Nvidia (NVDA) to the moon! | Pixels & Profits #9

hello everyone and welcome back to pixels and profits it's been a minute since we recorded our last episode but Matthew and Brian you guys got some fans they're demanding more pixels and profits so here we are and today we are going to be talking about Nvidia yes Nvidia the AI GPU we'll go into more about the business company they do have a bit of a gaming business which we'll talk about it's kind of like a rounding eror compared to the data center stuff but certainly something that we will go into in a little bit more depth since we are pixels on profits but before we dive in guys how you guys doing and what what is it gonna take for us to do more pixels and profits to to satisfy your audience I think I think just yeah I think getting more suggestions on topics would be great I think that's probably the hardest part right I mean the preparation is not easy but also just hearing what people want it talk about absolutely if you guys

want to tell us what you want us to break down and dive into that can definitely help a lot because we always want to cover interesting topics but they need to be interesting to people that are paying attention so yeah let us know in the comments right i' also like to do more video game content so I can research by playing video games instead of going through financial statements would be sick yeah yeah let's definitely do that I've got some ideas as well but and maybe even like some private companies to some degree like epic supercell companies like that but just for our audience to kind of set the stage on what we're going to be talking about so we're going to start with an overview of nvidia's business we'll talk about future drivers of growth valuation discussion and basically the future outlook for the company NVIDIA is a real I mean it is one of the most important companies in the world right now so clearly a good company to cover and I think the gaming aspects of it are also interesting so but let's go ahead and Dive Right In I'm going to kick it over to you Matthew to start us on the overview of nvidia's business

absolutely so it's kind of important to think about Nvidia as it was and then how it is kind of has been over the last few years with this AI boom because you know to your point we all know Nvidia for gaming gpus prior to let's call it 2020 you know if we look at the nvidia's got one of those weird fiscal years their year ends in January 2020 so if we look at the three months ending January 2020 that's before covid so before we had the big boom in gaming and covid it's before the crypto bubble in 2021 and 22 that affected gpus the company NVIDIA generated about $ 1.5 billion dollar of total revenue from gaming which is about half their total revenue was a whole company they generated about a billion from data center sales they were still at that time selling you know gpus for AI workloads then but it was by no means as big as it is today and then they have these other segments they have professional visualization they have Automotive they were doing stuff with Tesla one point right before Tesla

started to do their own ships another and other automakers right so you know they have all these other segments that made up the rest of the pie if we fast forward to the most recent quarter which is the quarter ending January of this year Nvidia generated $18 billion of revenue from data center which is more than 80% of their total revenue with gaming at 2.9 billion so gaming has doubled in that period over the four years but data center has grown just absolutely exponentially in that period and completely overtaken everything else and like as as important as these businesses might be to relevant people in the industries professional visualization Automotive just not relevant anymore not relevant to the investment case we don't even need to spend time on it it's really just it's data center it's AI maybe gaming at some point again becomes important and then everything else is just kind of irrelevant and so what's going on and what's caused this massive Title Wave shift if you've been living under a rock right it's generative Ai and how that's taking off but I think it's also

just how intuitive we talked about this a while back right how the chat GPT moment was really more of a wakeup call because it was it was such an intuitive user interface for the application of AI that it got so many Executives thinking on how they can actually leverage this and that's I think what really has been different in this AI cycle is we have a use case we have a user interface that's easy to interact with and it's caused a huge demand Boom for AI if you kind of widen in the bubble and call it accelerated compute which encompasses all sorts of data center servers that don't just run on traditional CPUs but they run on other types of chips to help them do fancy workloads accelerated compute is better suited for specific workloads like AI than traditional CPUs because they're just more specific more defined for these types of applications historically roughly 10% of Data Center capacity was accelerated this is rapidly getting closer to 20% it might even be past that at this point point and the

way that things are trending we're pushing towards a world where 50% 75% of total data center infrastructure at some point will be accelerated so that's that's kind of where this ground swell of demand has come from from a fundamental level I think I'll pause there and and JK let you jump in yeah it's kind of funny that the way that Nvidia has described their business originally is that they make gpus to power the human imagination but now they're making gpus to power simulating human intelligence and clearly the human intelligence aspect is a much bigger Market than human imagination but I I think the the short of what we're talking about and what what you mentioned Matthew is that to some degree Nvidia kind of won the lottery right gpus are kind of perfect or very well matched for enabling llm capabilities for our training and inference tasks for AI and I would say like since for me it was like since the start of Tesla investor day but then where every everybody got excited as you mentioned

