Beth Kindig: NVIDIA Set to Double | Bitcoin to $200K
Summary
NVIDIA's Transformation: NVIDIA is evolving from a chip company to a rack-scale systems company, which positions it as a key player in the AI sector, with significant growth potential in the data center market.
Market Outlook: The data center segment is expected to grow to $500 billion by 2028, with NVIDIA potentially capturing a trillion-dollar share by 2030, driven by advancements in GPU generations and AI demand.
China's Impact: The loss of China revenue is considered overblown, with NVIDIA's growth driven by demand from big tech companies and new GPU generations, rather than relying on the Chinese market.
Valuation Concerns: Despite high valuations, NVIDIA's comprehensive AI systems and first-mover advantage in AI software and hardware make it a compelling long-term investment, though investors should be cautious of market bubbles in AI software.
Investment Strategy: Focus on AI networking and energy sectors as NVIDIA's GPU advancements increase demand for networking components and power, presenting significant investment opportunities.
Bitcoin and Crypto: Bitcoin is viewed as a sentiment-driven asset, with a strategy of setting price targets and profit-taking to manage risk, anticipating a potential top in the $200K range.
Portfolio Management: Emphasizing the importance of active management and profit-taking to protect investments, particularly in volatile sectors like crypto and high-growth tech stocks.
Transcript
Nvidia is no longer a chip company. It's a rack scale systems company. That is why the stock will remain investable for the foreseeable future. The math is there. If we get to a 500 billion data center segment by 2028, that would imply depending on when you buy it, you know, 100% roomish or potentially more in the stock. Uh now if you take additional comments from the CEO this uh this this market could be up to 3 to 4 trillion by 2030. What's Nvidia's percentage? Probably a trillion dollars. [Music] If you have any questions about how to navigate the current environment, Wealthon can help connect you with a vetted adviser to get a free portfolio review. Just click the link in the description below or head to wealthon.com/free. There's no obligation and it will just take a few minutes of your time. Again, that's wealthon.comfree. Hello and welcome to Wealthon. I'm Maggie Lake. Joining us to discuss NVIDIA earnings and the outlook for AI in the tech sector is Beth Kindig, the CEO and lead tech analyst at the IO fund. Hey Beth, thanks so much for being here. >> Yeah, thank you for having me. So it doesn't get any bigger than Nvidia when it comes to earnings because they become such a bellweather not only just for the tech sector but I really I think broadly the overall US equity market itself and we had the chipmaker still reporting huge numbers sales up 56% but investor reaction in the first 24 hours at least seems to be a little mixed. So I'm really curious to hear what you think of the numbers and what you heard on the conference call. >> Yeah, I love talking Nvidia and no better day to talk Nvidia than today. Uh, one thing I would say is if this was just based on Q2, this stock would sell off much more than what it did. Uh, Q2 did actually report a decline, minimal, but still a decline in compute, and we're talking about the global leader in compute. You cannot report a quarter over quarter decline. Now, of course, they did that because of China. That was what impacted it. Q3. So that forward look saved the day because Q3 is indicating we're at the low uh the cyclical low. It's a semiconductor company. Uh it's much better to focus on GPU generations than earnings >> and we're at a cyclical low for the hopper generation. Blackwell is going to start impacting earnings in a significant way starting in Q3. The evidence there to illustrate is we'll see a 15% quarter-over-quarter acceleration on data center and networking was exploding in Q2. I would imagine that continues to explode in Q3. That's an important clue because that is what separates the current generation of GPUs from the next generation is the amount of networking that is required for those systems across the board. Uh again, that Q2 is pretty weak. It really is this Q3 that's saying as a semiconductor company that puts out these larger generations of GPUs, it's coming. We're shipping and it's green lights across the board. Green light across the board. That's the piece that really saved this company. One thing that I was really focusing on going into this, it's always a great time in between GPU generations to go back and just look at like total addressable market. How big can the data center get? what are we expecting from this generation of GPUs versus the last which was clearly explosive so that people don't lose that line of sight around around what where this company's likely headed. >> Yeah. And it's easy to do that because there is there's you know everyone jumps on the headline of course and then the numbers are so big and there everyone's got a narrative that they're married to around this AI story which we're going to get to in a second. But you mentioned China which seems to be incredibly important. How do you or tell me how important this is going to be? They're not selling it to China now. Will they be able to, you know, there's a lot of politics involved in this. Will there have to be some revenue share? There's so many unknowns. How are you thinking about that piece of the puzzle? >> I feel like China is overblown. I think it's overexaggerated. And I can give you numbers on that. China in this last quarter was a 4 billion impact. that black wall generation that's just now beginning to ramp was roughly 28 billion according to our model and CFO previous commentary you know a over 100 billion run rate on one skew versus maybe you know 15 billion was the annual amount for China what we should all be focusing on blackwell and I like to really keep that in perspective for people because the sheer amount of noise around China is you would think this stock is over because they lost China revenue. For me, I think it's pretty insignificant. And I think there could not be a better time to lose China revenue than now. Uh because if they had lost China revenue uh during that trough, which was again when Hopper sales were starting to cool off, big tech is waiting for this new generation of GPUs. The impact would have been much more pronounced. Um that's like your May earnings call for example. So, it was nice to have China sustain and elevate those sales. Uh, but if you were going to lose China, now is a great time when those new systems are shipping. There's such outsized demand. They can't build them fast enough. And therefore, I and of course, I'm talking um, you know, generally here, but their numbers also support that. When you look at Q3, they're not recognizing any China revenue. it it was a sizable beat compared to analyst estimates because again Blackwell is shipping and we have that 15% quarter-over-quarter acceleration. They're saying in numbers that they're good to go. China at zero doesn't impact that quarter quarter growth that we want to see in the second half of the year. >> Well, that that's going to be reassuring to a lot of investors to hear. Is there a potential for China to become a positive surprise if somehow that market does open back up to them? Yeah, I think China will always be an opportunity for Nvidia, although if I were to just broaden this discussion and I actually do um talk a little bit about Asia. I'm on Bloomberg Asia. It's located in Hong Kong. And so I do think about the Chinese markets, you know, more deeply than maybe just a United States investor. And one thing that comes to mind for me when I think about China is they have some pretty big tech companies. They have Alibaba, Tencent, BU. Um, they've done well on EVs with BYD. They've done well on mobile with Huawei. This is a country that has always sought to be self-reliant on technology. Uh, I don't know why they wouldn't want to be self-reliant on a market that is larger than e-commerce, larger than social media, larger than EVs, and larger than mobile. Uh, I have about 3 to 4x larger market than mobile for AI. So I I don't know a path forward for China given their history that would indicate they're going to be dependent on United States silicon. >> Yeah. And certainly from a from a political geopolitical point of view that does not seem to be in the game plan at all. So uh it sounds like you're not you're not like counting on that to come back online. You're you're looking at that as a homegrown story and not something that's necessary for Nvidia or some of these other international chip players. I would say the supply chains are going to are uh onshoring supply chains is a significant undertaking. Um it would be that the manufacturing side that could be impacted more than Nvidia's line item on their sales onshoring that will take maybe a decade and you know entirely and how the global tensions approach that piece is probably the biggest question mark for the semiconductor industry. Mhm. >> Um I see that as se a separate issue than did Nvidia get to recognize H20 revenue from China this quarter. It's more of this bigger discussion of like how do you onore these deeply entrenched supply chains that are global uh and very reliant on China. As part of our silver interview series this month, we're bringing you top tier research in a free silver investment report. To claim your free silver investing report, simply click on the link in the description below. And if you're looking for a simple, secure way to invest in physical gold and silver, check out Hardass Assets Alliance at hardassetsalliance.com. Speaking of silver, Wealthon will be on the ground at the SCP Resource Finance Global Silver Conference this October in Toronto where Eric Sprat will deliver the keynote address. It's going to be a major