And I think part of that fear is created by these tech companies themselves in the AI revolution. We're still in the third inning. We're talking about something that's going to change the way we live. There's a fourth industrial revolution, and it's the US that's going to dominate in this tech revolution.
Where do you think that the bears are right?
So this is Unusual Bets. I'm your host Arthur Bayisteros, and every episode I sit down with an expert who has a bold or contrarian bet on the future, and they're staking their business, their money, or their reputation on it. And so my job is kind of to dig in here and explore their idea and what it means for you and where those opportunities might be. And the goal here is to find the bold ideas that aren't really on your radar yet, but they probably should be.
So my guest today is Dan Ives. He's partner and senior managing director at Yorkville Ives. He's been a tech analyst on Wall Street for over 25 years, and he's one of the most sought-after tech experts in the entire finance world. You've seen him on CNBC, Bloomberg, CNN, just about everywhere. And now he's on Unusual Bets. So thank you, Dan, for coming on the show, and I welcome you here to place your bet.
In the AI revolution, we're still in the third inning. Like this is just a massively bullish market that we're so positive on overall tech. But I'd say if you say like bet or the things that maybe like would concern us is really just the politicization of data centers and what we see in terms of playing out, especially going to the midterms. I think that's a risk that needs to be mitigated. But it doesn't look, it doesn't change my overall view that this is less than 15% through the bull market.
Yeah. So when you talk about that 15%, we're about maybe two, three years since ChatGPT became a consumer product here. So you're saying that we still have 85% of this whole buildout to go. That's, that's your bet here.
This is year three of a 10 to 12 year buildout. It's building the Vegas Strip in 1955. Now I'm sure when you built the Vegas Strip, there were people in 1956 being like, oh, that building didn't work out. This whole strip thing is not going to work out. If you showed them the Sphere today, what would they have said back then? My whole point is, is like for the first time in 30 years, the US is ahead of China when it comes to tech because of AI. And I still believe investors are underestimating the scale and scope of what this is going to look like, not just from a market perspective, but how it's going to impact typical consumers and businesses.
I think that you're definitely feeling the nervousness from retail investors because a lot of them went through the tech bubble, they went through the great financial crisis and things like that. For people who are maybe a little bit younger and who didn't go through those things, should they be worried about what history taught us from some of those lessons, or is this just completely a different ball here?
Look, I get the comparison, and obviously we hear it all the time, right? It's for someone like me that covered tech stocks during the bubble and the burst to where we are today. So let's just like compare and contrast. Like back then, average tech company traded 30 times revenues, basically none were profitable. Today, the average tech company's trading at 21 times earnings. Back then it was VC-funded PowerPoint presentations, business models that were all on the come. Today this is being driven by big tech with a trillion dollars in the balance sheet, 300 to 400 billion a year free cash flow. And then it comes down to like the use cases, the why data centers are getting built out, and why GPU, and why Nvidia.
We're talking about something that's going to change the way we live eventually as it goes to physical AI and robotics and productivity. In my view, like more jobs would be created by AI than taken away over the next 10 years. But as you go into it, look, the bears have called 10 of the last two downturns. The problem is is that you have a lot of keyboard warriors, catastrophists. They usually come out on Saturdays typically, um, because bulls are just enjoying family and, you know, touching grass, whatever, the bears are building their conspiracy theory.
The point is, as someone myself has traveled three and a half million air miles, like I just got back from Korea this week, like the demand is unlike anything we've ever seen, and you have three to four trillion of capex that's going to be spent. So that's why I think it's just like it's very important sometimes where narratives could get you off your game as an investor that ultimately creates opportunity.
Where do you think that the bears are right in this situation, though? I mean, when you look at a lot of this and you look at the narratives, as you say, and you look at the circular financing and the issues with CoreWeave bonds, and you look at Oracle's, the fears around Oracle potentially not being able to pay a lot of this out, do you think that they're right when it comes to any of that, or do you think they're right anywhere along with their thesis?
No, I think they're right as will there be some issues if yields continue to go up, debt financing. You look at Oracle being a good example. They have to like cut cost, cut cost, cut basically muscle and bone to be able to then build out the data centers. That's why Oracle is like maybe like the poster child. The view that like neoclouds are not going to exist and there's more supply than demand, that's a false narrative. There's some fears that could be accurate in terms of finding winners from losers, but I think they're misplaced relative to just painting the whole sector with the same brush because it goes back to like let's just go over like the last few years: financing, geopolitical, US-China, like, you know, AI, the use cases aren't there, China with DeepSeek, that's going to take down our model. I mean, I could go SaaS apocalypse. Like there's many good examples, and they basically missed Nasdaq doubling.