chat GPT there's just been like this explosion of required infrastructure investment into buying their A1 Nvidia a100 h100s so that you can develop AI capabilities so it's kind of like they struck lightning right time right place they had the right thing there weren't they spent years developing this technology there weren't many other people who could uh develop that technology and so they just they absolutely struck lightning and in terms of the customers they've got a lot of huge big Tech firms that are really investing heavily into this space so companies like Tesla for vision for their full self driving things like that open AI of course with for chat GPT meta and Microsoft is other big examples of the kind of major customers that they have and many of these companies are basically building whether it's open AI like conversational text based or dialogue based research capabilities generative AI and vision as some of the

primary applications that these companies are working on but you also have you know hundreds maybe thousands of AI startups working on various capabilities leveraging Ai and so clearly we're kind of maybe early may maybe even mid I I don't know stages of like this massive infrastructure deployment phase and on top of that I would say Nvidia benefits from a competitive standpoint because they're more Hardware focused and if this where software company you could kind of just hire some programmers buy some laptops or PCS do some code throw some stuff up on AWS and you're ready to go compete against another software company but I'm not a hardware expert here but building a Foundry or even if you partner with like a tsmc it's just much bigger competitive moat so that kind of takes us to where we are today in terms of like nvidia's business and how they're actually pretty well positioned

yeah I would just add and then Brian love for you to jump in also but like Nvidia is an overnight success made over 30 years right they've been investing in GPU technology for the 30 plus years that they've been a company I'm excited to actually read take Kim's new B book coming out later this year which is all about the history of Nvidia the history of Jensen Hong so very excited to read that personally but I think if I think the company almost went bankrupt twice building gpus because like nobody wanted them and it was a useless you know task for a while but it just shows that there's a lot of like entrepreneurial spirit and a lot of real drive that took this company from a vision to where they are today is by no means was an accident this was very much the plan and very much okay maybe they didn't predict when chat GPT was going to blow up and when Enterprise demand for AI was going to happen but they were prepared for it and I think that's you know there's a going to be a lot of long-term learnings and that's not the purpose of this podcast but important to kind of when you think about it it's not like it's just h and then they're at front and center they've

been here for 30 years it's also kind of cool to think about just the birth of the semiconductor industry a lot of that in the valley here was really brought out from just kind of like nerdy Game Dev companies right like AMD Intel all of these companies at the ear onet were developing just basic chips to play video games and then it exploded and for like a 30-year Trend if you want to back track a little bit if you take a Long View on proliferation of data and the understanding that there's there's going to be more data to store to aggregate to analyze the only thing that could really do that was a you know a chip that was made to play a video game because of all the inputs that's just that's just data right it's just zeros and ones being calculated to to display pictures and interconnect with a gameplay back but it all kind of started From yeah just a bunch of nerds making video games which is kind of cool yeah and maybe to kind of dive in a little bit more deeply on the game or

more game development parts of nvidia's offerings and I know this is again a rounding era compared to the data center business but probably something for our audience in terms of the game development game if you're a game developer that you should be aware of and I I won't talk about the consumer side of the business like the graphics cards and GeForce now cloud gaming but in terms of the more game developer focused offerings probably two things to be aware of is NVIDIA Ace and Omniverse and so a stands for Avatar Cloud engine and this basically a set of tools to create and operate digital avatars with generative Ai and these AI models can be run both locally on a PC or in the cloud and the focus for Ace is to help developers customize and tune AI models to make it as easy as possible to create digital avatars and it basically Ace is kind of like a catch off or a suite of tools like audio to face which basically

allows you to create facial expressions that match audio live portrait is animation of 2D portraits or stylize human faces using a live video feed Maxine which helps enable augmented reality effects using AI Reva is a speech and translation AI SDK Nemo is a cloud framework that helps developers build and deploy their own generative AI models and finally I think game developers should also check out the chairo demo from Nvidia where they have partnered with convey to enable conversational AI capabilities for in-game NPCs and there's some other stuff in that mix as well but I believe those are like the main set of tools under Ace and then there's Omniverse and I think like some of the metaverse stuff has died down a little bit but eventually I think metaverse stuff will likely come back but Omniverse is a compute platform to create 3D worlds and it's also based on a 3D framework called open USD USD standing for Universal