event for silver investors. So stay tuned for more details in the weeks ahead. >> Yeah, absolutely. And this and this is where why sort of macro matters more now right and certainly geopolitics because of because of the sort of just scope of that challenge. Uh so the other issue I think that that is really front and foremost for investors with Nvidia is valuation right and so how are you balancing and nobody knows the sort of the the details of of I think this this sector and the Nvidia story better than you. So how are you balancing the fundamentals versus this valuation question? And I'll I'll share an anecdote. We we just had Jim Bianco on and he was talking about um you know just broadly the equity market and some valuations and he was at a a cocktail party and uh someone they they were all talking about how they own Nvidia and he just asked them like oh what do they do? They're like I don't know but it's a five billion it's up 500% this year and he's like yeah but what do they do? and they had no idea nor did they care you know so it's just like sort of retail chasing the momentum of a really really high-erforming stock and I think that makes people worried about some of the valuations around Nvidia so how do you balance those two things >> yes well in brief uh Nvidia is no longer a chip company it's a rackcale systems company um that means that it's a little bit of software in those systems there's networking it is how GPUs communicate to other GPUs, that is why this stock will remain investable for the foreseeable future. Um, and the more GPUs you can combine that think like one GPU, the higher up you can go in training models. Um, now you can break through from let's say a couple hundred billion parameter model to a one trillion parameter model. AI models can get smarter and smarter. um they are the parallel processing um king and that means that they introduce something called tensor cores that allows things like matrix multiplications um floatingoint precision these little features or not features these massive features transformer engine they have created a lead around training that is insurmountable >> uh and the reason that they were able to do all of that is they were so close to parallel processing from day one with gaming and then they just built out more parallel processing power for AI. And uh now when you look at what's different about Nvidia and how the AI economy will be built out versus something like mobile or maybe very similar to mobile actually maybe I can make an analogy here if I think about it. Uh I believe those that are that deeply entrenched in the hardware stack like Nvidia will also be dominant players in software. Um you can think of this as Apple being very dominant with the iPhone but also dominant with iOS and then having an app store something similar to that. I I don't want to pull a perfect analogy other than uh the fact that there's more to this story than just uh you know a single chip. And when I think about valuation, that's just like a broader thing. Like NVIDIA is not a chip company. Um, I just want to throw that out there. It's an AI systems company. Uh, it's powering all of AI training because of those significant advantages on product. Uh, specifically parallel processing power and they have a leg up, a first mover advantage when it comes to AI software, which will be as big as hardware. So, uh boy, like how could I how could I >> underscore exclamation point exclamation point like yeah this this is so and that's a huge statement. So and I think it's really important because I just sort of shared an anecdote about how many of us do not really have a a complete understanding of what's going on um underneath the hood which is why we're so happy to have you here today. And so that that that seems like a very different beast when you're talking about the ability to sort of scale and grow with this monster revolution, which is what I'm hearing you say. But so so again, it's this sounds like this is a stock that we all should have in our portfolio if they're going to be that critical to this to this big revolution. But how do we think about that in terms of valuation? Because I think that's what has people nervous. Okay, maybe I own it. Do I add to it? Maybe I never got in on it because I was burned by the dot bust and I was too scared. Like, do I do I start to buy it here or do I wait for pullbacks? Like, how do you guys think about that strategy management verse of of the sort of valuation versus that incredible fundamental story you just laid out? >> Yeah, that is the most important question is that we have the world's most valuable company here. Can it continue to grow? And I can go through that for you. I do want to say that our process is extremely detailed. A stock like Nvidia, we go through over 100 point checklist at least eight times a year before earnings and after earnings. I probably write maybe 25,000 to 40,000 words on that stock every year. Um, you know, we're talking very very big deep dives that constantly question, should this be a top position of ours? What's the allocation? And the way I answer that for my team, for my research members, and when I can, you know, on conversations like this is is here we go. When we look at the data center today, it's about 48 billion expected next quarter. Let's just call it 50 billion per quarter uh round up. So that's a 200 billion run rate. The reason why this stock has room is because analyst estimates are too low. when you try to do financial modeling on this uh opportunity Nvidia it's it it just crashes the models and once we hit that 50 billion data center quarter which I had predicted six quarters ago we would glad to know we're there that's a little mile marker for us I believe by the end of next uh calendar year we'll be at 75 billion that creates room in the valuation because anal estimates are pretty low. Now, let's broaden the discussion that Blackwell and Blackwell Ultra, the next two generation of GPUs should bring us to a 300 billion data center, up from about a 200 billion. Now, where the market and or analysts are getting it wrong is they don't see Nvidia hitting 500 billion until like 2034. That's where the e that's where the most room is in my opinion. I believe Nvidia will hit a 500 billion uh data center by 2028. And the reason and it was really interesting because I had just talked about this and Jensen Wong echoed it and I love when like you know I'm so deep into a stock that the CEO is putting out something around the same time. He was saying you can expect about a 50% kagger. That's what will bring us to 300 billion in six quarters. So now the math is just there is what I'm trying to explain. the math is there >> and analysts don't right now can't model that uh because they would have to rely on so many assumptions. If we get to a 500 billion data center segment by 2028 we're at roughly 200 billion now you know that would imply depending on when you buy it you know 100% roomish potentially more in the stock. Uh now if you take additional comments from the CEO this uh this this market could be up to 3 to 4 trillion by 2030. What's Nvidia's percentage? Probably a trillion dollars. They're about 33% of the market somewhere in there. Uh now here's the thing and this is where I can help people reframe. This isn't actually Nvidia doing this. It's big tech that is doing this. They are spending. they are opening their wallets and they are saying I don't want to be left behind and for example Blackwell if you don't get the Blackwell system let's say you're one of the the three bigger cloud infrastructure companies or maybe you're something like um you know XAI if you don't get the NVL72s you're going to be left behind because they are combining 72 GPUs versus eight what you can do is significantly better four times better training 30 times better inference at they're going to open up their wallets to make sure they get those systems. And it's really about big tech. That's where I would remain focused. And when big tech starts to cool off, we still have enterprises that aren't able to even get these systems. Um, so to put it into perspective, capex just grew 23% quarter-over-arter. Um, so I would think of it as truly the buyers driving this forward and not put so much pressure on how could Nvidia be that great? Could Nvidia really XYZ? it it's it's it's a marketplace where the buyers are scrambling to get the critical advantage of this next generation of GPUs and they'll do it again when Reuben comes out. Uh so I would think of it more broadly focus on the demand side. Is there anything that you see in the future that would cause that to cool off or change >> because Nvidia for sure we actually saw this last year. I I don't know how much the news picks up on all of this in granular detail. We actually saw what Nvidia's bottleneck is, which is that the product roadmap is moving so fast. This company is so aggressive in what they really want to achieve that they can run into this last year was thermal and power management issues. The Blackwell generation was supposed to have been here already. We're looking at roughly a six-month delay and they resolved those issues and they're moving forward. Now I would just say uh by breaking the upper limits of what is possible and having so much pressure on you as a company that engineering side doesn't have the time to fix everything that could go wrong. So that would be one. Mhm. >> Secondly, uh where we are starting to position is energy because if you track this by GPU generation, which I kind of, you know, started with in our conversation is maybe Wall Street is forced to track this company by quarterly earnings. If you can track it by GPU generation, I think you'll be a much more successful investor because it's each one is unique in their own um you know what they're achieving. And what is unique about uh Blackwell and Blackwell Ultra? I would say if I had to choose one