Speaker 1So absent having a purpose, I was like, man, I guess I should get rich. That's maybe the next best thing to having a monastic calling that you're devoted to. The infrastructure to support AI wasn't going to be centralized. It was going to be distributed where energy was low cost and abundant. Energy prices actually come down. It's actually the opposite of the narrative that's being told. People are very emotional about data centers. There's really three products that we ultimately sell to customers where we're making money. We can sell data centers, we can sell GPUs, and we can sell tokens. Well, you know, the GPU is actually the most valuable thing in the entire data center. Most moats are an illusion. Most moats don't exist.
Speaker 2This is 20VC with me, Harry Stebbings. Now, my job is to bring you the most relevant interviews in technology. Right now, we are in a race for data centers, for compute capacity, for energy. And there is not a more pertinent guest that could join me in the hot seat today than Chase Lochmiller from Crusoe Energy. They provide several different. layers of this very important business, which is they provide the data centers, they can provide the compute, and then they can provide the inference, three of the most valuable pillars of this value chain. He sits down today to join me following their $3.9 billion Series F, which was raised at a $30.9 billion valuation. Get the notebooks out. This one went very granular into the build outlets required over the next few years to supply the incessant demand for AI. But before we dive into the show today, founders and investors. I'm going to start off by talking about the new Pod 6, which has a new hub that's 50% smaller than the last one. And starts at just $1,999. So, let's get started. Visit 8sleep.com slash 20VC. That's E-I-G-H-T-S-L-E-E-P dot com slash 20VC. While 8sleep helps you sleep better, MongoDB helps you build better. MongoDB has always been the database developers love. Well, now it's the data platform AI agents need. Agents need accurate context, fast. MongoDB stores, searches, and reasons over your data in real time. Jason Native Database, Vector Search, and Voyage AI Embeddings all in one place. One system instead of 10. No data pipelines to maintain. Ugh, this sounds too good to be true. Build and scale from your first user to billions of vectors. Run on any cloud, on-prem, or your laptop even. That's why 75% of the Fortune 100 run their most critical apps on MongoDB, moving trillions of dollars every single day. And it's why AI-native companies like Eleven Labs run 40 million apps every single day. So if you're building an AI, MongoDB for Startups helps you move faster with Atlas and Voyage AI credits and dedicated support. Don't build agents that answer once and forget. Build agents that remember and learn from your real-time data. Go to mongodb.com/agents. That's m-o-n-g-o-d-b.com/agents. While MongoDB scales the product, Framer scales the story. If your team wants a website that looks and feels handcrafted but is still fast to ship, Framer is built for exactly that. Framer is the AI website builder that helps creators, teams, and businesses ship production-ready sites faster than ever while getting every detail right. Prompt, inspect, edit, and publish in one place at a whole new pace. Agents and humans work in tandem. Agents bring speed and scale. You bring taste, judgment, and control. The work lands on the canvas and stays editable. Build custom code components, manage CMS content, optimize SEO, and audit for issues all in one place. Enterprise-grade hosting, security, and 99.99% uptime SLAs trusted by leading brands like Perplexity and Miro. Learn how you can get more out of your site from a Framer specialist or get started building for Framer. free today at framer.com slash 20VC for 30% off a Framer Pro annual plan. That's framer.com slash 20VC. Rules and restrictions may apply. You have now arrived at your destination. Chase, I am so excited for this, dude. I told you before, I have been stalking the shit out of you. I even heard from performance coaches. This has been an amazing experience doing the research, so thank you so much for doing this with me. I'm pumped to be here, man. I am just going to start with a totally weird one that I never expect to start with. Mountaineering. I hear that you are an experienced and skilled mountaineer. I haven't had that often in the hot seat. What do you learn from mountaineering that you think makes you better as an entrepreneur?
Speaker 1It's a great question. We've actually ingrained mountaineering into the culture of Crusoe, and we have a core set of company values. One of our core values is actually to think like a mountaineer. And so what does that mean? You know, mountaineering is a practice where things are subject to change, and you need to be prepared for change. With it, it really embeds this sense of cultural resilience that I think is really, really important for a business like ours that is doing stuff both in the digital world as well as the physical world. But to further the analogy here, for mountaineering, you come up with a plan to basically set out from your camp, go some at the peak, come back safely, and take your photo, and your glory is written in the history books. But you have to be prepared for things to go wrong, right? You have to be prepared for the weather to change. You have to be prepared for gear to fail. You have to be prepared for your partner to get sick or an avalanche to occur. And the idea is really having here's a plan A, and then I have a plan B, a plan C, a plan D. I have plans all the way down in case different things go wrong. You know, additionally, like you have to have a sense of willingness to endure difficult challenges and difficult times. You know, this notion of enduring the expedition. The expedition could be long and hard and arduous. And, you know, you have to really dig deep at those moments. And I think that's a really important part of mountaineering. And I think that's a really moments of pain to work through those challenges. That's a different aspect of thinking like a mountaineer. And then, you know, another big aspect is thinking with a safety culture, having this notion of safety as a huge priority, which is important when you're doing things at the scale we're doing from a construction standpoint, thinking with the safety mindset of, you know, safety first, second, and third. There's this saying in mountaineering that getting up to the top is optional, getting down is mandatory. We try to practice that safety culture here at Crusoe. And really, you know, try to avoid any sort of physical injuries or challenges with digital attacks on the
Speaker 2platform as well. I have so much respect for mountaineers. It's one of those ones where I almost can't get my head around the complexity and the mental fortitude it takes to get through the change of plans, the weather, the partner that's sick, the endurance, as you said. I met an entrepreneur yesterday that climbed Everest, and he showed me pictures of the dead bodies along the way. I mean, to see that is just so harrowing.