scene description which is a format that was originally proposed by Pixar but it's kind of like a fa format but more than that in that it can describe and enable interoperability of 3D content in in Virtual Worlds but anyway not to go into too much more detail than that but for game developers who want to build 3D worlds or metaverse applications Nvidia has their omniverse kit which is similar to Nvidia Ace as SED of apps tools and capabilities to in this case with Omniverse help speed up workflows and enable interoperability for 3D worlds yeah USD is important it's been likened to you know HTML is the is the language of of the 2D internet USD is the language of the 3D internet the metaverse and so as that continues to build and we get closer to it it's a very powerful tool kit that Nvidia sits at the center of as well so these are totally really cool things and I also think it's important I mean this is where you get a little

bit big brained and this before all of the AI boom even when Nvidia was people were already alleging that the stock was overvalued before all of this part of the bull case of the story was not just that they're a chip company but they're an integrated systems and software company as well right and they have all these option alties with Enterprise software subscriptions across AI game development metaverse and Industrial internet and things like that right and I think when you put it all together what you really need to think about and this is kind of a a quote that I took from there was a long article on the New Yorker I think it was late last year or early this year from about Jensen Hong and the end of it was a quote from him that basically went along the lines of the future language of computing is human um and so all of these Technologies all of these chips and software and systems together The Guiding vision is always even I think Jensen's even been quoted as saying recently don't send your kids to learn coding anymore because in the future they're just going to talk to computers computers are going to write code for

themselves right and so I think when you think about it in that lens and you start to think about where all these pieces fit together you really get a you know not just a picture of a company that's a critical supplier of chips for AI but a company that's really reimagining how we interact with technology in the digital world right and just kind of Shifting Gears to Future drivers of growth but but maybe like a highlevel recap overview in terms of their lines of business and the way that Nvidia breaks out their lines of business their financial is data center which we've talked about and is like the biggest contributor at this point gaming which is largely their consumer gaming stuff and then you got these small areas right professional visualization Auto and other which what we're saying is like compared to Data Center and to a lesser degree gaming it's basically I think the growth and the value of this company is really going to be driven based upon the data center stuff or how they imagine in enabling the simulation

of human in intelligence in some way so iess it's like when you look at Microsoft and you say gaming Microsoft is relevant to gaming and even more so after buying Activision but gaming is still kind of not relevant for the investment case of Microsoft like it's a rounding it's an afterthought right it's kind of the same thing here is that this data center business has just gotten so large so fast that that's just where you need to pay attention right and so when we think about like the growth of the company with the key focus on data center I think the the big question is really around so how much more growth is there right so we're talking about huge infrastructure deployment to enable a lot of these capabilities such as what Tesla is trying to do with vision what chat GPT is doing and so what does like the longer term steady state revenue and margin picture look like and I think one thing for our audience to remember is that with most new technology we're probably going to see some form of Technology hype cycle where we kind of

get really excited about something over invest then there's kind of like this depression and then after the Depression we kind of figure things out and resolve over time how to fill that capacity that was over invested into and as somebody who many years ago worked in the optical networking industry as a Management Consultant I I saw this play out firsthand it's like everyone's building out all this fiber all this internet capacity it was over invested in and over time it just kind of filled itself out so anyway what do you guys think Brian you get do you have any thoughts in terms of future growth I just got some PTSD from Optical look I mean the acceleration of AI has ballooned super quick and and pretty Broad in terms of Industries it's happened just suddenly the financials Industrials Automotive consumer technology healthc care Pharmaceuticals just name a few and so consequently we're seeing a huge refresh of modern data center architecture and while gpos are huge part of that it's

also an interconnect and networking with with things like infiniband which you know they acquired with the melanox transaction that spending binge you just referenced it can be reactionary because many companies see AI as one of those Do or Die scenarios and we haven't really seen one of these at this scale probably since the early days of e-commerce but nearly every management team around today lived through that experience it wasn't that long ago and and they're spending hand spending hand over fist but this is just for like absolute size this is is A1 trillion doll installed base of Hardware that needs to turn over almost completely in the not too distant future so there will be some deceleration at some point and I don't think folks will necessarily spend blindly that's or seeing an Roi but we're already seeing a lot of green shoots where this is making an impact a good one is recently Clara suggesting that an AI assistant is now settling two-thirds of customer service requests a substantial amount of long-term savings even near-term savings or or American Express for instance and ring fraud detection accuracy by 6% using AI

um the notion of like as6 hitting the market and competitive products from AMD or Intel is certainly worth taking into consideration but there's a pretty severe first mover advantage in this and a lot of lessons learn from already being engaged with dozens of the biggest best companies in their own respective Industries and I would say like CTO generally don't get fired for buying the best of breed Solutions and these competitive threats come up with nearly every new business for Nidia at one point it was perceived to be a big threat for crypto mining with as6 and things like that but I think that's a little different because the spend binge popped before it really mattered but also in in the context of a one trillion dollar refresh there's room for more than one competitor out there but that's gonna be a hard one to knock down yeah I would add a couple things to Echo what Brian said but he covered it really well is if we're going from a market where maybe accelerated it was like plus or