thing, it'd be networking. And then if I had to choose what's unique as we continue to move down further generations, I'm going to go energy on that. Um, can we when when will the grid be maxed out? >> U, what generation of GPUs will do that? My best guess is it could be Reuben. Um, just because we're looking about between Reuben and Ruben Ultra from where we are today, let's, you know, 3 to 5x more power is required to power those systems. How are you going to get the power? Can you get it fast enough? Um, the transmission of the power, is it close enough? A lot of questions there and that would pro probably be the bottleneck would be just the size of this market. How do you deliver these systems that are um I I call it like Nvidia 3.0 uh if Nvidia 1.0 was gaming and then Nvidia 2.0 is tensor cores with you know ampear hopper you know this is Nvidia 3.0 now which is these rack level systems. Uh just continuing to break the upper limits of how many GPUs can you put together and be successful at it uh without breaking something. And then uh from there it's and and having to work through that challenge. And then from there it's um how do you power it? >> It's so you're you're hitting on a topic that's near and dear to my heart. I um a few months ago did an energy conference where all we talked about a Reuters conference central to that in every conversation was the relationship between energy and hyperscalers and how people are going to try to solve for this enormous problem and it's just wild some of the some of the things that are going on there. So, it's so interesting that you're focused and and zeroing in on that as um you know, listen, it's a hurdle, but it's also an opportunity, right? Um if you're if you're watching the right players, um I'll ask you about that more in a second, but just just to sort of talk about this this sort of AI spend and and what might slow it down. Um are there does everybody win in this? I don't I'm I'm trying to think of a really simple way to say it because I know I'm I'm thinking about people listening to this conversation and I keep coming back to the sort of valuation side and it's not just that but there are people and there are plenty of people littered all over any kind of you know financial media or finit or all over the place who will make an argument that this is just like an enormous AI bubble reminiscent of dot not the same right doesn't repeat but it rhymes and that while this might be a disrup technology. There's a lot of hype and there are going to be a lot of losers in this and there's a lot of money chasing just anybody who's throwing AI in the conversation and I know that you understand this space so well. You know, does anything feel bubbly here? You know, for people who don't don't subscribe to you or don't understand this space as well as you do, but they're trying to think about exposure to tech. Is there any part of this universe that feels like a bubble? >> Yes, absolutely. I I do agree with that. I would say um like AI software is really in this R&D stage and there are there's a lot of potential. Uh but until we start to see real revenue growth and AI numbers uh discussed in these earnings calls, that's where I think the bubble is primarily is at the software layer. uh where the picks and shovels like Nvidia, Nvidia suppliers, which there are many Nvidia suppliers, they they really are in uh that that they really are down uh downwind from the massive capex spend where others are just trying to figure out where they fit in. And distinguishing that is key. It's always been key. And I really try hard to not move into a position unless they are participating um participating right now. >> And so so two things are going on. There's there's so much going on with this AI market. One is >> the quants and those looking for a quick trade will rush into a stock um that will pop the stock price. Those who aren't looking deeply at the fundamentals will say, "This must be an AI stock. It's up 30% today." And they'll run into rush into it. That's not our process, of course. We are, like I said, over 100 point checklist. We are going really, really deep here. And uh off the top of my head, I can usually rattle off those with not only the most AI revenue, but AI growth because that's something that really helps inform our portfolio positioning. And so I think that piece can get really confusing. >> I would say it's not unique to AI. We saw that with cloud. Um I remember gaming, there was a moment the gaming industry, everyone thought it was going to last forever. It didn't. Um, so people will talk about the dot because it really was a public market uh driven um you know tech trend. There's almost every tech trend 90% fail. That's just tech. And whether you were in the private markets and you see it or if you're in the public markets and you saw limited trends like which would be internet, mobile, cloud, uh it's happening all the time I would say. So nothing unusual there. uh except knowing in or having that information as an investor would mean that your process should really be I need to find the 10% and that mindset >> of I need to find the 10%. Now, when it comes to is there actual is there actually going to be applications and you know is is there going to be what I would call product market fit. Is there going to be a moment where AI is making a meaningful impact besides just a chatbot? Um where that impact will be and why this is so attractive to companies like big tech and different from the other trends we've seen on the public market. So again, that's internet and mobile is that this is not a consumer technology first and foremost. It's an enterprise technology. A company like Meta can spend I I think they're at a I think they're spending almost 25 billion a quarter right now. Um it is mind-blowing. uh they can do that because they're going to cut back on developers internally as well and they can optimize internal processes and then they can optimize their own product which is their their ad engine and by the time that you do all of that you are not looking to get consumers excited about AI uh that's not first and foremost you can instead just make your own product products better and you can work hopefully get developers to build foundation models with Llama for Meta's example >> and you can cut back on your developers and your headcount. So, by the time that you do all that, they're they're not trying to come out with like a, you know, a social media um app feature. They they do have an AI app that they're trying to incorporate, but they'll even tell you we're not expecting any revenue from that this year or next. no meaningful revenue but they will at the same call earnings call say but we do think that by the time of next year AI will be able to replace a mid-level engineer so they're they're telling you the our first move is to replace mid-level engineers and we don't know when the consumers will will adopt our AI app that so it's does that solve the moat question because it like I hear that come up all the time well they're spending all this money they're not going to be able to build a moat where they can get all the revenue from it Sounds like you're describing they're spending all this money to really transform themselves into this super company um efficient um better better able to operate um and so that maybe the moat question isn't as relevant or does it matter? >> They are definitely building their fortresses. I guess that's that that's a big deal for them for big tech. They're all about the fortress and what AI can do is just make their ad engine for example meta. I'm just using meta. similar for all of them. So superior to the mid-level players um that they continue to take, you know, some kind of market share. So in this case, let's just use ads. uh the targeting is is so much better and the visibility around the conversion is so much better with AI that this company that you might have said is going to plateau which you could argue it plateaued >> uh somewhere about you know um 2018 2020 uh if you look at some of their key metrics average revenue per user daily you know that kind of thing it was was plateauing a little bit there they could now um continue to grow because their ad targeting is just so much sharper with AI. Um that piece of continuing to grow when most of them, let's just assume had plateaued. We know Azure would have plateaued because when they break out the AI revenue, if you were to remove the AI revenue, let's just call it Azure core that was plateauing and not keeping up growth rates. Um so they get to ex start growing again as well. In addition to building more of a fortress, more of a moat, I would say though that uh I guess while we're on this topic, I would say that AI is actually a threat to their moat. Uh because not necessarily on their core products, but they are pretty concerned in my opinion as to if they can become the foundation model that developers use to build their applications. That that piece is front and center. OpenAI's done a great job there uh in terms of their API usage and that piece can really tip the scales in terms of moat uh if you just have so many developers using you know a specific R&D company's APIs um that that could threaten a moat. So I would also say it's a little bit of both. It's offense but definitely defense as well. Hm. So, as you look at this, it still it sounds like you're still believe hardware is is sort of the critical space here. Um, if if if someone has new money to put to work, um, are there other parts of the ecosystem that look poised to be interesting or where you see opportunity or some like smaller names that are not on people's radar? Where are you looking for like the next opportunity? Yeah, if I were to like start to talk about how how am I positioned and there's a few ways that I'm positioned. I suppose if I think about the portfolio, the number one way is AI networking and it goes