Welcome to Okay Computer. Now you're saying to yourself, "Wait a second. Either that's Guy Adami's voice that I'm hearing or it's Guy Adami that I'm seeing." That is in fact the case. Dan Nathan is off this week. So filling in the large shoes of Dan Nathan will be myself. Fortunately, Gene Munster agreed to do this, much to his dismay, I'm sure. But Gene, of course, is a managing partner and co-founder at Deepwater Asset Management. Somebody that I've known for a very long time and somebody that I deeply admire. Gene, how are you?
Honored. This is super fun to join you, Guy. I always love joining Dan and just like this, like extra drive mode. Hyperdrive. Hyper mode.
We're in hyperdrive mode. And that's the way we'll handle it. And obviously, listen, we're talking, Gene and I, on Monday afternoon. This is November 9th. As you're listening to this, it's Tuesday, but we're obviously talking in the wake of what's been a fascinating weekend in terms of the news flow. So I want to bury the lead. Let's start with Nvidia and the news about Jensen and not the, you know, not all the nasty stuff because that's for the tabloids, but what do you think it all means? Because in terms of the stock, obviously the market is sort of taking this as a negative. In terms of the broader market, it says nothing ever happened.
Exactly. If anything, in the broader market it might have been endorsement that we still have room to go on the AI trade. In terms of the actual Nvidia news, kind of the way I've been thinking about it, as the stock as we were speaking here is down about 5%. I would generally say one way to think about like what's priced in right now is a question of if in fact Jensen had to leave, because that's the central question now. If he was asked to step down, what would happen to the stock? And I've surveyed some of the other partners at Deepwater and our general sense is probably around down around 20% if that would happen. It's a pretty big haircut, pretty aggressive, but just given where the stock has been over the past five years, I think you can build a case that you could have a big drop. I think that means that essentially the stock right now is factoring in maybe a 25% chance that he's not going to be around. And maybe there's other small pieces around what it means for the China potential. Maybe that gets diminished a little bit, but I don't think investors have owned Nvidia for the past couple years based on any China opportunity. And so I think that the market's kind of assuming that this is about a 25% risk. I think the risk is actually lower. You know, obviously I don't have any knowledge about what the investigation looks like, but it's probably more like a 10 or 15% chance that he's not going to be leading. And so my sense is that, you know, this is the market kind of acting responsibly. And ultimately the bigger question that we debated this morning was how frequent is it that China kind of puts these efforts against a company? If you believe that this is common, that let's say the top 50 most influential companies in America, that it's common that China would use these tactics to try to infiltrate them, then you really can't read much and beyond this is just standard operating procedure. But if you believe that they only do it to Nvidia and to Google and kind of these forward tech companies, which that's the camp that I'm in, I think it really speaks to what's at play here and just how significant the AI infrastructure whole conversation is relative to the future of geopolitics.
No, it's fascinating. And I mean, I watched enough episodes of Lioness to understand exactly what's going on. If you haven't seen it, I encourage you to do so. It's a wonderful show. But the, I guess, the real question is, to your point, what are the chances of him leaving? According to you, I think your numbers are spot on. It's sort of somewhere between, I guess, 25 and 30%, whether that means he does it of his own volition or steps down or if he moves to another role. I mean, the key man risk is obviously significant, but the levels that you talked about in terms of what the haircut would be if in fact that were to happen. Think about this for a second. It probably gets us back to the prior all-time high that we made this time in 2024, which I think was about a buck 45 or so, somewhere between 145 and 150. So it's not out of the realm of possibility. And by the way, as you know, I mean, this is a stock that had not been trading well leading into this. I mean, the stock made an all-time high, I think, in early October of around 243, and we've been sort of trending lower since. So maybe there was something out there in the first place. I don't know. But let's talk about this because I look at it two ways and I follow you obviously on Twitter. You put out a tweet: the fact that China targeted Jensen underscores the geopolitical significance of AI leadership, 100%. But now this is like, I think, really the important part of the tweet. That means both Democrats and Republicans will need to build lots of data centers here in the United States. That to me is really the, I think, the important thing, and to your earlier point, that's maybe why the market is trading the way it is.