minus 10 in the tens of billions of dollars and it's going to be hundreds of billions and Nvidia had 80% of it before even if that market goes to 100 billion whatever the number is from in the tens 20s whatever and their market share goes from 80 to 60 they're still growing exponentially right and so it really doesn't like it I think people expect them to lose market share expect there to be more competition but this is where I think you have to look at more than just a chip because already today there are chips that on paper are more powerful than Nvidia gpus for data science and AI right like on paper they can do things better than an h100 can but the problem is an h100 is not a standalone thing you don't just plug it in and poof you have an AI machine right you need networking you need storage you need memory you need all the different parts of a server you need you know fiber optics again right and and so when you start to think about how nid has built systems and not just chips how

they've added the software they control the Cuda ecosystem which is the default for AI researchers and a engineers in terms of the software interface to interact with these chips when you start to put it all together melanox being which in hindsite might end up being one of the greatest Acquisitions of all time when you think about how it's positioned inidia to control infiniband here as opposed to ethernet networking I mean all of this together is what differentiates Nvidia and so when people come out and say oh Broad com's got a chip coming out and it's more powerful well it that doesn't really mean because unless you can you can show me how that that's going to fit into a system and be more competitive on a price per price per performance basis because that's really the trade-off right yes broadcom can make a chip they can make an excellent chip hwan's an excellent operator and that company runs like a machine but what's the performance look like and what's the cost benefit and out what's the cost tradeoff for the performance trade-off of what you're getting by by buying that solution right and when you look around the landscape no one can compete with

Nvidia on a c per performance basis yet and for me the way that I'm kind of thinking about future growth for NVIDIA is really in in three areas and the first is really around current applications Brian you mentioned a lot of different verticals that AI is being applied to but and Matthew you mentioned kind of this notion of having this kind of conversational or dialogue based AI interface to data and to people including customer service like the CLA the the Clara service that that you mentioned Brian but there's also data analysis creation of charts and deep analysis over big chunks of data generative AI around content you mentioned fraud detection computer aided drug Discovery and Healthcare in automation of previously very manually oriented tasks so there there's like this current set of applications that we're already seeing that we're deploying against and it's got a lot of room to run and then I think the tricky part then is is to think about the second area which is around new applications and I think this is what's going to determine the difference between how much overhang there is before we kind of catch up when when we

for example when I brought up the example of having over invested in like Network bandwidth and fiber and things like that and so I think we're in the early days of figuring out what AI can do but I I think in a similar way we likely will see crazy new applications that nobody has thought of before and so when we think about the hype cycle and the evental additional growth that Nvidia has to run and to deliver GPU us and chips to enable the new applications that's the big question mark how much more is that going to be and then the third area I think about is around replacement and Clara is the example that that I really think about when I think about replacement so you mentioned the replacement Brian of uh customer service Representatives I think in if I remember correctly it was around like 700 customer service representatives were replaced and in fact AI actually outperformed the their human counterparts in several areas and so this kind of use case of human

replacement because AI fundamentally is about the ability to mimic human behavior and speech I think we should expect workplace replacement as a major application across a potentially huge part of worldwide GDP and on top of this it's not just like it's also going to be with things like Tesla Optimus and Boston D Dynam out there as well we're now also going to be able to consider real world interfaces with Bots meaning real physical Bots and once you add thinking capabilities that human intelligence to Locomotion and fine motor skills with fingers and hands and things like that all of a sudden there's a significantly sort of Greater Universe of human replacement that becomes possible and so then that should drive from an Nvidia perspective a big need

for vision and instead of llm models for a huge push into like imitation models to deliver against human replacement Tesla FSD V12 is a good real world demo of what's possible with vision and imitation model so then that should also Drive continued growth of nvidia's business I think the short of all this is that it's not clear how soon this run for NVIDIA hypergrowth ends without significant competition but I think the full extent to scale before we over invest could be a a really big run I mean in my opinion I'm I'm more optimistic on Nvidia but I think that we've got a really long way to go and on top of that as per some of the things that you just mentioned Matthew as far as like Infinity band and some of the other stuff where that's creating a bigger competitive mode which places Nvidia in a pretty Prime position to like really hold a fairly Monopoly