back to what we first discussed with Nvidia GPU uh G you know Nvidia moving from a chip company to a rack scale um company which J basically just means that they're moving from eight GPUs communicating to 72 GPUs communicating. And in a nutshell, that's I I really try to keep things simple because tech you can get in the weeds with technical jargon, but to keep things really simple, uh that's a networking problem to solve. Uh among many things, but networking for sure. Uh to put it maybe in some rough numbers, you're looking depending on the component within the networking um stack, you're looking at 5x to 9x more of you know whatever those components are, whether it's rettimers, fabric switches, whatever that could be. Um and the reason why that needs so much more comm um more networking components and just even cabling is because the GPUs have to communicate to each other. there's just a lot more GPUs now to communicate. Uh now, how do you do that without signal loss? Um there's just lots of um there's lots of questions to to to solve >> in terms of how do you how do you move from that system to that and we know it was a complex problem to solve because Nvidia had a six almost ninemonth delay and had to really double down and figure out how to release a system that didn't overheat. Anyways, with that said, uh AI networking is where I would say we have outsized allocation right now. As we move into the second half of this year and into the foreseeable future, I don't know when we would change that piece other than now is a great moment to look at AI networking because if you go back to that comment that the best way to look at Nvidia and even AI is GPU generation and GPU generation for now. Um that is what is um front and center in terms of the biggest opportunity in my mind. >> Yeah. Um I'm I'm I'm sure everyone's reckracking this and taking notes because I think we all need to reset the way we we approach this. Um so I'm so happy that you're giving us some of this language to really sort of educate ourselves and dig in. So are you looking at since you see energy as a big issue here? Are you are you also looking at that space from an investment point of view or are you just looking at it from the sort of fundamentals? I need to watch what's happening here. >> Yes. Where we started to move into AI networking almost a year ago and uh meaning early. We try to be you know reasonably early here. That is the way you you know what real wealth can be made. I we're in the early stages of moving into energy which means we've dabbled with a couple of stocks. Some have actually been quite successful because these big tech companies are gobbling up whatever um alternate power alternative power they can find. >> Um and then you know just going back to what are we talking about here? >> Um I would model at least 10x growth for the right energy stocks between let's just say Blackwell and 2030. Um, Blackwell is about 100 kilowatts more or less depending on how you build the system. It's breaking the it's breaking through 100 kilowatts could be 140 kilowatts. We will end the decade with one megawatt. So there's about 10x growth in there and just energy requirements on GPUs. So you could, you know, reasonably say it should be about a 10x move. Uh so if if an energy stock is chosen correctly, I think it could have a 10x move in it. And it's really that urgency that's kind of interesting for an investor, which is those systems I was describing. They as that are shipping now, they cost about three million. Um you know, if you're buying multiple racks, you're into the billions. Uh potentially every week you're paying for a couple billion in racks. um or you know every week Nvidia right now is shipping three billion in racks. So, but let's say every month you're you're spending that and it just continually accumulates th eventually those would have to sit on the shelf and or sit stagnant. These extremely expensive systems if you don't have power, if you're not supplying power um so big tech is a big tech and I would call the neoclouds, it's like your core, they're going to want to get there early. They're not going to wait until it's too late and they now can't power these GPU these GPU systems. they're going to want to think, you know, a year or two out as well. And then if you think about Reuben, which should be about 3 to 5x from where we're at now, I think if I had to guess, uh, that's where we run into serious problems. Therefore, any, you know, accumulating of the power right now is probably preparing for Reuben if I were to guess because that's about one to two years out >> which is so interesting because right now if you just look, you know, broadly, the energy sector is pretty beat down and investors are kind of down on it. It's out of favor. So, you know, it's is why it pays to kind of keep your eye um really fixed on the future. Um I want to hit while we have time just another area speaking of energy demand. Um another area that I know is on your radar and you watch is uh which is crypto. Um how you feeling about Bitcoin at these levels is your interest elsewhere? How are you thinking about that space? >> Bitcoin has uh Bitcoin and crypto has a completely different process at our company. uh we feel quite strongly that it's sentiment driven and the difference between something that's sentiment driven and I'm just talking crypto broadly of course Bitcoin being the heavyweight there is there's no earnings reports there's no financials you can definitely pull some key metrics uh and try to compare them and people do but there's no real moment that you can say this intrinsic value is changed in a uh standard way. Therefore, um technicals work really well, we've found in in terms of really riskmanaging crypto and we firmly believe in riskmanaging crypto. Uh because even your juggernaut of Bitcoin, it can go down 60%. Will it go down 60% this next time with ETF participation? Maybe not. But there's a lot of volatility in crypto and people forget that. And therefore, we are big advocates of setting a price target years in advance and literally years in advance. Uh and when it hits that target, you go ahead and you trim with the goal of buying lower. Uh we believe a uh near-term top could be seen in the 130s. Uh that's a moment where our firm who's up about you know 600% some somewhere in there we're going to take some profits profit taking and we will put it back in to work lower now this is extensive work between an AI engineer um named Vincent Duchain who was early to AI I mean he's he's the OG AI kind of type where he was selling like companies to terodine uh before Nvidia's big moment he's OG kind of AI, not AI is here. I'm just going to jump on the bandwagon, which a lot of software developers have done. He's he's like OG. And then Knox, who's constantly working with technicals. Those two have combined forces to really figure out how do we risk manage Bitcoin? There's we have too much money in it to ignore that. and they're coming up with uh the strategy that profit taking in the 130s is not a bad idea with the uh with a with a minor pullback on the horizon. Load back up, go into the 200s. What's interesting is both of their models are saying right around in the 200s, this thing should put in a meaningful top, >> which I find interesting because you'll hear people say 1 million or People go big on those price targets. >> Yeah, they certainly do. Are you are you um is that in your portfolio because you were early and like you said you just have too much money in it and you have to manage it or or do you see this space as opportunity because there's a lot of conversations about ETH now that we have institutional participation you know people really trying to look at these these ecosystems which is really what they are to see functionality and what might you know you get you have a lot of enthusiasm and people building earnestly in that space. So do you see that more broadly or are you mostly positioning managing that like early right call for for Bitcoin? >> No, we participate in in additional coins. I would say our crypto mix to stocks is usually 70 stocks, 70% stocks, 30% crypto. Uh we've always been actually quite aggressive in that way. This is before there were ETFs. The minute we launched our premium site, we recommended Bitcoin and told pretty conservative investors, you you have to put Bitcoin in your portfolio. Uh, and that was so and Warren Buffett was calling it rat poison, that kind of thing. uh I feel there is enormous opportunity in crypto where we are a little bit different is we think it's prudent to book profits particularly in a volatile asset class like like crypto. I would say the same for uh some of the hyperrowth networking stocks. we are not um we are not interested in holding the bag. And if you can put that first when you're a tech investor, it seems it seems in the reverse, right? It's like I got to make money. I got to make money. That's how what I'm doing as an investor. It's it's actually smarter if you could say I want to protect my money. I want to protect my money. Uh it's just a little bit of a different mentality. and we've put it to the test um where we said can an actively managed portfolio keep up with a long-term buy and hold plan and we have exceeded those portfolios in terms of performance that go long-term buy and hold because like you're talking about now we have a lot of stretched AI valuations what's the plan >> are you going to continue to hold are you going to book profits are you just going to take that hit uh you know those are questions that I think are as important if not more important than the stock picking side and and and especially with crypto. >> Absolutely. I mean, you can't say it any better and I think this is a super important time for for people to plug in on this. Beth, we I love your level of rigor on all of this. It's so fantastic to be able to talk to somebody who's been following space and so smart um in how to look on it and I think uh look at it rather and I think we can all learn a lot from that. So, thank you so much. Always such a pleasure to catch up with you. >> Thank you so much, Maggie. really appreciate the conversation today >> and like I said everyone's I know going to be rewinding and taking a lot of notes on this and when you do if you need or want to take a fresh look at your allocation to tech and now that you have some really good questions to go and talk to an adviser about just a reminder you can get a free portfolio review with an adviser in the wealthy network just head over to wealthon.comfree. Thanks everybody for watching. We'll see you next time. [Music]
