisition goes through, ARM’s architecture, will still progress because of Apple. So Qualcomm is being threatened.
Reggie: Yeah. Nice. That’s good to know. In this space, okay… I think you’ve shared a lot. I want to sum up… in this space of high level computing devices, every single CEO will come out and tell you they want to do the next big thing. Every single CEO, every investor day, they will come out and tell you there’s this new big thing that they want to do, this new big thing they want to do, but are they really doing it? That is the big question, right?
They can promise you the world, but are they really making it happen? Because by now, after this one discussion only, you should be able to see that it is not a simple thing to just start another thing. Not so easy to just open other product, not so easy to just open a supply chain, not so easy to just stay with the market or catch up with the market.
It’s a very complicated and intricate space and we can over time talk about all these other things, but don’t be fooled by all these CEOs that will come out and tell you “oh, I got a plan.” Everybody tell you they got a plan. So look at whether their products are working, look at whether there’s market adoption and then from there you can filter and decide which company actually fits your investment palate and which management actually works for you. Cool. Any other things you want to add?
Thomas: No, man. Well said.
Reggie: Okay, great. Thank you. That’s all for today. I hope you guys learnt some good stuff and let’s keep chugging along. Take care guys. Bye.
Thomas: Awesome stuff, Reggie. Bye bye.
Reggie: I hope you learnt something useful today. Definitely recognize that investing is a personal decision. We’re not giving you recommendations here, but we are always happy to geek out with you about different, interesting companies and trends for the future. This series definitely has a lot more depth and terms, so please head over to TFC Stock Geekout topics tab on your members backend to have a great discussion about this company. Share your opinions, disagree, debate and ideate to strengthen your investment thesis and become a better investor on this journey with us. If you have any other interesting companies you would love us to explore, drop ideas under feedback and suggestions or through our socials.
Meanwhile, remember to stay invested.
Sign up for our newsletter to get the latest updates. You may even find limited opportunities not shared anywhere else.
Find out what makes Google one of the most valuable companies in the world!
Companies with high growth and healthy profits are uncommon but TransferWise has managed to find the sweet spot in the fintech industry, beating out competitors like PayPal and even traditional banks.
Despite being a Singaporean brand, this company is world-renowned and many regard it as a great way to fly.
This Company Has Outperformed The S&P 500 For 13 Years & Running $BIP [SGO 42 with Timothy Phillips]
Fancy investing in a company that has beaten the S&P 500 ever since it went public 13 years ago?
A Macro Outlook On What’s To Come… And One Investment Idea [SGO 41 Sponsored By ProsperUs] With Say Boon
Will there be another Asian financial crisis if the Japanese Yen passes a certain threshold? Where are the opportunities? Will the recession hit this year or 2023? What’s the best investment advice?