Speaker 1It's tragic, just kind of, you know, seeing when people take unnecessary risk because, you know, actually spending a lot of time within high altitude mountaineering culture and communities, you do see oftentimes people taking unnecessary, very careless risks. It's actually quite shocking. Oftentimes, tragedy is avoidable. It is preventable. And, you know, I think having that mindset of thinking like a mountaineer and planning ahead is what enables mountaineers to do it for a long period of time.
Speaker 2Can I ask, you said about resistance. Resistance to change. What change did you find
Speaker 1the hardest in your life? I think maybe one of the most difficult moments of change for me was in my life, I always had sort of the next thing planned before I left the previous thing. So what I mean by that is when I left high school, right, I knew exactly where I was going to college before, you know, the end of high school. When I left college, I knew exactly where I was going to work and, you know, what I was going to do. And when I left my first job, I knew I was, you know, going to work and going to grad school. And I had a job lined up even after that. When I left to go climb Mount Everest in 2018, I did not have a plan after that. You know, I had been, had a reasonable amount of financial success in my life to that point, but I didn't have the thing planned. And actually having that, that void, that sort of emptiness really was both challenging and empowering for me because it felt like I could do anything. It was sort of like a blank slate. I had thoughts of like, maybe I'm just going to invest my own money and just kind of like enjoy the rest of my life and not do much, but I could also start a company. And if I were going to start a company, what are the most, most challenging things that I could, you know, think about tackling that are interesting and that I would, I would be excited about pursuing on a daily basis. And that was kind of like the emergence of Crusoe sort of came through that chapter of difficulty and challenge of just nothingness, right? Just having an empty, empty slate.
Speaker 2My job is to find signals that will determine a potentially great entrepreneur. And one thing that I find now is actually often richer entrepreneurs are better entrepreneurs because they're not worried about downside protection and they just see what is possible. How do you think starting Crusoe with a little bit of money and success already changed how you think about building it?
Speaker 1Oh, I think it absolutely changed the way I think about building it. There's a bunch of like psychological aspects to it. I mean, you could look at like Maslow's hierarchy of needs, you know, and just like, it is empowering to know that like, if I take risk and this all goes to zero, I'm going to be able to do it. I'm going to be able to do it. I'm going to I'm still going to be fine. I'm going to be able to cover my rent and, you know, feed my family and cover these base needs that I may have in my life. When I was in university, I mean, I was convinced I was going to become a theoretical physicist and likely a professor. And I was spending time doing research. in basic science, right? So I was doing research in physics, both at MIT and then at Los Angeles National Lab. So early on in that journey, I was like pretty committed to living this monastic life of like not really earning a bunch of money, but just having this life of discovery where I get to go discover the mysteries of the universe and that would be my life's work and pursuit. At some point, I realized it's a very slow moving industry, this industry of discovery. It may be changing today because of AI, but at least at that point in my life, it was a very slow moving industry. And I felt like I wanted to do something very fast paced. And I felt like I had sort of lost my life's purpose. Like I was like, okay, I don't, I don't feel like I have a purpose anymore. So absent having a purpose, I was like, man, I guess, I guess I should get rich. Like that's maybe the next best thing to like having a, you know, a monastic calling that you're devoted to. You know, I was looking around for different ways to make money and this headhunter found me and recruited me into this field of quantitative finance. And it was like, hey, you're really good at solving math problems. If you come work in this industry, they just pay you a lot of money to, you know, solve math problems. And I said, this sounds great. Sign me up. But I think once that need was met of like, okay, I have enough money. Okay. What, what next? And, and I think to your original question of like, you know, starting from this position of strength of, you know, being an entrepreneur and actually having had some financial success, it's empowering to me. And I think that's what I'm trying to do. And I think it's empowering to take bigger swings, empowering to take bigger risks, which can ultimately lead to bigger outcomes.
Speaker 2The final one before you actually get into what I plan to get into. Yeah. It's just, it's just picking up on, you said kind of the endurance element of mountaineering. Endurance is great in some respects, but it can also lead you to continue doing something you maybe shouldn't. And sometimes there is a right time to change course rather than persevere. You went through a very big change in the business, seemingly so from the outside. A lot of people are like, oh, we're not going to change where we are today. But seemingly it's like a Bitcoin phase one. And then phase two is AI. Yeah. I don't want to kind of really actually get into that maybe weirdly, but I want to ask your advice, which is when asked, how do you advise someone when to endure, keep going, persist versus when to change tact and go down a different route?
