The Shaky Economics Behind CoreWeave (ft. Gil Luria)

Welcome back to The Rundown, interview edition. Today, we are talking to Gil Loria, the Managing Director of Technology Research at DA Davidson. Gil has been covering tech for a long time and he has a really interesting perspective on the AI trade. So in today's conversation, we covered a ton of topics, including Corweave's earnings and whether the NeoCloud model can compete with the hyperscalers. We also get into NVIDIA's partnership with Wall Street to finance more than $500 billion of AI infrastructure and whether this AI spending cycle is becoming too circular. Finally, we end the conversation talking about Palantir and why Gil calls it the best story in all of software. This was a fantastic conversation. Gil brought a lot of great insights and I think you guys are going to really enjoy this one. So let's get into it. All right, guys, today we are joined by Gil Loria, the Head of Tech Research at DA Davidson. This is Gil's second time on The Rundown. So Gil, welcome back. Thanks for having me. I'm super excited for today's conversation. There's a lot going on right now in the market, especially with the NeoClouds. We had Corweave report earnings this week and also Nebius. So I want to start with the conversation there. Nebius earnings were solid. Revenues doubled compared to a year ago. Backlog grew to $130 billion. Margins are finally starting to move the right direction. Stock popped 20% post earnings. I was reading your note on it, though, and you still have a neutral rating on the stock. What is keeping you from getting more bullish on Corweave? They still don't make any money. Let's start with that. We're not sure that this category is even a worthwhile category, right? The NeoCloud category is the marginal provider of compute. And we talk about tides, right? When the tide is rising, guess which boats get lifted highest? The smallest boats, the ones that have the most leverage to an incremental buyer of compute. That's Corweave, Nebius, and other NeoClouds. So of course, right now, when there's an unbelievable shortage of AI compute, there's an enormous gap between demand for AI compute and supply of AI compute. These are the companies that are going to benefit the most. That's why they're doing so well right now, when there's such a shortage of compute. The longer question is, are these worthwhile businesses? Will they generate the types of economic returns that can justify their cost of capital? The jury is out on those things. And I believe that right now, the evidence is that they will not, over time, be able to even return the rate of capital that they borrow at, that they raise at. Certainly not above that. I'm so happy you're bringing this up. I'm so happy you're bringing this up because I made the same point when I was talking about Nebius's earnings. We covered it on the show. Sorry, not Nebius's earnings, but Corweave's earnings when we covered it on the show a couple of days ago. And I've been saying the same thing. I'm a little skeptical of the whole neocloud model. Just long-term, I'm not sure how they're going to be able to compete with the hyperscalers, Microsoft, Google, Amazon. Those hyperscalers have access to much cheaper capital. So I don't really know what is the long-term moat that these neoclouds like Nebius and Corweave have when it comes to these hyperscalers. They don't. Let's go back to finance 101, right? What's the job of a business? The job of a business is to generate return on invested capital that's higher than their cost of capital. So if I'm investing in something, if I'm borrowing to invest in something, the return I get needs to be higher than the cost of borrowing. Otherwise, I am not a worthwhile endeavor. I should not exist as a business. It's so vivid that it's probably an even better example than Nebius. Nebius is actually a little bit more of a nuanced story. But let's talk about Corweave. Corweave's cost of borrowing is now 9%. To get capital, Corweave needs to raise it 9%. Their return on capital right now as of last quarter is less than 1%. By the fourth quarter, they're saying their margins are going to expand dramatically by the fourth quarter to what they're saying is mid-team margins. But if you take that and you look at what that means for return, that's still around 4% return. So they're borrowing at 9% to generate a 4% return, right? Your viewers and listeners are pretty sophisticated. I don't think any of them would take a margin loan at 9% to buy treasuries. Investing in Corweave is taking a 9% margin loan to invest it in treasuries, at high-risk treasuries at that. That doesn't make any sense. Now, the argument that Corweave shareholders and Nebius shareholders make is, oh, don't worry, we'll grow into it. And I think it's possible. Again, if the demand for AI compute continues to strip supply for the next five years, maybe. But it's not that these are small businesses, right? These are already over 10 