status over a pretty long period of time in my opinion yeah I mean I I agree with Brian though like I do think that enterprises will start to need to see an Roi on these Investments within the next 12 to 18 months yeah or that investment Cadence will slow like for example it sounds like Microsoft co-pilot's kind of been slow out of the gate for Enterprise or at least slower than some people expected those are kind of the incremental data points that you need to kind of look out for and say like we're building this infrastructure to support the software applications to run on top of it if the software applications aren't scaling people are going to slow down their infrastructure Investments right and so that's kind of where it's at so I fully agree there I'll also note that a lot of the supply chain bottlenecks that existed 12 months ago for gpus and specifically AI chips and is pretty much negated at this point I mean I think lead times are still longer than they traditionally were like pre chat GPT but they're still way below where they were like six months ago at

this point right we've had a lot more hbm3 memory Supply come online a lot more Coos packaging I mean I we can go deep into all the technologies that are involved in the supply chain but the pack the advanced packaging specific types of memory that needs to be used some of the networking components even some of the assembly by the O oems and the and the SMC of the world that are making the actual systems for NVIDIA a lot of those bottlenecks have been cleared up enough where Supply constraints are less of an issue as well and so that's also something that I'm watching and this kind of gets into the Outlook and and everything but when demand always outstrips Supply it's impossible to know where things end because keeping up with the supply chain and how much they can deliver is always a guessing game people try to do all their surveys but it's always a guessing game once Supply catches up that's when you're really guessing on demand now and it's not an easy thing to look at and that's why all these Canary data points like co-pilot and things like that are

important indicators of of where demand is going to continue to be right and so I I think if if I'm understanding you correctly Matthew like the two points you're bringing up one is like and I guess you originally brought this up Brian is like the economic case because if the economic case isn't there for a lot of these new applications and and for whatever reason you can't build a business out of these capabilities that you're enabling with AI then people aren't going to be buying more Nvidia chips I'm a little more optimistic on that end just because some of the specific use cases I'm seeing especially in from a game development perspective the value contributed by AI seems to be so large that it seems like at least for these current sets of applications that there are going to be a number of applications where the value is clearly there and you know exactly where that's priced and how much margin flows back to the developers of those capabilities it seems like it's I I'm I'm kind of optimistic there and do your second Point since be about bottlenecks and I

think besides Supply the other thing I would think about would be from an operational perspective and like do we have enough energy to power all of this stuff in the future and I know that's something that Elon Musk has brought up as a potential problem and I think he mentioned like Transformers as well as energy and things like that so we'll see if we run into some of those bottlenecks but that could be something that stunts their growth in the future as well well I will just add and I think the reason that Brian and I both brought that up is we come from the financial background where we always think about the worst case scenario and especially when a stock is up 7x in 12 months the the negative reflexivity of the stock price to any inary bad news is going to be huge right and so that's why you need to think these things through you know for example we're recording this it's March 14th happy Pi Day by the way and Adobe just printed and everyone thought adobe's going to be a big AI winner because like look at all the creative power that's GNA be Unleashed by AI but

it actually turns out that maybe AI is a huge competitive threat for them as well as an enabler and so Invidia stock actually did a I mean after hours is always after hours but Nvidia stock did a little drop and rally back after hours on that adobe print right and so you know that's kind of like that negative reflexivity that you have to be cognizant of that's just a good example of that as well yeah I think nividia is a really good and a really interesting demonstration on valuation there's a lot of ways that management teams try to drive shareholder value and share price increases but the number one always has been and always will be growth and we have not seen growth at this scale ever that the one trillion dollar market cap figure is still reasonably rare with only seven companies breaching that level it's it's Microsoft it's Apple Nidia Saud aramco Amazon Google meta and then just a little bit behind there is Berkshire tsmc and ilil rounding at the top 10 so so Nvidia at this point in

time is outgrowing all of those by magnitudes it might be a smaller company from a revenue perspective but the revenue growth is still you know usurping all of those things so over like the next 12 months we'll grow slow so almost absolutely but AI spend isn't going to dissipate in a crypto fashion and a theme and a trend as important as AI isn't going to be a one andone upgrade right like Nvidia has developed a really intense sales motion over the years from its gaming routes and pushing for freshes with stronger chipsets and software they not just standing still it's also really important to understand that we're in a bull market valuations are stretching higher across a ton of sub sectors that I follow with in Tech so if you're of the belief that we're due for a correction of course Nvidia will pull back on Big Market drawdowns but you know PMS will start looking for more value and turnaround stories whatever but fundamentally Nvidia is firing on all cylinders and when that appetite resumes those are usually the first companies to get bought so this boils down to really two main risks for