Speaker 1It's interesting investors that you spoke to validated this claim, but, you know, from the outside, it does appear to be a pivot. Internally, when we started the company, the goal was always to build an AI platform. That was what I had sort of spent my career working on with AI machine learning algorithms. And, you know, I've been a big user and early adopter of deep learning for certain tasks. And I felt like it was this meta science that ultimately was going to transform every industry. And, you know, at that scale, compute becomes the bottleneck. Compute and data become the two bottlenecks. And, you know, what is the bottleneck for compute? Energy. So that was kind of the premise of starting the company. Now, we obviously weren't dedicating all of our resources towards building an AI platform day one. We were dedicating actually the majority of our resources towards monetizing waste energy with Bitcoin mining. That was kind of the philosophy of starting the company and finding and seeking out these low cost, abundant energy resources that we could monetize with compute. Bitcoin was the best monetization engine we had at the time, but we were building this AI platform from very, very early days. So we were building this AI platform from very, very When you think about at what point do I shift the resource allocation to this other thing that maybe I'm working on on the side, I think about the world in a very Bayesian approach. I don't think the future is deterministic. I think it's very probabilistic. And I sort of have a view of what the future is going to look like. It's more certain if you go out just a little bit of time. If you go out a further amount of time, it gets a little bit hazier. But I still have a view of sort of what things are going to unfold. And what was happening with this AI platform was it was sort of a project that Chase had devoted some time and resources and engineering towards internally. And we were committed to exploring this pathway. But we had launched the Crusoe Cloud platform in early 2022 with our first set of paying customers. Before that, we were working with a lot of researchers at MIT and Stanford and different groups trying to understand what's actually valuable for AI researchers from an infrastructure platform. So we finally launched in early 2022. But the moment that changed everything was obviously November 30th, 2022, the launch of ChatGPT. This was like the tweet heard around the world, the moment that just shifted everything in terms of like, OK, AI is here. And there are really interesting use cases that are going to explode from sort of this moment. And that was kind of like a moment where my view of the world shifted quite a bit. How so? I just felt there was going to be far more demand for AI compute infrastructure solutions. And we were very well positioned by actually having launched that platform at that point.
Speaker 2What did you do then differently immediately as a result?
Speaker 1After the chat GPT moment happened, we had sort of watched this evolution of the data center to support Bitcoin. And if you look at the history of Bitcoin mining, it's pretty fascinating, right? It started with like hobbyists on their laptops in this sort of decentralized way. And then people started to use GPUs. And then people started to lease data center capacity from more traditional co-location facilities that have like five nines of reliability. People started to develop ASICs. And then at some point, people realized, you know what? We don't need five nines of reliability to mine Bitcoin. We can do it with three nines. Oh, wait, no. What about one nine? Like how much does that reduce the actual data center cost? And they were able to eliminate something like 98% of the total data center cost with like these like chicken coop type data centers. You know, very, very low cost, no redundancy, but, you know, ultimately brought down the payback period for this investments in Bitcoin mining. And, you know, we felt always that something similar was going to play out in AI. And we're long-term partners of NVIDIA. And we'd been working with NVIDIA. And I was looking with them at the roadmap, you know, many years ago. And I was like, okay, our current generation of chip is, you know, 150 watts or 200 watts. And, you know, the next generation is going to be like 300 watts. And then maybe, you know, there's going to be a future generation that's maybe 600 watts per chip. And I was like, okay, well, you know, if as the power density goes up, it's going to change what the data center looks like to actually support these things. And we also had a perspective on the compute workload, the workload itself for AI. It didn't need to be in these centralized facilities. I think so long people have been anchored on these like centralized hubs for data center computing infrastructure. So Northern Virginia is like the hub that runs most of the internet. Most web applications run out of Northern. But it didn't have to be that way with AI because, you know, so much of the time to serve a neural network is actually the compute that's happening in the data center, not the time to get to the data center. And so that I felt like opened up new geographies. And we felt like the infrastructure to support AI wasn't going to be centralized. And so when we think about things that we did differently, it was like investing more heavily in how to design. And develop the data center to support AI and investing heavily in how do we find and source energy resources to support the scaling of AI.
Speaker 2I want to kind of unpack those a little bit and kind of going back to some form of structure that I did plan. I promise that's great. If we start on data centers, we then have Crusoe Cloud, managed inference, energy kind of runs throughout it all. So I'll kind of pick that apart, too. On the data center side, we always hear about the supply constraints today. How do we do that? How truly supply constrained are we?