billion. Nebius will be 10 billion next year. Corweave is already at a 10 billion run rate. If you can't generate good returns when you're 10 billion and managing dozens of data centers, why should I believe you that you'll do it when you're bigger than that? And that's the essence of the problem. Now, does it matter right now? No, because these companies are showing huge upside and they're selling. And the price for compute is at the spot market is probably four times what it was at the beginning of the year. And this is a good place to be. But you always have to remember that there's another side to this kind of spike. I love how you broke that down. And actually, I was reading your note on it. You said that Corweave continues to destroy value with every data center they build because, again, it costs them more money to borrow capital and they're not making enough on the return from the data centers. So you broke it down really great there. But I want to switch from Corweave to Nebius because I think you are a little bit more warmer on Nebius. They reported earnings this week and their stock jumped 34% following their earnings report. You have a neutral rating on Nebius. But again, reading the note from Nebius, it seems like you're more warmer on them. I guess, what is Nebius doing that Corweave isn't? They're very savvy operative. So if you're going to bet on somebody in this space, you want to bet on the smartest guy in the room. And that's Arkady, the CEO. This is a guy who built Yandex. To build Russia's Google in Russia, gosh, that's a high degree of difficulty. And then he left Russia because of the Russia-Ukraine stuff and got to delist. And when he relisted, it's his Nebius. It's the same super savvy guy. So yes, it's a tough business, the neocloud business. The burden of proof is on him. But he's doing super clever stuff. Like, you know what? I can provide capacity to my long-term customers. But the price for long-term contracts is actually quite low compared to what I can get in the open market, in the spot market, in the short-term contract market. So let me go around and just gather some bits and pieces of supply, sell it in the short-term market, and I can make a killing doing that. This is the kind of savvy operator that you want to be in business with. Now, again, I'm not sure that the business is long-term a winner. But if you're going to be in this business, be with the person who knows what they're doing. That's a great point. And I'm happy you kind of explained that, the difference between the two. I want to kind of move on a little bit and talk about the role that NVIDIA plays in all of this. The big story this week was that NVIDIA, they are talking to a bunch of Wall Street banks. They signed a memorandum of understanding to help finance more than $500 billion of AI infrastructure. Now, there's still a lot of unknowns. We don't have all the details on what this deal actually is. At the simplest level, the way I understand it is that NVIDIA is helping their customers get access to capital, whether it's cheap capital or not. They're getting them access to capital to help build more data centers, buy more chips. What did you make of that when that headline hit your phone? Is there any concerns there? So many concerns. But let's just start with NVIDIA. Jensen Wang is in charge. Jensen Wang is the puppet master. Jensen Wang does not see the world like other CEOs do, which is, okay, how do I improve my research and development and sales and marketing to sell more chips this year? Jensen Wang sees the world as something that he can construct in the manner that will favor NVIDIA. That's a very different approach to the world. And the reason he can do that is that he has $60 billion on his balance sheet. And in the next 12 months, he will generate more than $200 billion. Now, he uses some of it to buy back stock. That's very shareholder friendly. But then he has a lot left over. What can I do with it? I'm not going to be allowed to buy any other chip company. Regulators aren't going to allow that. So what am I going to do with that cash? Let me use that cash to build a world with a lot more AI demand, with a lot of investment in generating AI supply, so I can continue to perpetuate this wave, this build out. And when that happens, I'm going to sell a lot more chips. And that's what Jensen Wang does. Now, as part of doing that, he is funding his customers. That's what we're referring to as circular financing. That is not healthy. I want it full stop. Not healthy when you are financing your customers and they're buying your product. We can hold both thoughts in our head at the same time. That is not healthy behavior. That is not real demand. We can hold that thought with the second one, which is Jensen Wang has been and continues to be successful at creating a world where there's a lot more AI demand. And his ecosystem is building up the ability to generate the supply to meet that demand that will, at the end, end up with