Yeah, I think so. You know, the move we're seeing is, I think, in part with a little bit of recovery from last week, but also in part if you look at some of these like really AI-first companies like Coherent, for example, that's up 17% today. It was up big on Friday. Like some pretty wicked high moves, and they're basically an optical supplier for the data centers. So I think that that really does speak to that this probably is an endorsement at the end of the day. Is that there's the, I mean, this is, there's politics that are going on, but I'm not talking about what's going on with Nvidia. I'm talking about what's going on around data centers. There's an environment and there's also the reality of how the US needs to prepare itself for the future. And I firmly believe that the Democrats are going to want to build a lot of data centers. And is it going to be the same number as Republicans? I actually think it's going to be more and or be at least what the Republicans thought. Now, I'll put this out there. This is going to be a hard one to measure. But the real question here when it comes to the data center buildout and the capex is on the 2028 number. So we're seeing about roughly 80% growth this year in capex and the hyperscalers as a proxy to the data center growth next year supposed to be right around high 30s. It was around 42 before the election. So that's come back a little bit. But for 2028, Street's looking for 15%. And so I think that we'll be well above 15% for next year. I suspect that, excuse me, for 2028, I suspect that number is probably closer to 30 or 40%, much higher, which is either a clear view in terms of how important everything is going on, the need for inference, how much demand is out there, or it is someone who's delusional about what is going on in terms of the significance of AI. It's going to be one of those two. Guy, let me throw this at you. It's something that I've read. I actually sort of thought about it earlier. This news, and I'm just putting it out there. I'm not suggesting it's going to happen, but this new move, this news potentially, Gene, could give companies an out on the capex side of things. In other words, if you've been, you know, pouring your money into Nvidia in a meaningful way, and if Nvidia is at the epicenter of this, if companies have been looking for an out, and again, I'm not suggesting they are, but if they're looking for an out in the form of we're not seeing the return on invested capital and we need an elegant way to get out of it, this is a pretty elegant way to do so. Am I reading too much into this?
I mean, if they are being forced to spend, it is absolutely a way out. There is this prisoner's dilemma factor that's going on between all the different hyperscalers, or this nuclear race, arms race, whatever the analogy. But ultimately is that I think that at the end of the day, I mean, this could be, yes, that's true, this could be flagged as a reason we don't support Nvidia for these reasons. It would beg the question, do you go for GPUs from AMD then potentially? But I think that the reason why the hyperscalers are so bulled up, I mean, we heard it on the earnings call. There was a disconnect between what we heard on the earnings call and how the market reacted. The market was not buying what the CEOs were selling. And I believe the CEOs when they say how early we are, when Zuckerberg talks about the potential around Meta longer term, when Elon talks about what potential they want to do in terms of how much capacity to build. So yes, it is an out, but I don't think it will be an out. I think they fundamentally want to build these data centers.
Look, you know this a lot better than I do. I'm not going to pretend. But when you hear Facebook for the first time, I think in 10 years went cash flow negative, and that's on top of all these other companies that historically have been cash rich, they're pushing all their chips in the middle of the table. And to your point, probably rightly so given the importance and where we are in terms of the cycle. So that's out there. What I found fascinating, and I'd love for you to opine, post-election when there was some concern around data centers and stuff, all those stocks got a huge bid on the fact that maybe, again, maybe there's an out for them. So I don't know how to read this other than trying to sort of game it out in terms of stock volatility, and it's been in a word, as I used before, fierce.
The way I've been thinking about it is there's the market today, kind of the tone of the market, which is it's nervous. Market's nervous, and I think when you talk about slower capex, when the market's nervous, that's viewed as a near-term positive. There's also the question about like the fundamentals, like what is the reality of growth going to be over the next two to five years, and will the market pay, what will the market pay for that growth? So these three things and what we see in this week and these immediate reactions, whether it's Nvidia today or what with elections last week, is the market is operating on its current mode, and again, that mood is nervous. We saw going into it. Maybe some of that was priced in. Maybe the market was ahead of what the oddsmakers were in terms of how these elections were going to turn out. But either way, I think it speaks to kind of the nervousness, some of the euphoria, but then ultimately nervousness. And historically, when I say historically, I mean kind of over the past 40 years, is that fundamentals should move the market. And of course, these stocks have all moved up in anticipation of great fundamentals. And so that begs the question, will the substance of demand exceed the hype? And that's where our conversation typically comes down to. I think the market should be nervous about cutting capex. And again, I'm bullish. We use these tools. We're not just observers. We're practitioners. And I think that it should be rewarded. These companies should be rewarded for the spend. I don't know how the market would react if we were to hear a pullback in capex from a Google or Facebook or you name the hyperscaler. And see, but you know, let's just play stock market a little bit to the extent that you're interested in doing it. Google, for example, I think they made their all-time high in the spring of this year. We recently traded down to the lows that we saw in late March. So the move lower has been over the course of a period of time and then dramatically a couple weeks ago. So for whatever reason for that, I can't fully understand other than the slowdown in their, the concern around their search business, Google's being punished. Is that slowdown around search, which by the way, people were talking about the existential risks associated with this a year ago, were they just early or is that real?