Speaker 1Where it's actually manifesting is there are not places to plug in GPUs. So that's ultimately the supply constraint. There's just not places where you can plug in GPUs and turn them on and run AI workloads. So how does that actually manifest across the supply chain? You know, the supply chain to support large scale AI data centers is sort of like a game of whack-a-mole. At different points, different bottlenecks come into play. And I think that's actually one of the critical aspects of being vertically integrated. Integrated is that we're able to navigate a lot of those critical bottlenecks and situations. I'll give you an example. When we set out to build the first two buildings in Abilene, it was a little over 200 megawatts of compute capacity. We had sort of committed to doing this in one year. And the next closest bid was two and a half years. There were 34 other data center developers. And, you know, but speed was of the essence. So I said, sure, we can do it in a year. One of the key bottlenecks was what's called a power distribution. So that's the power distribution center. It's where, you know, you receive medium voltage power and then you, so at around 34, 34.5 KV, and then you distribute it to low voltage transformers that ultimately go into the data center and power the racks. When we went out and canvassed the market for vendors that were supplying these medium voltage power distribution centers, the lead time for this one component was a hundred weeks. And I said, well, I don't have a hundred weeks. I committed to doing this in a year. So I went to our internal team. And we had. We integrated electrical manufacturing. We said, how quickly could we make this ourselves? The team did a bunch of work, sourced components to make these things. And we were able to do it in 28 weeks. And so being able to navigate those types of key bottlenecks by actually having the resources internally gives you ultimately a lot more flexibility to unblock these key bottlenecks. Number one, and number two, it gives you a perspective on like, what, what is the true cost of a lot of these different things that are being built? Elon has this very like interesting. Framing of this, he calls it the idiot index in terms of, you know, what, what is the, what is the cost of raw materials and then what is the end product relative to that cost? cost of, you know, just the raw materials going in. By being vertically integrated, we have a perspective on like the end-to-end cost of everything going into building and operating an AI factory.
Speaker 2How much extra margin do you juice out by having that internal build process yourself? Like what is the idiot index chasm?
Speaker 1Look, electrical manufacturing is not like an ultra high margin business. Like today it's actually spreads have gone up quite a bit. Margins have gone up quite a bit just because they're so short in supply, but it's not where like the critical alpha of, it's not the main reason we're doing it. It's like the stacking margin. It's more availability and getting availability, being able to deliver on time. And then also it gives us a critical innovation platform by being able to, you know, start from first principles and say, okay, we can bring in raw materials and we can build anything we want in the world. What is the thing that we want to build to ultimately serve the workload? And so when you look at the campus analysis, it's like, okay, we can bring in raw materials and we can build anything we want in the world. And so when you look at the campus analysis, it's like, okay, we can bring in raw the inspiration for that was like, okay, we want to build a one gigawatt scale computer where you can interconnect a giant cohesive cluster of GPUs that can operate one cohesive workload across the entire campus. And what does it look like to actually build that thing from, from the ground up? That's how, you know, the cores are designed. That's how the GPUs are designed. That's, you know, all of it is designed to run on one cohesive RDMA fabric. When we think about scaling of inference, you know, 2026 is really about the scaling of utilization of AI or utilization of LLMs and foundational models for useful tasks. You know, it's been this era of, you know, scaling of agents and tokens, token utilization. And, you know, during that moment of scaling infrastructures, people are finding utility in, in these AI models. It's driving a lot more demand for inference. It's the era of the agent, the era of, of inference. And for that, you don't need a gigawatt scale computer to serve these models. You can do it with a much smaller cluster. What matters is that you don't need a gigawatt scale computer to serve these models. What matters is actually how quickly can I go from not having anything to being able to deliver tokens? Like, so what is my time to token?
Speaker 2I love the analogy you said there of it's like whack-a-mole. What is the biggest supply constraint that we have today?
Speaker 1Energy is definitely a key constraint. Power that's available for compute is a key bottleneck. Alongside that, labor is a very key bottleneck. There are a finite amount of skilled trade workers, electricians, welders, in the United States. And then when you overlap, trying to find if I have access to energy in this unique place, can I get the labor to that place to actually bring the AI factory to life? You know, that, that, that is a big challenge.
Speaker 2With the greatest of respects, if power and labor are the two biggest constraints or challenges, the UK is totally fucked. We have energy prices that are four times higher than anyone else. And we don't have labor and we have humans. They just don't work. You in the US have both, but significantly more expensive than China. If we're honest.
Speaker 1Power is not significantly more expensive.
Speaker 2Really? No. You don't think that the government is subsidizing data centers in China with much cheaper energy to power? I'm asking him, not telling.
Speaker 1There are subsidies going in place, but in the United States, we are globally competitive with power pricing. Labor, I'll give you. Labor in China is far cheaper than the United States.
Speaker 2What about policy and regulation? Everyone I speak to is like, oh, it's the red tape. It's the policy and regulation that stops us in the US being able to build out. Is that a problem in the way that it's told to me?
Speaker 1I wouldn't classify it as a problem. I would classify it as something that needs to be navigated. When we think about policies, we want good policies to be in place. We don't want just the wild west, everybody out doing whatever they want, building whatever they want with no care for anybody else in the world. These are big investments that are being made. They have to be done thoughtfully. We are supporters of the right policies being in place, but it does cause friction in terms of developing things faster. That's actually one reason we also like the idea of manufacturing a lot more of this infrastructure as opposed to making everything a giant construction project.