Beth Kindig: NVIDIA Set to Double | Bitcoin to $200K
Summary
Transcript
Nvidia is no longer a chip company. It's a rack scale systems company. That is why the stock will remain investable for the foreseeable future. The math is there. If we get to a 500 billion data center segment by 2028, that would imply depending on when you buy it, you know, 100% roomish or potentially more in the stock. Uh now if you take additional comments from the CEO this uh this this market could be up to 3 to 4 trillion by 2030. What's Nvidia's percentage? Probably a trillion dollars. [Music] If you have any questions about how to navigate the current environment, Wealthon can help connect you with a vetted adviser to get a free portfolio review. Just click the link in the description below or head to wealthon.com/free. There's no obligation and it will just take a few minutes of your time. Again, that's wealthon.comfree. Hello and welcome to Wealthon. I'm Maggie Lake. Joining us to discuss NVIDIA earnings and the outlook for AI in the tech sector is Beth Kindig, the CEO and lead tech analyst at the IO fund. Hey Beth, thanks so much for being here. >> Yeah, thank you for having me. So it doesn't get any bigger than Nvidia when it comes to earnings because they become such a bellweather not only just for the tech sector but I really I think broadly the overall US equity market itself and we had the chipmaker still reporting huge numbers sales up 56% but investor reaction in the first 24 hours at least seems to be a little mixed. So I'm really curious to hear what you think of the numbers and what you heard on the conference call. >> Yeah, I love talking Nvidia and no better day to talk Nvidia than today. Uh, one thing I would say is if this was just based on Q2, this stock would sell off much more than what it did. Uh, Q2 did actually report a decline, minimal, but still a decline in compute, and we're talking about the global leader in compute. You cannot report a quarter over quarter decline. Now, of course, they did that because of China. That was what impacted it. Q3. So that forward look saved the day because Q3 is indicating we're at the low uh the cyclical low. It's a semiconductor company. Uh it's much better to focus on GPU generations than earnings >> and we're at a cyclical low for the hopper generation. Blackwell is going to start impacting earnings in a significant way starting in Q3. The evidence there to illustrate is we'll see a 15% quarter-over-quarter acceleration on data center and networking was exploding in Q2. I would imagine that continues to explode in Q3. That's an important clue because that is what separates the current generation of GPUs from the next generation is the amount of networking that is required for those systems across the board. Uh again, that Q2 is pretty weak. It really is this Q3 that's saying as a semiconductor company that puts out these larger generations of GPUs, it's coming. We're shipping and it's green lights across the board. Green light across the board. That's the piece that really saved this company. One thing that I was really focusing on going into this, it's always a great time in between GPU generations to go back and just look at like total addressable market. How big can the data center get? what are we expecting from this generation of GPUs versus the last which was clearly explosive so that people don't lose that line of sight around around what where this company's likely headed. >> Yeah. And it's easy to do that because there is there's you know everyone jumps on the headline of course and then the numbers are so big and there everyone's got a narrative that they're married to around this AI story which we're going to get to in a second. But you mentioned China which seems to be incredibly important. How do you or tell me how important this is going to be? They're not selling it to China now. Will they be able to, you know, there's a lot of politics involved in this. Will there have to be some revenue share? There's so many unknowns. How are you thinking about that piece of the puzzle? >> I feel like China is overblown. I think it's overexaggerated. And I can give you numbers on that. China in this last quarter was a 4 billion impact. that black wall generation that's just now beginning to ramp was roughly 28 billion according to our model and CFO previous commentary you know a over 100 billion run rate on one skew versus maybe you know 15 billion was the annual amount for China what we should all be focusing on blackwell and I like to really keep that in perspective for people because the sheer amount of noise around China is you would think this stock is over because they lost China revenue. For me, I think it's pretty insignificant. And I think there could not be a better time to lose China revenue than now. Uh because if they had lost China revenue uh during that trough, which was again when Hopper sales were starting to cool off, big tech is waiting for this new generation of GPUs. The impact would have been much more pronounced. Um that's like your May earnings call for example. So, it was nice to have China sustain and elevate those sales. Uh, but if you were going to lose China, now is a great time when those new systems are shipping. There's such outsized demand. They can't build them fast enough. And therefore, I and of course, I'm talking um, you know, generally here, but their numbers also support that. When you look at Q3, they're not recognizing any China revenue. it it was a sizable beat compared to analyst estimates because again Blackwell is shipping and we have that 15% quarter-over-quarter acceleration. They're saying in numbers that they're good to go. China at zero doesn't impact that quarter quarter growth that we want to see in the second half of the year. >> Well, that that's going to be reassuring to a lot of investors to hear. Is there a potential for China to become a positive surprise if somehow that market does open back up to them? Yeah, I think China will always be an opportunity for Nvidia, although if I were to just broaden this discussion and I actually do um talk a little bit about Asia. I'm on Bloomberg Asia. It's located in Hong Kong. And so I do think about the Chinese markets, you know, more deeply than maybe just a United States investor. And one thing that comes to mind for me when I think about China is they have some pretty big tech companies. They have Alibaba, Tencent, BU. Um, they've done well on EVs with BYD. They've done well on mobile with Huawei. This is a country that has always sought to be self-reliant on technology. Uh, I don't know why they wouldn't want to be self-reliant on a market that is larger than e-commerce, larger than social media, larger than EVs, and larger than mobile. Uh, I have about 3 to 4x larger market than mobile for AI. So I I don't know a path forward for China given their history that would indicate they're going to be dependent on United States silicon. >> Yeah. And certainly from a from a political geopolitical point of view that does not seem to be in the game plan at all. So uh it sounds like you're not you're not like counting on that to come back online. You're you're looking at that as a homegrown story and not something that's necessary for Nvidia or some of these other international chip players. I would say the supply chains are going to are uh onshoring supply chains is a significant undertaking. Um it would be that the manufacturing side that could be impacted more than Nvidia's line item on their sales onshoring that will take maybe a decade and you know entirely and how the global tensions approach that piece is probably the biggest question mark for the semiconductor industry. Mhm. >> Um I see that as se a separate issue than did Nvidia get to recognize H20 revenue from China this quarter. It's more of this bigger discussion of like how do you onore these deeply entrenched supply chains that are global uh and very reliant on China. As part of our silver interview series this month, we're bringing you top tier research in a free silver investment report. To claim your free silver investing report, simply click on the link in the description below. And if you're looking for a simple, secure way to invest in physical gold and silver, check out Hardass Assets Alliance at hardassetsalliance.com. Speaking of silver, Wealthon will be on the ground at the SCP Resource Finance Global Silver Conference this October in Toronto where Eric Sprat will deliver the keynote address. It's going to be a major event for silver investors. So stay tuned for more details in the weeks ahead. >> Yeah, absolutely. And this and this is where why sort of macro matters more now right and certainly geopolitics because of because of the sort of just scope of that challenge. Uh so the other issue I think that that is really front and foremost for investors with Nvidia is valuation right and so how are you balancing and nobody knows the sort of the the details of of I think this this sector and the Nvidia story better than you. So how are you balancing the fundamentals versus this valuation question? And I'll I'll share an anecdote. We we just had Jim Bianco on and he was talking about um you know just broadly the equity