stocks and absent General market malays numbers risk like deteriorating fundamentals which is hard to see in the medium term with demand still delicately outstripping Supply it's narrative risk too or saying that another way if you believe AI is too hyped and not a durable theme but at this point I can't really get behind either one of those and Matt kind of alluded this already but the double-edged sword is if you're tethering your company to a hype train that comes with increased expectations if you want to receive the premium afforded to story stocks you need to accelerate and maintain above market growth and if you don't that premium can compress awful quick uh this isn't really relevant to Nvidia given they're the tip of the spear but I do think we'll see more single name dispersion going forward as some companies will inevitably miss that Mark an earlier Matt referenced an8 billion do Q4 and data center revenues that's up 27% sequentially in 4X year-over-year Nvidia last quarter if you go back to Q3 they

guided Q4 to Total revenues of 20 billion Street estimates pushed that to 21 billion right so like analysts over the course of the last three months raised those numbers so that the average was 21 billion so a$1 billion dollar beat and then the buy side was expecting a further 1 billion beat on top of that which Nvidia beat that beat is bigger than some Revenue lines for a lot of software companies out there that that are worth $20 billion so this again this is just like not something you really want to in front of and and just something that we haven't really encountered before at this scale you'd have to go back to like the tech boom and things like that the early one 2000s to see growth of this in any sort and much smaller companies mind you and then sometimes you'll see 400% growth in like a series a software company or something like that but not in the public markets and not 30 years after Inception it's just absolutely insane let me put the growth figures that you referenced and I referenced earlier just in maybe a perspective that other listeners of this podcast might better grasp they added roughly four

billion of data center Revenue sequentially in the last quarter that's two Monopoly goes so it's fairly profitable sorry had to yeah and uh compared to other companies though what in terms of like from a multiples valuation perspective Matthew you mentioned when we first started that they're not that expensive at at over two trillion in market cap I think just looking here over 70p ratio like what are the specific multiples or how should we be thinking about valuation for NVIDIA I mean I'll probably get on and this is gonna be probably a quotable moment when the stock drives from here but valuation really doesn't matter and here's why and Stacy rasan who's a very well-respected chip Analyst at Bernstein has has made this point for several years now you know through the cycles with Nvidia going up and down going back a decade the valuation has never mattered because it's always been wrong in the sense that even if you're looking

at forward multiples consensus numbers that go into forward multiples have always been lagging what they actually do for example if you look at how not only are they beating consensus numbers every quarter but there's like the whisper bogey numbers that like the buy side puts together and it's usually you know in the case of Nvidia it can be 10 to 20% above like what those consensus sell side numbers are and they're even beating those bogey numbers that people try to Benchmark to right and so you can look at it and say oh it's trading at 40 times forward earnings still that's expensive yeah but it's also growing earnings 400% year on year right and the and that 40 times is based on estimates that are probably too low anyways right and so that when I say the valuation doesn't matter I say that in the sense that the inputs to calculate that valuation metric aren't necessarily reflective of the reality that that that has come to be and many foresee will be right and so I know that's a very quotable line to say the valuation doesn't matter but in the sense of looking at these specific metrics when you pull up Yahoo finance or whatever site you're using you need to understand

that that's not telling you the full picture and it's not fully encompassing right if you kind of step back and you say how many companies in the world are gonna let's get through this hypergrowth cycle and just think can they grow 20% annually at some steady state maybe can they do that at 65% 70% operating margins which by the way are basically software margins actually potentially because they have this technology mode right and you start to think about what that looks like on a long-term basis the valuation is not egregious at that point yeah it it's kind of very speculative right because as we mentioned before when we think about future growth there's so many variables that can go into how much room to run do the current applications have nobody really knows what are the new applications that are going to be that could potentially Drive future growth and sales of G GPU capabilities and so yeah it does seem like to your point evaluation not that it doesn't matter in

my perspective but it's just like it's so speculative and so you would have to have a strong P to to I think to make that determination you'd have to have a strong perspective in terms of the growth and the overall scale of the business in a Ste steadier State longer term and then to the point on competition that we talked about before you know how much market share will Nvidia be able to keep in this major massive potentially once in a lifetime kind of like industry that that we're seeing and so maybe I could bring that point up again in terms of competition uh we have already talked about some of the things that you talked about Brian and Matthew in terms of you know whether it's Infinity band or Kudo and you know in their financials Nvidia talks about they talk about two things specifically with respect to competition which is like having a platform strategy bringing together the hardware systems software algorithms libraries together is kind of like a full stack offering and then they also mentioned that they've invested