Speaker 2If I were to put you in charge of policy to enable and encourage this ecosystem to flourish, what would you do? What would you do differently?
Speaker 1The thing about policy is it's hyper-local. Politics are local. What people care about is they care about, are there jobs being created? Is this going to drive up my energy costs? Is this going to take all of my water? Is this going to pollute the air so that, you know, my kids get, suffer from some disease as a result? Like, you know, they care about, they care about, are the data centers being good stewards and good citizens of the community? And are they actually creating value for the community? And then what are other ways that the data centers are sort of giving back? And so, you know, this comes in the form of things like tax revenues and things like that. So I think one of the challenges for AI data centers broadly right now is that I think there's a lot of misinformation out there about what the impact of a data center is. People that have told me that data centers use all the water in the world is crazy. It's also just false. When you look at all of these modern AI factory designs, they use pretty close to zero water. One of our giant buildings in Abilene, Texas, call it 140 megawatts of power is like what's budgeted for each of those buildings on the first eight buildings. They use about the same water annually as about 10 single family homes. The primary water usage there is like the staff using the bathroom, washing their hands. Us watering the plants on the property. It's very, it's very de minimis water usage. And there is water in the building. We are using water to cool the GPUs. The critical aspect here is we've actually designed a closed loop architecture. So the cold water is entering into the racks and then it's going out to a chiller outside and we're sort of exhausting the heat. So the water requirements argument is just wrong. The water requirements argument is just wrong. Yes. Okay. What about energy prices going up? So, I think this is an important aspect for the data center industry to get right and to storytell effectively. We like to look at data for these things. When you look at markets where data centers have made investments and have built big data centers, typically energy prices for communities have come down because they're, it catalyzes more investment in energy generation technology and energy generation capacity. And, you know, with more megawatts being amortized over the same transmission and distribution infrastructure. And as a result of that, people's energy costs end up coming down. When you look at some of these big builds that need a lot of net new generation capacity, we are very supportive of, you know, the data center industry helping to bring online new power production in order to support those energy requirements.
Speaker 2So actually that too is misinformation. Energy prices won't go up for local towns and villages.
Speaker 1The data has shown that energy prices actually come down. So what is a legitimate concern? Anytime there is a big construction project, it does cause other aspects, right? There's, there's benefits, there's people with money in town that, you know, are spending. And, you know, if you talk to any local business owner in Abilene, times have never been better, right? The amount of money being spent at restaurants and coffee shops and hotels and all the local services is, you know, is, is more than it's ever been in the history of Abilene. The challenge is with that comes traffic, with it comes some dust from construction and some, you know, just noise that, you know, is, is involved with a very large construction project, but that will go away and there will be long-term permanent jobs in place that, you know, is continuing to support the local community. And there's tremendous tax revenue, right? If you look at the tax revenue that's coming from this, we're going to account for more than a third of the tax revenue in Taylor County and Abilene, the town. And this is transformative for local services, like the police, the fire department, the roads, the schools. For the school system, we're more than done. We're more than done. We're doubling the tax receipts that are going to the schools.
Speaker 2I completely agree with the ripple effect that you see. I think it's one of the worst boring things when we see the amount of millionaires leaving the UK. People don't consider the ripple effect of them leaving in all of these different ways. We're seeing a lot of data centers that are planned not being built out. We've seen very prominent people say, oh, we expect 50% of data centers that are planned not actually to be built and ready to go. What percent of data centers do you think that are planned will not actually go online? 50% seems reasonable.
Speaker 1There's, it's one of these things, you know, it comes back to thinking like a mountaineer. You can't give me a decimal point.
Speaker 2Why do you bother? Let's end it here. Neil told me you'd give me precision.
Speaker 150%. You know, it's one of these things where coming back to this notion of thinking like a mountaineer, when you're going through these planning processes, a lot of things can go wrong. What's the number one
Speaker 2thing that goes wrong
Speaker 1in Provenza? Getting permits, getting entitlements, getting land. You know, you're trying to acquire a certain piece of land and, you know, the person you're trying to buy it from ultimately doesn't want to buy it. Getting a large load interconnection agreement done with the local utility, getting an air permit if you're bringing online new generation.
Speaker 2With the greatest respect, I don't come back to this and I'm not an agent of the CCP, I promise. But in China, I mean, they'll move the 3 million people out the dam for you with a forklift and they'll just give you a green light on day one.
Speaker 1Yes. It's a benefit to the data center developers if you're viewing it through a purely Machiavellian lens. Yes. I am
Speaker 2a venture capitalist. I give zero shits about the sentimentality of your home. You will be rehoused somewhere better. Again, charity is not our endeavor. Okay. When we look at these data centers, now they're being used as like a political instrument in a way that we didn't expect. Does that worry you?