market and some valuations and he was at a a cocktail party and uh someone they they were all talking about how they own Nvidia and he just asked them like oh what do they do? They're like I don't know but it's a five billion it's up 500% this year and he's like yeah but what do they do? and they had no idea nor did they care you know so it's just like sort of retail chasing the momentum of a really really high-erforming stock and I think that makes people worried about some of the valuations around Nvidia so how do you balance those two things >> yes well in brief uh Nvidia is no longer a chip company it's a rackcale systems company um that means that it's a little bit of software in those systems there's networking it is how GPUs communicate to other GPUs, that is why this stock will remain investable for the foreseeable future. Um, and the more GPUs you can combine that think like one GPU, the higher up you can go in training models. Um, now you can break through from let's say a couple hundred billion parameter model to a one trillion parameter model. AI models can get smarter and smarter. um they are the parallel processing um king and that means that they introduce something called tensor cores that allows things like matrix multiplications um floatingoint precision these little features or not features these massive features transformer engine they have created a lead around training that is insurmountable >> uh and the reason that they were able to do all of that is they were so close to parallel processing from day one with gaming and then they just built out more parallel processing power for AI. And uh now when you look at what's different about Nvidia and how the AI economy will be built out versus something like mobile or maybe very similar to mobile actually maybe I can make an analogy here if I think about it. Uh I believe those that are that deeply entrenched in the hardware stack like Nvidia will also be dominant players in software. Um you can think of this as Apple being very dominant with the iPhone but also dominant with iOS and then having an app store something similar to that. I I don't want to pull a perfect analogy other than uh the fact that there's more to this story than just uh you know a single chip. And when I think about valuation, that's just like a broader thing. Like NVIDIA is not a chip company. Um, I just want to throw that out there. It's an AI systems company. Uh, it's powering all of AI training because of those significant advantages on product. Uh, specifically parallel processing power and they have a leg up, a first mover advantage when it comes to AI software, which will be as big as hardware. So, uh boy, like how could I how could I >> underscore exclamation point exclamation point like yeah this this is so and that's a huge statement. So and I think it's really important because I just sort of shared an anecdote about how many of us do not really have a a complete understanding of what's going on um underneath the hood which is why we're so happy to have you here today. And so that that that seems like a very different beast when you're talking about the ability to sort of scale and grow with this monster revolution, which is what I'm hearing you say. But so so again, it's this sounds like this is a stock that we all should have in our portfolio if they're going to be that critical to this to this big revolution. But how do we think about that in terms of valuation? Because I think that's what has people nervous. Okay, maybe I own it. Do I add to it? Maybe I never got in on it because I was burned by the dot bust and I was too scared. Like, do I do I start to buy it here or do I wait for pullbacks? Like, how do you guys think about that strategy management verse of of the sort of valuation versus that incredible fundamental story you just laid out? >> Yeah, that is the most important question is that we have the world's most valuable company here. Can it continue to grow? And I can go through that for you. I do want to say that our process is extremely detailed. A stock like Nvidia, we go through over 100 point checklist at least eight times a year before earnings and after earnings. I probably write maybe 25,000 to 40,000 words on that stock every year. Um, you know, we're talking very very big deep dives that constantly question, should this be a top position of ours? What's the allocation? And the way I answer that for my team, for my research members, and when I can, you know, on conversations like this is is here we go. When we look at the data center today, it's about 48 billion expected next quarter. Let's just call it 50 billion per quarter uh round up. So that's a 200 billion run rate. The reason why this stock has room is because analyst estimates are too low. when you try to do financial modeling on this uh opportunity Nvidia it's it it just crashes the models and once we hit that 50 billion data center quarter which I had predicted six quarters ago we would glad to know we're there that's a little mile marker for us I believe by the end of next uh calendar year we'll be at 75 billion that creates room in the valuation because anal estimates are pretty low. Now, let's broaden the discussion that Blackwell and Blackwell Ultra, the next two generation of GPUs should bring us to a 300 billion data center, up from about a 200 billion. Now, where the market and or analysts are getting it wrong is they don't see Nvidia hitting 500 billion until like 2034. That's where the e that's where the most room is in my opinion. I believe Nvidia will hit a 500 billion uh data center by 2028. And the reason and it was really interesting because I had just talked about this and Jensen Wong echoed it and I love when like you know I'm so deep into a stock that the CEO is putting out something around the same time. He was saying you can expect about a 50% kagger. That's what will bring us to 300 billion in six quarters. So now the math is just there is what I'm trying to explain. the math is there >> and analysts don't right now can't model that uh because they would have to rely on so many assumptions. If we get to a 500 billion data center segment by 2028 we're at roughly 200 billion now you know that would imply depending on when you buy it you know 100% roomish potentially more in the stock. Uh now if you take additional comments from the CEO this uh this this market could be up to 3 to 4 trillion by 2030. What's Nvidia's percentage? Probably a trillion dollars. They're about 33% of the market somewhere in there. Uh now here's the thing and this is where I can help people reframe. This isn't actually Nvidia doing this. It's big tech that is doing this. They are spending. they are opening their wallets and they are saying I don't want to be left behind and for example Blackwell if you don't get the Blackwell system let's say you're one of the the three bigger cloud infrastructure companies or maybe you're something like um you know XAI if you don't get the NVL72s you're going to be left behind because they are combining 72 GPUs versus eight what you can do is significantly better four times better training 30 times better inference at they're going to open up their wallets to make sure they get those systems. And it's really about big tech. That's where I would remain focused. And when big tech starts to cool off, we still have enterprises that aren't able to even get these systems. Um, so to put it into perspective, capex just grew 23% quarter-over-arter. Um, so I would think of it as truly the buyers driving this forward and not put so much pressure on how could Nvidia be that great? Could Nvidia really XYZ? it it's it's it's a marketplace where the buyers are scrambling to get the critical advantage of this next generation of GPUs and they'll do it again when Reuben comes out. Uh so I would think of it more broadly focus on the demand side. Is there anything that you see in the future that would cause that to cool off or change >> because Nvidia for sure we actually saw this last year. I I don't know how much the news picks up on all of this in granular detail. We actually saw what Nvidia's bottleneck is, which is that the product roadmap is moving so fast. This company is so aggressive in what they really want to achieve that they can run into this last year was thermal and power management issues. The Blackwell generation was supposed to have been here already. We're looking at roughly a six-month delay and they resolved those issues and they're moving forward. Now I would just say uh by breaking the upper limits of what is possible and having so much pressure on you as a company that engineering side doesn't have the time to fix everything that could go wrong. So that would be one. Mhm. >> Secondly, uh where we are starting to position is energy because if you track this by GPU generation, which I kind of, you know, started with in our conversation is maybe Wall Street is forced to track this company by quarterly earnings. If you can track it by GPU generation, I think you'll be a much more successful investor because it's each one is unique in their own um you know what they're achieving. And what is unique about uh Blackwell and Blackwell Ultra? I would say if I had to choose one thing, it'd be networking. And then if I had to choose what's unique as we continue to move down further generations, I'm going to go energy on that. Um, can we when when will the grid be maxed out? >> U, what generation of GPUs will do that? My best guess is it could be Reuben. Um, just because we're looking about between Reuben and Ruben Ultra from where we are today, let's, you know, 3 to 5x more power is required to power those systems. How are you going to get the power? Can you get it fast enough? Um, the transmission of the power, is it close enough? A lot of questions there and that would pro probably be the bottleneck would be just the size of this market. How do you deliver these systems that are