over 45.3 billion in R&D with a bunch of inventions and things like that that are required for their current service offering and so from a competitive SP standpoint like if you're if you're a venture investor thinking to fund a company to go up against Nvidia do you need $45 billion dollar to compete with them do you guys have any further thoughts on competitors can anybody compete with Nvidia longer term How would how how much should investors be scared about competition for NVIDIA well it won't come from Venture right it's it's too big companies for every cycle of chips in the last yeah four decades well it might venture because remember trying to raise like a trillion yeah whatever that's not I mean whatever if Masa wants to do Masa things that's fine but I mean at at the end of the day there's only so much and a lot of that trillion would just go to asml

probably right because I mean to build a three what is it to build a three nanometer Fab it cost $30 billion now and I think two nanometer is going to be closer to 50 billion and they're talking one nanometer is going to be 100 billion so like most of that would just go to semic cap anyways which I've talked to you guys at length about why you know I have a love affir with semic cap but that's another conversation but in general like for all of this to happen all this competition it none of that changes the software ecosystem none of that changes the networking none of that changes the memory none of that changes the packaging the relationship with tsmc all of these things that exist today that individually might not be the biggest otes but collectively together is what differentiates Nvidia and you know I in the long run is it going to be 100% is it going to be 80% no I said that I think the market share of accelerated compute goes to something like 40 or 50% just like we saw with Intel right in the data center with CPUs

right but the Market's also going 10x larger so that's still very good and there will be Google A6 and Microsoft A6 and meta A6 and Chinese A6 and all these other chips and there will be an AMD chip and there will be a broadcom which goes into an Asic package right all of these things will exist but that doesn't matter if we don't answer all of the other questions collectively it's hard to think about many software sorry semiconductor companies that have successfully raised Venture and had some sort of public entree and and a huge market cap at the end of the day and it hasn't really changed over much in the last you know 20 years or so there's some things like mobile eye or something like that that that kind of grew and then got acquired but at some point in time they just get acquired there's there's that walkway check where it no longer makes sense to compete um I mean part part of that's also due to the just the state of the the semiconductor industry right you have

these large companies that you know for example take broadcom right fantastic company it's run operationally it's probably one of the best forign companies in the world in terms of how they manage the p&l but they gut like R&D and everything and then they just they buy R&D effectively right and so there's also an argument to be made and we saw this in the gaming industry too like in the late 2010s right there was a ton of m& which soaked up a lot of the talent and a lot of these upstart companies before they went public before or even before they hit significant scale as private companies right and so you've seen that in the chip industry as well and that's that's a consideration as well that like if one of these Venture back companies has a you know Eureka moment and something amazing they're gonna get acquired they like some Intel will spend God's amount of money to buy that or someone will yeah and the broadcom model is generally not startup driven it's more so finding the number one or number two competitor in any space that is kind of sleepy and doesn't really have a lot of competition and then raising the prices of shitload

and firing everybody but that's not how this is going to go I think broadcom's the wrong example there but I it just kind of was in my mind but Intel would definitely pay up for it yeah Intel do it you've seen other companies where they pay up for for other Technologies to buy versus build yeah yeah I mean competition it's funny because like if that's the major knock on Nvidia now this is the least controversial I've seen this stock in 15 years which is part of the reason why it's so extraordinarily large now but there was always something else competition for every single business year they've been in but this is the first theme I've seen that isn't being you know pooed away and and it's just like it's a very loved stock at the moment and and those things kind of tend don't not to reverse themselves in an instant and certainly not from like an upstart or a competitor but could be wrong all right well maybe we can end with just your thoughts I know Nvidia is

highly speculative there's a lot of variables surrounding the company but well more with respect to like valuation but in terms of the future outlook for the company what do you guys think where do you see this company do you believe in its future growth and continued increase in valuation moving forward you want to start Matt I'll maybe I'll give you my answer after we go after I go to GDC next we can see how many black letter jackets are worn around same right because if everyone's dressing like Jensen Wong then it's then we're in a bad spot right but no I think I mentioned kind of what I read in some of his interviews and everything and kind of his vision for technology and it is a big brain what is this guy smoking type of concept but we're already starting to see it right did you ever think that you could just type into a computer and it you know English and it could write code for you right like two years ago that would have been a foreign concept and it's like no that's the Matrix like that's sci-fi and here we are today and that's reality chat GPT and all of these AI chat Bots can take