Speaker 1Yeah. I think I've never wanted to be the main character of a political policy debate. And, you know, data centers have come front and center to the upcoming midterms in the United States. And they've been doing it for a long time. And I think it's a good thing. They've become this very polarizing topic. Why that is, I think they are kind of this – I think there is concern around this era of AI. And I think people are legitimately concerned around, like, am I going to have a job in the future? Am I going to be able to provide useful work to the world that I'm going to be able to get compensated for? So I do think that that is a concern on people's minds. And data centers are sort of this physical manifestation of AI. And putting all of that to side, the biggest aspect is people are very emotional about data centers. I think when you look at the facts of what data centers are doing from job creation, from driving down energy costs, from being water neutral, from generating long-term tax revenue that is transformative for the communities that we're investing in, I think the facts are all on our side.
Speaker 2Do you know what I slightly blame? With respect to Dario and a lot of other leaders who've said for the last few years, we're going to replace all jobs, we're going to replace all jobs. And then surprisingly, they hate you. Gosh, cake or death. Cake, please. I wasn't expecting that. Oh, crikey. You won't have a job. Oh, you don't want the data center?
Speaker 1Yeah. The interesting aspect to all this is that data centers are this massive employer. They're leading to this massive economic boom for the entire blue collar labor economy in the United States. This massive resurgence in re-industrialization in the United States. We're hiring people in factories, in the field, and these are- Tons of folks that are skilled trade workers, right? People working with their hands to bring the infrastructure of intelligence to life. I think the interesting aspect that people are concerned that AI is going to take jobs, but yet it is so far only created tremendous amounts of jobs and tremendous amounts of economic development. So again, I think it's one of these moments where people are having an emotional reaction to something, but if you look at the facts of what's actually happening, the data is on the side of data centers. It's actually a very positive investment for the communities.
Speaker 2If anyone's doing data center security, by the way, I would love to invest in that business. It's a really good- It might be a more PE play, to be honest, not a venture play. It's a phenomenal business. Another part of your business is obviously providing GPUs and compute. Can you help me understand the economics of that layer? When you look at it, I'm a venture investor. How long does it take to pay back on the hardware? I would say the
Speaker 1price of a GPU. It's kind of like a traded commodity. You're even seeing these commodity platforms creating compute as a tradable asset. I think people are thinking through this lens of compute is going to be this next great commodity. So there's a couple of platforms like the Compute Exchange, Orn. I think there's a handful of others that are creating these tradable futures products. So I view it as that's like a product that we're producing. And there is like, you know, there's a lot of product that we're producing. And, you know, there is like, you know, there is like, you know, commodity price fluctuation in it. When I think about payback periods for this, I think when we are taking debt on against it, there's different ways we think about like managing that risk. And we sort of think about a portfolio of different services and a portfolio of different returns that we're going to get for this.
Speaker 2What does that mean?
Speaker 1So we have a range of different deals that we're doing with the compute we're buying. We are renting capacity on a long-term basis, you know, call it five-year contracts to credit quality customers that, you know, we expect to pay. Those are going to pay for it. They're going to pay back within, you know, that timeframe, and they're going to cash flow during that timeframe. There are shorter term contracts that we'll do at higher margins, but they're riskier because at the end of the contract, it's like, is there a renewal? Is there another thing that you're going to, you know, earn from that resource? And then finally, there's other services that we provide in package. So things like Crusoe Managed Inference, things like our serverless fine-tuning product that actually helps people, that's more of a developer tool that ultimately enables developers to, you know, fine-tune, some of these open source models, these leading edge open source models, so that they can get better performance for maybe some domain-specific task. Or actually, as their applications are scaling, they can actually serve the tokens needed to, you know, from an inference basis to, you know, enable that application scaling. Those contracts are typically much shorter term in nature, and they end up being higher margin to Crusoe.
Speaker 2And so you want a portfolio of those different margin profiles to make up what you deem as like a healthy margin across and blended.
Speaker 1Correct. So one of the analogies I've used in the past is, when we think about Crusoe and sort of our vertically integrated strategy, there's really like three products that we ultimately sell to customers where we're making money. Those are the three core products that we sell. And why does that matter? You know, if you look at a different market, I think that there's very interesting parallels for, you know, AI infrastructure and AI compute is the energy market. If you look at the oil and gas market, and you look at the value chain from, you know, the oil and gas market, and you look at the value chain from, you know, the wellhead to the product, you know, the gas station or, you know, the plastics that are being sold to people, there's a lot of different ways to make money across the value chain. But typically it gets broken down into the upstream business, which is drilling wells and sort of, you know, producing oil. There's the midstream business of like transporting oil and gas to, you know, a refinery. And then there's sort of the downstream business, which is typically refineries, gas stations, basically getting it into these finished products that you're ultimately selling to customers. Well, what Crusoe's focused on doing is what we believe the opportunity is, is to build an AI super major, right? So there are these super majors like Exxon and Chevron in the traditional oil and gas industry that are vertically integrated across upstream, midstream, downstream. Where that becomes actually important is that Exxon very famously doesn't hedge their oil exposure. And you say, how could that work, you know, during this, you know, very crazy commodity price cycles? Well, they're sort of inherently hedged by being vertically integrated. And I think that where the margin accrues is going to move around. And I think you're seeing a very similar phenomenon play out in the oil and gas industry. It's going to play out in AI infrastructure. But when oil prices come down, Exxon makes less margin at the wellhead, but their margins, their downstream business actually go up. So, you know, because the oil prices are lower, the margins on gasoline, on plastics, on all these finished products, they actually go up. And I think we're seeing a similar phenomenon play out in AI, where it's like, you know, our margins are going to move around across electrical data centers, chips and services. What's the best margin layer today? I mean, the best margin layer managed GPUs, like managed compute clusters is like incredibly high margin, like right this instant. You know, the reason for that is just a massive shortage of supply.