um I I call it like Nvidia 3.0 uh if Nvidia 1.0 was gaming and then Nvidia 2.0 is tensor cores with you know ampear hopper you know this is Nvidia 3.0 now which is these rack level systems. Uh just continuing to break the upper limits of how many GPUs can you put together and be successful at it uh without breaking something. And then uh from there it's and and having to work through that challenge. And then from there it's um how do you power it? >> It's so you're you're hitting on a topic that's near and dear to my heart. I um a few months ago did an energy conference where all we talked about a Reuters conference central to that in every conversation was the relationship between energy and hyperscalers and how people are going to try to solve for this enormous problem and it's just wild some of the some of the things that are going on there. So, it's so interesting that you're focused and and zeroing in on that as um you know, listen, it's a hurdle, but it's also an opportunity, right? Um if you're if you're watching the right players, um I'll ask you about that more in a second, but just just to sort of talk about this this sort of AI spend and and what might slow it down. Um are there does everybody win in this? I don't I'm I'm trying to think of a really simple way to say it because I know I'm I'm thinking about people listening to this conversation and I keep coming back to the sort of valuation side and it's not just that but there are people and there are plenty of people littered all over any kind of you know financial media or finit or all over the place who will make an argument that this is just like an enormous AI bubble reminiscent of dot not the same right doesn't repeat but it rhymes and that while this might be a disrup technology. There's a lot of hype and there are going to be a lot of losers in this and there's a lot of money chasing just anybody who's throwing AI in the conversation and I know that you understand this space so well. You know, does anything feel bubbly here? You know, for people who don't don't subscribe to you or don't understand this space as well as you do, but they're trying to think about exposure to tech. Is there any part of this universe that feels like a bubble? >> Yes, absolutely. I I do agree with that. I would say um like AI software is really in this R&D stage and there are there's a lot of potential. Uh but until we start to see real revenue growth and AI numbers uh discussed in these earnings calls, that's where I think the bubble is primarily is at the software layer. uh where the picks and shovels like Nvidia, Nvidia suppliers, which there are many Nvidia suppliers, they they really are in uh that that they really are down uh downwind from the massive capex spend where others are just trying to figure out where they fit in. And distinguishing that is key. It's always been key. And I really try hard to not move into a position unless they are participating um participating right now. >> And so so two things are going on. There's there's so much going on with this AI market. One is >> the quants and those looking for a quick trade will rush into a stock um that will pop the stock price. Those who aren't looking deeply at the fundamentals will say, "This must be an AI stock. It's up 30% today." And they'll run into rush into it. That's not our process, of course. We are, like I said, over 100 point checklist. We are going really, really deep here. And uh off the top of my head, I can usually rattle off those with not only the most AI revenue, but AI growth because that's something that really helps inform our portfolio positioning. And so I think that piece can get really confusing. >> I would say it's not unique to AI. We saw that with cloud. Um I remember gaming, there was a moment the gaming industry, everyone thought it was going to last forever. It didn't. Um, so people will talk about the dot because it really was a public market uh driven um you know tech trend. There's almost every tech trend 90% fail. That's just tech. And whether you were in the private markets and you see it or if you're in the public markets and you saw limited trends like which would be internet, mobile, cloud, uh it's happening all the time I would say. So nothing unusual there. uh except knowing in or having that information as an investor would mean that your process should really be I need to find the 10% and that mindset >> of I need to find the 10%. Now, when it comes to is there actual is there actually going to be applications and you know is is there going to be what I would call product market fit. Is there going to be a moment where AI is making a meaningful impact besides just a chatbot? Um where that impact will be and why this is so attractive to companies like big tech and different from the other trends we've seen on the public market. So again, that's internet and mobile is that this is not a consumer technology first and foremost. It's an enterprise technology. A company like Meta can spend I I think they're at a I think they're spending almost 25 billion a quarter right now. Um it is mind-blowing. uh they can do that because they're going to cut back on developers internally as well and they can optimize internal processes and then they can optimize their own product which is their their ad engine and by the time that you do all of that you are not looking to get consumers excited about AI uh that's not first and foremost you can instead just make your own product products better and you can work hopefully get developers to build foundation models with Llama for Meta's example >> and you can cut back on your developers and your headcount. So, by the time that you do all that, they're they're not trying to come out with like a, you know, a social media um app feature. They they do have an AI app that they're trying to incorporate, but they'll even tell you we're not expecting any revenue from that this year or next. no meaningful revenue but they will at the same call earnings call say but we do think that by the time of next year AI will be able to replace a mid-level engineer so they're they're telling you the our first move is to replace mid-level engineers and we don't know when the consumers will will adopt our AI app that so it's does that solve the moat question because it like I hear that come up all the time well they're spending all this money they're not going to be able to build a moat where they can get all the revenue from it Sounds like you're describing they're spending all this money to really transform themselves into this super company um efficient um better better able to operate um and so that maybe the moat question isn't as relevant or does it matter? >> They are definitely building their fortresses. I guess that's that that's a big deal for them for big tech. They're all about the fortress and what AI can do is just make their ad engine for example meta. I'm just using meta. similar for all of them. So superior to the mid-level players um that they continue to take, you know, some kind of market share. So in this case, let's just use ads. uh the targeting is is so much better and the visibility around the conversion is so much better with AI that this company that you might have said is going to plateau which you could argue it plateaued >> uh somewhere about you know um 2018 2020 uh if you look at some of their key metrics average revenue per user daily you know that kind of thing it was was plateauing a little bit there they could now um continue to grow because their ad targeting is just so much sharper with AI. Um that piece of continuing to grow when most of them, let's just assume had plateaued. We know Azure would have plateaued because when they break out the AI revenue, if you were to remove the AI revenue, let's just call it Azure core that was plateauing and not keeping up growth rates. Um so they get to ex start growing again as well. In addition to building more of a fortress, more of a moat, I would say though that uh I guess while we're on this topic, I would say that AI is actually a threat to their moat. Uh because not necessarily on their core products, but they are pretty concerned in my opinion as to if they can become the foundation model that developers use to build their applications. That that piece is front and center. OpenAI's done a great job there uh in terms of their API usage and that piece can really tip the scales in terms of moat uh if you just have so many developers using you know a specific R&D company's APIs um that that could threaten a moat. So I would also say it's a little bit of both. It's offense but definitely defense as well. Hm. So, as you look at this, it still it sounds like you're still believe hardware is is sort of the critical space here. Um, if if if someone has new money to put to work, um, are there other parts of the ecosystem that look poised to be interesting or where you see opportunity or some like smaller names that are not on people's radar? Where are you looking for like the next opportunity? Yeah, if I were to like start to talk about how how am I positioned and there's a few ways that I'm positioned. I suppose if I think about the portfolio, the number one way is AI networking and it goes back to what we first discussed with Nvidia GPU uh G you know Nvidia moving from a chip company to a rack scale um company which J basically just means that they're moving from eight GPUs communicating to 72 GPUs communicating. And in a nutshell, that's I I really try to keep things simple because tech you can get in the weeds with technical jargon, but to keep things really simple, uh that's a networking problem to solve. Uh among many things, but networking for sure. Uh to put it maybe in some rough numbers, you're looking depending on the component within the networking um stack, you're looking at 5x to 9x more of you know whatever those components are, whether it's rettimers, fabric switches, whatever