I'm just saying like it's now at a time you'll get to a point where like there could be names that get ahead of themselves, but my whole view is like there's a fourth industrial revolution and it's the US that's going to dominate in this tech revolution. And that's why I say I think it's just been very important as an investor to embrace the technology change, understand how you could ultimately invest and be opportunistic rather than being someone like 1905, 1910, be like, dude, this car thing is a joke, horse and buggy is where it go. 1995, dude, that internet, Windows, Microsoft, that's a joke, I got to stick with the typewriter and word processor.
You bring up something interesting with the SaaS apocalypse, and you saw a lot of fear at the beginning of the year, especially when this first started in February, and then you saw another sharp sell-off in April with the release of Mistral there. And I remember listening to you and reading some of your writings early on there, and you were basically saying that everybody was wrong at the time about that. And when you look at it now, the IGV, I think, is up almost 40% since its lows then. So what were people missing at the time when they thought that software was dead because of the advancements in AI?
So to that point, it was like when you talked understanding the technology, it was, do Anthropic, they're going to eat software. What, they're going to get into the enterprise, you're not going to need a Salesforce.com. If you talk to any IT engineer, any CIO, any CTO just during that period, you realize like that was a false narrative. Now, are there software companies that have structural issues, including Salesforce? Of course, Intuit, Adobe, I mean, software, it's a challenge, but the stocks got to a point where they basically were saying that the software sector is structurally broken. And I view it on the opposite, where it's like these companies are going to, they are going to, the use cases will be much more software-driven, which will expand surface area, which will increase the need for cybersecurity. And it just speaks to like you look at so many, from CrowdStrike to Palo Alto, there's so many others. Now again, I'm not saying like we're going to go through these white-knuckle moments, and it is uncomfortable for many investors to navigate through it because a lot of times just the noise is so loud out there. But the reality is like for every dollar spent on capex, there's five multiplied across the rest of tech, and that now will drive the next few years this tech bull market.
Now, I know you've said that, you know, that the software names were going to aggressively catch up to some of these AI players, and I think that you have seen that, and I've read and seen you say that they're going to continue to aggressively catch up over the next year. Do you still believe that in our current climate right now?
Yeah, because Arthur, my view is the value over time. Even if you look like Anthropic, OpenAI, the frontier model, that's like, that's only like the tip. They're focused on building out enterprise and building out use cases, and that's what they're going after. My view is that data is the new oil and gold. Data sovereignty, AI data sovereignty is going to become more and more important. That's why Palantir went from, you know, call it elementary school, junior high type of stock to, you know, 200 despite many of you go through, right? So many like, oh, Palantir, it's a services company, and because they don't understand like Karp is, you're talking like Mount Rushmore of CEOs understand where technology is heading. And it was those investors, if you talk to CIOs that are basically like, this company's able to do something I've never seen that was possible, you don't have to be Stephen Hawking to figure out like that they were going to ultimately, if they cracked data and enterprise, the stock was going to be meteoric, right? And I just think like Salesforce, ServiceNow, Snowflake, I go through just so many others where they have the install base. Some might not be successful, but that will be the opportunity in terms of the use cases.
So where do you think the losers sit right now when it comes to software in particular? Because you look at some of these companies, like you bring up Salesforce, and I think that they're still down for the year, while you see companies like CrowdStrike and Palantir that have made big runs and aggressively caught up, but Salesforce still struggling even though they're adopting this AI and they're changing their business model to where they're charging for the work that the agents are doing as opposed to some of these seat licenses and trying to change that. So do you see them as a potential loser here on the software side, or are there other names that you're looking at that might not catch up?
Like I think like ServiceNow is a clear winner. Like that's one, like if I'm betting on someone, McDermott, like they will get onto the other side of this.
Because I think like if you look at their install base and where they sit in the data stack, I'm not saying they're not going to have to do some acquisitions, but ServiceNow is a good example, like they will be a much bigger company from a revenue and market cap, you know, in the coming years than they are today. Because Salesforce is a good example, one like it's been probably the biggest headwinds it's ever had, but they have installed base. If you're able to monetize successfully with agents and where this could all go, then over time this will be an inflection point in the Salesforce growth trajectory.