Speaker 2There's a phenomenon or a kind of a statement of take or pay in this business. Can you explain take or pay to me and to anyone that doesn't understand it?
Speaker 1Sure. So, take or pay is essentially where a customer is signing up to pay for something, whether or not they use it. So, you know, there are take or pay contracts in energy oftentimes where you know, if I commit to 100 megawatts of capacity, I'm paying for 100 megawatts of capacity, whether or not I use that.
Speaker 2Are all of your revenues take or pay?
Speaker 1The GPU rental agreements are typically take or pay.
Speaker 2Yes. Because I think a lot of people are looking at going, well, the first sign of trouble would be a crack in consumer demand. And if that crack in consumer demand permeated down and it was take or pay and they went, ah, don't need to take now. Is that correct?
Speaker 1I don't really think so. I don't think of it that way. Because the thing about AI is it's transformative for so many areas of the economy. It's not a single thing that's benefiting every single industry. So, you don't agree that the
Speaker 2first crack would be consumer demand softening? Consumer demand in what though? I think it would be consumer demand. Like chat GPT or? Yeah, consumer demand across chat GPT and core.
Speaker 1Yeah. I mean, I think we're honestly just scratching the surface. I think like, you know, what's exciting to me is, you know, seeing, you know, the step function improvements that we've seen from some of the more recent models that are fundamentally enabling new discoveries that weren't possible. Model capabilities are only going up. When people think about the utility of these models and of this type of infrastructure, it's the worst it's ever going to be for the rest of the future of humanity. When you sort of take that perspective, it is inspiring to what's actually possible. If we saw consumer demand softening for one very specific thing, I couldn't be more optimistic or I couldn't be more bullish for like how this is going to transform scientific discovery of new materials, of new drugs, of new things that are ultimately going to make people's lives better.
Speaker 2When you're buying chips to the scale that you are, you have to think about life cycle. How do you think about chip depreciation risk?
Speaker 1Well, the way we depreciate the assets today is we use a six year depreciation cycle. That's sort of like the standard across the industry. Now, one of our philosophies as a business is like, look, we are going to be long infrastructure. We're going to be long data centers. We're going to be long energy. We're going to be long chips. And so the more ways we have of turning those long investments into money, the better off our shareholders are going to be and the better off our business is going to be. And so that's actually what caused us to investing in this suite of managed AI services. And when you look at that business where it abstracts away the the actual underlying chip, the underlying compute from the service that people are receiving, it opens up new monetization engines that can persist for a much, much longer period of time. And there's always going to be demand for the frontier of silicon. There's always going to be demand for the frontier of, you know, model capability. But there's also going to be demand, especially for for much cheaper alternatives and, you know, being able to provide intelligence at a lower cost with a older generation of chip that may be slower is still going to have value. And so, you know, abstracting away some some of the difficulties by actually, you know, having a suite of services that can monetize these chips for longer will ultimately we believe. extend the depreciation cycle beyond six years.
Speaker 2What do you think everyone gets wrong about chip buying, chip depreciation, that they should know?
Speaker 1I think when we started to make really substantial purchases with hoppers, and this was in 2023, the feedback we got was like, well, we don't even know if this is going to be valuable after year three. Here we are three years later, and the prices being charged for utilizing hoppers is higher than the rates that were being charged three years ago when they were brand new. Early financings was you had to get payback very, very quickly, and you had to have a lot of debt service coverage. I think people underestimate the ingenuity of applications and application developers of turning compute capacity into value and into valuable services for the economy.
Speaker 2I spoke to Lee Jacobs before, and he was talking to me about the risk mindset that you had to ordering millions of dollars of GPUs from NVIDIA well before it was cool, as he put it. My question to you is, how on earth do you think about forecasting demand today? Is it just, I'll take as much of it as I can?
Speaker 1No. I mean, we build up forecasts from having conversations with customers. I think we are a very customer-centric organization. We want to deliver ultimately solutions for customers. The challenge today is that increasingly customers are being asked to forecast their demand further and further out because the timeline to bring online AI infrastructure is extending. And that's problematic for the industry writ large. One of the things that Crusoe's focused on is actually trying to reduce that time to token or time to delivery of a managed cluster of GPUs by actually deploying small modular manufactured AI data centers that can deliver more just-in-time AI infrastructure, particularly for smaller clusters. And we think that's going to be transformative for the industry. We think it's a very important shift in terms of, you know, especially like when you look at earlier stage companies, earlier stage startups that are being asked to commit to compute, compute in 2028. It's like, that is an eternity from now, right? You know, this is, this is like a. We might have RSI by then. Exactly. That might change it. Exactly. Exactly. So.