that could be. Um and the reason why that needs so much more comm um more networking components and just even cabling is because the GPUs have to communicate to each other. there's just a lot more GPUs now to communicate. Uh now, how do you do that without signal loss? Um there's just lots of um there's lots of questions to to to solve >> in terms of how do you how do you move from that system to that and we know it was a complex problem to solve because Nvidia had a six almost ninemonth delay and had to really double down and figure out how to release a system that didn't overheat. Anyways, with that said, uh AI networking is where I would say we have outsized allocation right now. As we move into the second half of this year and into the foreseeable future, I don't know when we would change that piece other than now is a great moment to look at AI networking because if you go back to that comment that the best way to look at Nvidia and even AI is GPU generation and GPU generation for now. Um that is what is um front and center in terms of the biggest opportunity in my mind. >> Yeah. Um I'm I'm I'm sure everyone's reckracking this and taking notes because I think we all need to reset the way we we approach this. Um so I'm so happy that you're giving us some of this language to really sort of educate ourselves and dig in. So are you looking at since you see energy as a big issue here? Are you are you also looking at that space from an investment point of view or are you just looking at it from the sort of fundamentals? I need to watch what's happening here. >> Yes. Where we started to move into AI networking almost a year ago and uh meaning early. We try to be you know reasonably early here. That is the way you you know what real wealth can be made. I we're in the early stages of moving into energy which means we've dabbled with a couple of stocks. Some have actually been quite successful because these big tech companies are gobbling up whatever um alternate power alternative power they can find. >> Um and then you know just going back to what are we talking about here? >> Um I would model at least 10x growth for the right energy stocks between let's just say Blackwell and 2030. Um, Blackwell is about 100 kilowatts more or less depending on how you build the system. It's breaking the it's breaking through 100 kilowatts could be 140 kilowatts. We will end the decade with one megawatt. So there's about 10x growth in there and just energy requirements on GPUs. So you could, you know, reasonably say it should be about a 10x move. Uh so if if an energy stock is chosen correctly, I think it could have a 10x move in it. And it's really that urgency that's kind of interesting for an investor, which is those systems I was describing. They as that are shipping now, they cost about three million. Um you know, if you're buying multiple racks, you're into the billions. Uh potentially every week you're paying for a couple billion in racks. um or you know every week Nvidia right now is shipping three billion in racks. So, but let's say every month you're you're spending that and it just continually accumulates th eventually those would have to sit on the shelf and or sit stagnant. These extremely expensive systems if you don't have power, if you're not supplying power um so big tech is a big tech and I would call the neoclouds, it's like your core, they're going to want to get there early. They're not going to wait until it's too late and they now can't power these GPU these GPU systems. they're going to want to think, you know, a year or two out as well. And then if you think about Reuben, which should be about 3 to 5x from where we're at now, I think if I had to guess, uh, that's where we run into serious problems. Therefore, any, you know, accumulating of the power right now is probably preparing for Reuben if I were to guess because that's about one to two years out >> which is so interesting because right now if you just look, you know, broadly, the energy sector is pretty beat down and investors are kind of down on it. It's out of favor. So, you know, it's is why it pays to kind of keep your eye um really fixed on the future. Um I want to hit while we have time just another area speaking of energy demand. Um another area that I know is on your radar and you watch is uh which is crypto. Um how you feeling about Bitcoin at these levels is your interest elsewhere? How are you thinking about that space? >> Bitcoin has uh Bitcoin and crypto has a completely different process at our company. uh we feel quite strongly that it's sentiment driven and the difference between something that's sentiment driven and I'm just talking crypto broadly of course Bitcoin being the heavyweight there is there's no earnings reports there's no financials you can definitely pull some key metrics uh and try to compare them and people do but there's no real moment that you can say this intrinsic value is changed in a uh standard way. Therefore, um technicals work really well, we've found in in terms of really riskmanaging crypto and we firmly believe in riskmanaging crypto. Uh because even your juggernaut of Bitcoin, it can go down 60%. Will it go down 60% this next time with ETF participation? Maybe not. But there's a lot of volatility in crypto and people forget that. And therefore, we are big advocates of setting a price target years in advance and literally years in advance. Uh and when it hits that target, you go ahead and you trim with the goal of buying lower. Uh we believe a uh near-term top could be seen in the 130s. Uh that's a moment where our firm who's up about you know 600% some somewhere in there we're going to take some profits profit taking and we will put it back in to work lower now this is extensive work between an AI engineer um named Vincent Duchain who was early to AI I mean he's he's the OG AI kind of type where he was selling like companies to terodine uh before Nvidia's big moment he's OG kind of AI, not AI is here. I'm just going to jump on the bandwagon, which a lot of software developers have done. He's he's like OG. And then Knox, who's constantly working with technicals. Those two have combined forces to really figure out how do we risk manage Bitcoin? There's we have too much money in it to ignore that. and they're coming up with uh the strategy that profit taking in the 130s is not a bad idea with the uh with a with a minor pullback on the horizon. Load back up, go into the 200s. What's interesting is both of their models are saying right around in the 200s, this thing should put in a meaningful top, >> which I find interesting because you'll hear people say 1 million or People go big on those price targets. >> Yeah, they certainly do. Are you are you um is that in your portfolio because you were early and like you said you just have too much money in it and you have to manage it or or do you see this space as opportunity because there's a lot of conversations about ETH now that we have institutional participation you know people really trying to look at these these ecosystems which is really what they are to see functionality and what might you know you get you have a lot of enthusiasm and people building earnestly in that space. So do you see that more broadly or are you mostly positioning managing that like early right call for for Bitcoin? >> No, we participate in in additional coins. I would say our crypto mix to stocks is usually 70 stocks, 70% stocks, 30% crypto. Uh we've always been actually quite aggressive in that way. This is before there were ETFs. The minute we launched our premium site, we recommended Bitcoin and told pretty conservative investors, you you have to put Bitcoin in your portfolio. Uh, and that was so and Warren Buffett was calling it rat poison, that kind of thing. uh I feel there is enormous opportunity in crypto where we are a little bit different is we think it's prudent to book profits particularly in a volatile asset class like like crypto. I would say the same for uh some of the hyperrowth networking stocks. we are not um we are not interested in holding the bag. And if you can put that first when you're a tech investor, it seems it seems in the reverse, right? It's like I got to make money. I got to make money. That's how what I'm doing as an investor. It's it's actually smarter if you could say I want to protect my money. I want to protect my money. Uh it's just a little bit of a different mentality. and we've put it to the test um where we said can an actively managed portfolio keep up with a long-term buy and hold plan and we have exceeded those portfolios in terms of performance that go long-term buy and hold because like you're talking about now we have a lot of stretched AI valuations what's the plan >> are you going to continue to hold are you going to book profits are you just going to take that hit uh you know those are questions that I think are as important if not more important than the stock picking side and and and especially with crypto. >> Absolutely. I mean, you can't say it any better and I think this is a super important time for for people to plug in on this. Beth, we I love your level of rigor on all of this. It's so fantastic to be able to talk to somebody who's been following space and so smart um in how to look on it and I think uh look at it rather and I think we can all learn a lot from that. So, thank you so much. Always such a pleasure to catch up with you. >> Thank you so much, Maggie. really appreciate the conversation today >> and like I said everyone's I know going to be rewinding and taking a lot of notes on this and when you do if you need or want to take a fresh look at your allocation to tech and now that you have some really good questions to go and talk to an adviser about just a reminder you can get a free portfolio review with an adviser in the wealthy network just head over to wealthon.comfree. Thanks everybody for watching. We'll see you next time. [Music]