This is so dynamic what's going on in the market. A year ago, we understood what's going on with chatbots and the impact potential. Sorry, a year ago, this is early 2025, and how Google search, the 10 blue links, could be negatively impacted by that. I was negative on Google at the time that I saw the March search results, realizing that they could take these AI overviews, AI mode, essentially funnel traffic from search into AI activity and draw greater monetization. That is good. They were in a good position. They were in a great position. But the challenge is they missed on the innovation side. And of course, what we've seen since the beginning of September, the rise of the intelligent bots, the personalized agents, call it what you like, basically changes the whole conversation and changes the conversation because how we use bots is different than how we use, or how we use personalized agents is different than how we use a chatbot. And so to answer your question is I think Google essentially did an incredible job of parlaying what was their 3 billion search users into AI traffic, but now has a new question about how do they compete in the personalized agents world, and I think that adds a whole new wrinkle. I think it's a fundamentally different way that people will be interacting. The usage, and we've talked about this on the show, very small today. We're probably at 25 million daily active users, 5.5 billion daily search or internet users. So this is the new reality, and we don't know what Google's response is going to be if in fact people start their day with their personalized agent.
Let me ask you this, because again, I don't fully, I don't understand it nearly as well as you do, but I've heard Jensen, you know, again, the pre-week news, Jensen talk about this technology that we talk about all day long being as important as electricity, which, you know what, that's probably true. I think you would agree with that. But if you think about what electricity is and what's happened over the years, electricity is a utility, it's become commoditized. Is that a concern or is that a good thing or a bad, a good witch or a bad witch when you hear that?
I think both can exist. I think you can have commoditization for tokens that are at the lowest price. So the average intelligence of an AI agent today is about 130 IQ. Those are higher price tokens. Those cost about 10x what a low-level token would be at around normal human IQ, about 100 IQ. And so the question is how much of those, I believe that there's going to be a lot of commoditization with token pricing. But I also believe that some of the biggest problems that take the most amount of usage is going to continue to increase, and I think that the need for these most advanced tokens will be there, these most advanced models, and I think people, businesses will be willing to pay for that. And so ultimately, I think that you can have commoditization as long as there are frontier models that are still desired. And I think that's the case. I would put it this way: 90% of the world's going to trade on commoditized tokens, 10% will be on the frontier ones, and I think those will be more than enough to offset the pricing.
History can repeat itself through this lens, to your point, where you know you have a commodity, you have 90% of something being a utility, commoditized, but the other 10% more than making up for it. So we'll see, but I think it is worthy of a conversation. I think over the last couple weeks, last week or so specifically, people are making a correlation between what happened in 2025 with DeepSeek and that selloff that we saw, where I think the NASDAQ was off about 20% or so in a month and some of the underlying stocks more than that, and what we've seen now, to the extent that you can compare, is it even worth comparing? Because I know you have some thoughts on that.
So DeepSeek at the time was in front of unknown about what these models, ultimately the pricing of these models, kind of that same conversation was also in front of the tariff concerns. So I think those were actually much more negative of environment, as you said, NASDAQ down 20, the NASDAQ was down 14, the most recent sell-off, so it didn't get that further down because we didn't have the uncertainty around tariffs. We didn't have the uncertainty around these models. We do have uncertainty around the pace of the AI buildout. So again, the market's actually responding appropriately. I think when you look at what the pullback was around DeepSeek versus the more recent pullback, it just all seems to line up. In both cases, we had the same outcome, which was the substance did not exceed the concern. And I think that's going to come down to, that's the fundamental question that people need to ask themselves today. Investors need to ask themselves: do you think that the number of data centers over the next two or three years is going to slow down? If that's the case, we should, if that is the case, then the market will continue to pull back. I would, just another kind of dynamic around this, I'm not saying that this new governorship, the Democratic governorship, I'm not saying that will have no impact. I think it does have an impact, but the simple math is to say that about 85% of the cost of a data center comes in the hard costs like chips. 15% soft costs are going up. Labor costs are going to go up, provisioning costs, potentially. What that means, subsidies to communities, all that's going to go up. If that 15% piece goes up by 30%, then the overall cost of the data center goes up by 5%. That to me is a trade that these hyperscalers are still going to want to make.