regular text interface and spit out python or whatever other code you're working in right and so like we're already seeing some of this come to fruition today is it perfect no it's we're very early in this right but I think when you start to kind of think about how we rethink how humans engage with technology and digital interfaces and when you think about nvidia's position centrally to so many aspects of that that's that big brain argument that's like wow this this is not just a chip company this is not just a GPU company this is a company for the next phase of the Industrial Revolution you know the the AI Revolution the how we rethink that connectivity with digital Technologies I think it's also just like who's buying Nvidia is if this is from like a professional standpoint it's a bit more nuanced than just being able to own Nvidia right that's not necessarily their job there's your you're balanced against an index your fighing volatility your hedging and all these other things

but you know as like a as like a PA or like a idea for a single person is there the opportunity to maybe own this five or 10 percent lower absolutely it could happen the next five six months but if you zoom out look ahead a year or two I don't see a reason why this can't be the most valuable company in the world mainly because just again it's just growing like a weed and I don't really see any signs of that stopping in over the next 12 to 18 months and like dollar for dollar would would you rather own this or apple or meta or everything has its own little quirks and things like that but I've never seen something growing this fast in in a state where you're entering into potentially one of the biggest themes that we've ever seen and we're old enough now all of us are to make it through e-commerce and the internet and all this other crazy but this is just uh this is just crazier yeah I'm not GNA sit here pretend that market Cycles don't happen like you said ignoring Market Cycles here is the barricade well Market Cycles can happen happen right we just saw bad inflation numbers this week and so who knows what

that's going to mean to the long-term outlook for macro things and Invidia is such a large component in so many indices right it's going to be affected by the macro but stripping that away and just looking at the risk specifically for Invidia I agree with Brian that in the decade plus that I've been covering stocks and stocks like Nvidia this is the least fundamental risk I've ever seen for it I do think you have to watch some of those ancillary kpis like rois on co-pilot and things like that that'll be a good proxy for Enterprise sentiment around investment in data center infrastructure for AI but like other than that this is just it's I I agree with exactly what Brian said like look at Apple right how many Regulators are coming for Apple right now there's so many more risks than so many other stocks if you want to talk about some of the more speculative like AI stocks like super micro and Big Bear AI like absolutely let's that's a completely conversation but have you ever seen a company of the quality of

fundamentals of Nvidia we're looking at a software company that makes chips in terms of the profitability profile of the company and and they are Central and they have significant market share and they have competitive Moes like it's anyone that just says oh it's a bubble like you know they that they just don't know what they're talking about yeah I personally believe that the way I Fally think about this is just in terms of the potential potential scale for growth and for me AI is similar to the internet and there's some people who think it's even a bigger Innovation than the internet and so when we think about the potential growth potent growth from here in that Nvidia and I don't understand the competitive Dynamics well enough but it seems like they've got like a iron grip lock on this Market that this we're we're talking about potentially a once in a-lifetime occurrence in terms of Technology Innovation and you have a company that

has strong leadership and potentially a very strong competitive mode so I don't know I'm very optimistic I think that of any company that can hit 10 trillion in market cap I think this is probably it I mean I factor in a little bit of inflation there too but like I think it's for me it's Nvidia or Tesla and as far as the leadership is concerned we were kind of joking in our private slack about how much nobody cared about what Hong said until until Nvidia had its run and to some degree I think people do focus a little bit too much on the messenger rather than the message but from all this stuff I've been seeing about Jensen the more I learn about him the more I am impressed by him dude's pretty wise he's practical he's a Visionary and I've spoken out against a number of management teams on previous episodes that we've recorded here so it's not like a fanboy at all of most companies but here I would say I'm

jumping on the jinen bandwagon and I think he's a really strong leader and I think that matters especially for these bigger for these companies where if you have a leader like him you can really help drive it to a lot of growth and to help them maintain their competitive mode so long story short huge fan super optimistic fingers cross that Nvidia Hits 10 trillion and and the only other thing I would say I hope I hope so too because then I can retire and the only other thing I would say is like I hope we can also just because our audience and because we're pixels and profits with the focus of our audience of game developers that we see more coming out of Nvidia Ace and other Partnerships to help enable game developers to do more stuff with with Nvidia but that's just a personal wish list from me so with that is that any any other

parting comments or is that is that a wrap guys so everyone when you go to GDC next week count those black jackets you know if we get a if we get a large majority of people walking around San FR wearing black leather jackets that's that's your signal that we've hit the top that's my all right cool well with that guys we will catch yall later and that there you have it that's our take on Nvidia thanks guys [Music] [Music]