a lot more NVIDIA GPUs. We can hold both thoughts at the same time. We can be worried about circular financing, worried about leverage, but at the same time understand that Jensen Wang is busy building that world where we're all using AI a lot more. And he's willing to invest in anybody that can help him accomplish that. And then in terms of what this particular deal is, you were spot on. He sees what we just talked about earlier, which is CoreWeave now needs to borrow at 9%. They were already down to seven, and that's now gone up to 9%. That's concerning to him because he understands that whole conversation we had about the NeoCloud business. And he says, well, what can I do to reduce their cost of borrowing? I want to bring it back down to 7% because I believe long term they will be able to generate returns better than that. But if I don't lower their cost of capital, they'll never get there. So how about I step in? Again, I'm going to generate $200 billion of cash flow in the next 12 months. And so I have the resources to step in and backstop these loans, make the lenders comfortable that they can lend not at junk rates, but at investment rate rates. And therefore, I can continue to make this ecosystem healthy, build out the supply, more supply, better models, more models delivered to consumers and businesses, businesses using models more, thus creating more demand, thus creating more supply. And that way I sell more chips. Yeah. The analogy that I was making when it comes to this is like when your parents co-sign a loan for you, a car loan or a lease or something, and like Jensen's co-signing loans left and right for all these Neo clouds and all these other people that can't afford his chips anymore, hoping that they can kind of get on their feet and then become sustainable businesses. For now, it's all great because the demand is crazy, but it does make me nervous because it's like, well, what happens if the music stops? What happens if there's a slowdown in demand? And the other thing that I want to get your opinion on is like, I saw that these banks are going to be using GPUs as collateral. I guess, how do you feel about that? I mean, don't these chips depreciate over time as new chips come out? I know they're revenue generating. That's what Jensen said, which they are, but I mean, it's kind of like a car. It's like, it's not an appreciating asset, like a house or a real estate or a commercial real estate. It's like it's a depreciating asset. So I found that a little concerning. Yeah. Leverage is bad, right? Leverage is bad. We should not be building out this highly speculative infrastructure based on leverage. Amazon, Microsoft and Google have plenty of cash. They generate plenty of cash flow. They can choose to make speculative investments in that. When we start introducing substantial amounts of debt to build what is a speculative asset, that is not a good thing. Again, we can hold both thoughts at the same time. AI is good. It's going to create a lot of productivity, generate a lot of economic growth. Building it on debt with using leverage, especially if we're not pricing that debt properly, is a bad idea. It's just not a good idea. And we have to be careful about that. We can fund some of the build out with debt. But when we start funding most of it with debt, which is the path we're headed, that's a bad idea. This is a speculative asset. To your point, how quickly the assets depreciate is of no concern here. It's that they depreciate. Whether it's three years or nine years, they depreciate. And borrowing against that, when we don't know if it's three years or nine years, is not a good idea. We should not be doing that at the scale we started to do. We haven't done it yet. We've done it. We've only lent a small portion of the CapEx today. But we are on a path to inflating this thing by lending too much money at too low of a rate for what is inherently a speculative investment. Right. And that's where blowups happen, right? Leverage and too much debt. And we saw this in previous cycles. We saw it with situational awareness. That's true. We saw this with a hedge fund a couple of weeks ago. Yeah. Yeah, yeah. A couple of weeks ago. You're right. So that's where blowups happen. The market is choosing to somewhat ignore that for right now. I do wonder when that starts getting priced in again. But for now, everyone's just hunky-dory and the AI rally seems to be fully back on. Yeah. Look, when Countrywide and AmeriQuest started making the loans isn't when the problem was evident. The problem was evident years later when nobody could pay back those loans. So we're starting to inflate above. People keep asking, is it a bubble? It's not a bubble. This is inflating a bubble. When you start building speculative assets using leverage, that's when you start inflating a bubble. Whether it lasts for six months, 12 months, five years, that's a really, really big question. But at the very least, this time, we're aware of it. I like