But part of like what Salesforce faces is like there's some of its business that's melted like an ice cube, other part that's accelerated. So then like look at IBM, AI piece that Arvind's built, massively growing, services, a lot of the mainframe under pressure. So I think these companies like they don't have all day to get it done, you know, so I think there's a timing of it as well, but it's just my view like the use cases will be driven by the software companies.
Now, what do you think of the situation that a lot of these software companies are faced with when it comes to AI and AI being able to kind of erase some of that moat that they had, where AI can go out and build its own software? And why wouldn't a company kind of abandon that model and those software services and the enterprise deals there and just go pay Anthropic, OpenAI, whoever it may be, to just be able to build their own software to replicate a Salesforce or an Adobe? What's missing?
It's a great question. Part of it is like there's security, there's guardrails, there's giving your data to frontier models that you could argue over time maybe even cannibalizes your business. And then it comes down to like they could replicate pieces of it, but when you think about the stack that a lot of these software companies have built, what could 10 to 15% of it be replicated? Yeah. But the reality is that CIOs are not just going to give the keys to the model players in this world, and that it's an opportunity to challenge for software companies. Look, Anthropic, OpenAI, they're very cognizant of that, which is why they're so aggressively trying to hire enterprise sales reps and engineers because they recognize that tokenization costs will go further down. The value is in the frontier model and open source. No different what Nvidia is doing. Look what Mistral is doing. The value will be on the data and the monetization on a use case perspective.
How do you see it playing out with these companies being able to protect that data? Like how do they protect that moat? Because the data is becoming so plentiful and so spread out, especially with data that a lot of even employees at some of these companies are feeding into AI. Like you don't know what your employees are doing with a lot of these frontier models and what they've already fed it. So how do they still keep that moat with the data?
I'll call it AI data sovereignty. That is the biggest debate. You a lot of people be like the debate be like open source versus frontier models, like that's kind of the Rocky Apollo type of battle. I disagree. I think the battle is more frontier model versus data at companies. They want to keep it in their moat. They don't want to give it out. They give it away. And if you look like Nvidia, Palantir have really been kind of well ahead when it comes to on the sovereign AI front. I think that's the biggest debate. I think that's going to be like the biggest sort of clash as you think about where this goes in terms on the data side.
The key is the sovereign AI for these companies, these software companies to build their own AI. Is that what you're saying? That's the key.
Well, it's really build on-prem, and you could have like some frontier models in there, but it's like for them, and who's going to do that? Software company. The reason Palantir has been a rocket ship is that they are the ones that you bring in to build out your AI strategy, but that data doesn't go to them. It's not a frontier model. They're agnostic about the model.
It's interesting too, also when you bring up the open source, like there's been a huge narrative that that is really going to start to make a real market impact here. Do you see any of these companies relying on that because it becomes just simply cheaper to rely on these open-source models, or is that something that you just don't ever see them doing because of the risks?
I think open source is definitely going to be an avenue. I mean, if you think about like Jensen with some of the ventures and revolution, I mean, they've supported open source, Hugging Face. They don't do the Hugging Face acquisition if they're not like bullish on the open-source model because it comes down to like if iPhones in 2007 when they went on sale were double the price, would you have 1.5 billion iPhones today for mass adoption? Tokenization prices need to go down. You need open source because you can't just have companies like 10 days into a calendar month be like, we've maxed out on our quota, like no more. That's not sustainable relative to where this goes. So that is really I think going to be the big focus going forward for a lot of these companies.
Yeah. Is to integrate on both sides. You're saying to...
100%. You have because you can't just go frontier model 100%.
Because ultimately it's going to be the models over time will become more and more commoditized. The value is in the data and what you do with it, right? But it goes back to from an investment thesis, it's following like the second, third, fourth derivative of this AI revolution. And it's very easy where you talk about SaaS apocalypse. I would say like with CoreWeave and like what we're going to neoclouds now, like that's almost like a SaaS apocalypse like 2.0.
Like it's like in terms like that piping, which is what's happening now, right? If you're able to kind of like see around the corners, like I believe like Nebius and CoreWeave, like are, you know, those like huge opportunities today. I'm not saying all neoclouds, but if you just look at the capacity, that's sort of my view in terms of where this all going to head.
So when you talk about looking around the corner here for a lot of investors, I think when you look at a lot of this right now, retail investors, they look at it and they just hear the narrative that the AI companies are expensive, even though Nvidia is cheaper than six months ago given the...