that people are asking if it's a bubble, that some people are asserting that it's a bubble because it keeps the conversation under control. It allows us to hopefully will prevent us from the types of excesses that got us in trouble the last time. Yeah, absolutely. I want to end on a more positive note. I want to talk about Palantir. Palantir reported earnings about a couple of weeks ago. Numbers, again, fantastic. Revenues were up 93%, which is accelerating, which is very encouraging. Stock is up 30% post-earnings. In your note, you called Palantir the best story in all of software, which is a very strong statement. What makes you that bullish on Palantir? Yeah, that's actually me softening my stance. I think Palantir may be the best company in the world, but I have to limit it somehow. By the way, this is the positive. This goes to that point I was making about artificial intelligence is a hugely productive technology that's going to make us all better as consumers, as employees, and as companies. The company that's doing the best job of making that happen is Palantir. If you talk to most businesses at enterprise, they'll tell you, we're trying really hard to figure out how to use AI to make our business more productive. We know it will. We know it can. We're trying really hard. We're spending a lot on tokens. We're not there yet. Then you talk to Palantir customers, and they're getting there. Palantir customers are able to use artificial intelligence to make their business better, to handle mission critical tasks in a way that's much better than they used to. That's why Palantir is special, because nobody else has been able to accomplish that, definitely not at the scale that they have and at the scale that they're going to. That's really important to know, that this is where we're headed. Artificial intelligence will make companies more productive. You just have to do it right. One of the ways to do it right is to work with Palantir. Are you worried at all that the frontier labs like OpenAI and Anthropic and maybe Google at some point, if they get their act together, could eventually move further into the enterprise and start competing directly with Palantir for some of their business? Less than I was before. What's happening is these companies are realizing, enterprise customers, companies employing artificial intelligence are realizing that it's not a good idea to work directly with Anthropic and OpenAI. This is early in that realization, but it's going to catch on. It's really dangerous to work directly with Anthropic and OpenAI. I'll give you a couple of reasons. One is Anthropic and OpenAI have ambitions. If I'm going to IPO at $2 trillion, I need to be in a lot of businesses. Right now, the whole world is handing me their intellectual property. Initially, I had to steal it. OpenAI and Anthropic stole all of our information to train their models. Now, they're settling lawsuits. It's just typical Silicon Valley regulatory arbitrage. Now, companies are handing them their intellectual property. Anthropic and OpenAI can study everybody's business and decide who to compete with. It's a bad idea to just hand them your information. That's one. Two is you can't be beholden to one of them. You can't just say, I'm going to build my business, my business application, my mission-critical system on top of Anthropic's Fable. Because guess what? The government pulled the plug on Fable and it was gone. Then when they tried to put it back, the company pulled the plug on it. If you had built an application directly on Fable, your business would be down. You wouldn't be able to operate your business. It's a bad idea to build your business directly on Anthropic and OpenAI. As that realization happens, it doesn't matter how many forward-deployed engineers they send your way. You're going to tell them, no, thank you. I'd like access to the model. I'm just going to do it through an orchestration layer, through somebody who can help me figure out for every AI task that I have, should I route it to Fable? Should I route it to OpenAI? Should I route it to an open source model from Meta or from Amazon or from Microsoft or from some small lab like Reflection? That's going to be somebody else. That's not going to be Anthropic or OpenAI. For Palantir customers, it's going to be Palantir. I was worried, but Anthropic and OpenAI have made it very clear and increasingly clear to their customers that it's probably not a good idea to work directly with them. Alex Karp had that viral rant on CNBC about a month ago or so where he talked about the role that Palantir plays and should you trust Anthropic. I can definitely see the point that he's making. I think the part that I and some investors struggle with is they get the Palantir story. They get the growth. They get the role that they're playing when it comes to this AI rollout. They do trade at 90 times forward earnings. Can they grow into that valuation? That's the key question. I'm grizzled to say the least. I have an aversion to paying high multiples for anything ever, but I cover technology and software. I've had to make myself comfortable. The way I do it is by saying, okay, well, how are they related to their comp group? That comp group is very rarified there. There's only a handful of companies that have accelerating growth over 30%. That's Datadog, Snowflake, CrowdStrike, and Shopify. I look at what those companies are trading at. They're trading at around 50 times next year's cash flow, not earnings, but cash flow. As long as Palantir is trading in that neighborhood and growing twice as fast as that peer group and with twice as high a margin as anybody in that peer group, I'm pretty comfortable with it. Got it. It'll be interesting to see what happens in the second half of the year. Before we leave, I have one last question for you. When you look across the whole AI trade, there's the chips, there's the infrastructure, there's the software layer. Where do you think the most value will ultimately end up being created over the next few years? Just based on this conversation, it seems like you think it's going to be in the software layer, but really curious to get your take there. I actually think the biggest pocket of profit will end up in the semi-pocket for the next at least two to three years. If that's the case, I look at a company like Micron and I scratch my head and I've asked people for weeks to explain to me, in what world are AMD and Intel worth 35 to 60 times earnings and Micron six times earnings? I still haven't been able to get a good answer from anybody because anything you can say that's a risk to Micron, you can say about CPUs as well. I'd argue that if you actually break the semi-business down to its parts, you'd actually come to the conclusion that the memory business is a better business than the CPU business. Yet Micron's trading at six times, AMD at 35, Intel at 60. That to me is the biggest opportunity. This is so interesting. I think you're first person that is advocating for the memory business because I think from what I understand is that memory is a commodity. Right now, what's happening with Micron is just a consequence of supply and demand imbalance. Once that gets figured out, whether it's in two years or five years, these margins, there's no way they can sustain 89% gross margins that Micron is currently doing. That's why I think it's currently trading at a much lower valuation than what AMD is doing. If you lived in the three years ago world, yes, memory was almost entirely a commodity. It is not. Today, it is not a commodity. The biggest part of it, the most important part of it, high bandwidth memory does not fit any definition of a commodity. It hasn't for a while and it definitely doesn't now. I understand three year ago thinking it's just not right now because three years ago, memory served a completely different role. Memory was used for storage three years ago. Now, memory is used to run AI models. Models run better with more memory. Models run faster with more memory. Models allow for a bigger context window with more memory. That's not storage. That's value add and high bandwidth memory is something that is designed into a data center and into a server years in advance. It's not bought a month to month right now. It's being bought five years in advance. It does not fit any definition of a commodity. That's why the stock's held back because people are still in 2023 mindset. That's okay. That's why opportunities are created. Without dislocations, there wouldn't be anything for us to talk about. If the market priced everything perfectly every day, our conversations wouldn't be nearly this interesting. Absolutely. Well, this is why we like having you on, Gil, because you bring some insight like this. I appreciate your time today and we got to bring you on hopefully closer to the end of the year. We can kind of recap some of the things we talked about and see where the market's at, you know, closer to the end of the year. Sounds good. Thank you. Awesome. Thanks so much, Gil. Well, all right, guys, I hope you enjoyed that conversation with Gil Loria. You know, I thought this was a really interesting discussion, especially Gil breaking down the I also thought Gil's explanation on why he's bullish on Palantir and Micron was very good. I mean, the idea that memory is now transitioning from a commodity business into a more strategic piece of AI infrastructure is a really fascinating angle. Let me know in the comments on what you guys thought about today's conversation and where you think the biggest opportunity is in AI right now. Drop your thoughts on Spotify and YouTube and while you're there, consider giving us a five-star rating as well. You know, all that engagement really does help us out and it helps other people find the show. Thank you guys so much for listening, watching and commenting. Shout out to Mike for all the work behind the scenes and we'll see you guys back here tomorrow.