Speaker 2Have you ever had a moment in your career where the future is so unknown?
Speaker 1I feel, I actually don't feel like it's super unknown. I actually feel like, you know, we've had this breakthrough in overall model development cycles. You know, I think the scaling laws have. We've continued to prove out. I think we've reached, you know, we keep moving the goalposts of what is, what is artificial intelligence? What is AGI? You know, you think about the, the Turing test. When I took my very first computer science class in high school, the Turing test was heralded as this, as this critical notion of true machine intelligence. And it seemed like an impossible task at that point in time. We blew past the Turing test and almost no one even celebrated. You think about, you know, the model capabilities you have today, and now, you know, we're solving millennium prize problems, right? This was something, you know, when I was an undergrad at MIT, these millennium prize problems sort of came about, and it was like this, wow, these are these unsolved problems where there's a real financial bounty out there. Like you can earn a million dollars if you can solve one of these problems. And they sat there for decades unsolved. Now we have AI solving these, these, these critical
Speaker 2problems. I agree with you. My question to you is, we see these swarms of rogue agents doing things that are nefarious. We see whole industries that are being put to question. My girlfriend's a lawyer. Six months ago, she didn't use the Agora. Now she hasn't written a document in five months. There is just such uncertainty across everything. We have RSI that could come and change absolutely everything.
Speaker 1So, sure. I get the comment that there's uncertainty around what the future of work will look like in certain ways. But I also feel like it's, it's actually not uncertain that AI is going to be a critical underlying engine for work and productivity. That, that is going to shift things quite a bit, quite dramatically. And so it's more about there is certainty that AI is establishing capabilities that are at human level or greater than human level in terms of overall productive capacity. So, sure. Is that going to change things? Yes, that is going to change things. So, so there's uncertainty in the change, but not, to me, it's certain that there will be change. Maybe that's a different way of putting it.
Speaker 2Kaz, we mentioned the managed inference being a great part of your business. Gavin Baker, one of your investors, has said that the lowest cost producer of intelligence wins. When you think about that, like what's the unit? Is it like dollars per token, per training run? What is that unit for you?
Speaker 1Dollars per token is a good measure in terms of this, but it's, it's not necessarily, you know, there is token efficiency. So, you know, just not all tokens are created equal. So it's not like every single token has the same, you know, unit of value, but I think it is a good indicator. I do think that, you know, there are different things that folks optimize for in terms of utilization. Whether it's if throughput, so the amount of tokens per second you're able to generate for, for, for a very specific model use case, as well as time to first token. So latency, time to first token, time to last token, you know, how quickly you're able to actually serve this. And a lot of that comes down to, it's a confluence of, of different things, right? So there's, you know, what is the cost that go into that? Obviously by being vertically integrated and owning the infrastructure, you're able to stack a lot of margin across the process. So it gives you more flexibility in terms of how you actually, you know, build for that. Additionally, I think utilization of the GPU itself is a critical aspect of being able to deliver tokens ultra efficiently and shine on metrics like latency and throughput. And so how does that actually manifest? Well, you know, the GPU is actually the most valuable thing in the entire data center, right? It is like the most expensive thing. So it's the thing you want to make sure you're keeping busy. If it's sitting idle, that's just, that's just money burning, right? If you think about that from, from that perspective, I think one of the key aspects is actually the management of memory and the management of the, you know, of the KV cache specifically. So the KV cache is the key value cache. It's really like, you know, you can kind of think about, you know, when tokens are fed into a large neural network, you can compute the output tokens by basically running this feed forward process and running all these matrix multiplications that takes time. You may also already know the answer of that output token before you put the, so if you, if you've actually put, put those same tokens in before you actually know the answers and you can look them up in this lookup table, the KV cache can get quite large. And so it can expand well beyond the amount of, you know, memory you have on chip in the HBM. And so actually being able to manage that both across multiple different GPUs, HBM layers, as well as system memory, DRAM as, as well as things like MVME. So any, any sort of, you know, solid state drive that you have on the system, and then sort of expanding that into the object storage layer of your cluster and being able to really, really efficiently move data around is actually a critical aspect to being able to serve inference and be able to keep those GPUs busy so that you're able to really shine on those metrics of latency and throughput.
Speaker 2What does no one see about this layer of the value segment that you provide that everyone should see? What do we not know?
Speaker 1Well, I think the big aspect is that not all inference providers are created equal. I mean, you, you have like, you know, great open source projects like, you know, SG Lang and VLLM, but you know, there are ways to, you know, those are, are very general, generalized or generalizable. When you look at like an inference provider, like Fireworks. Fireworks is great. They've done an incredible,
Speaker 2incredibly great job. To what extent do they compete with you and cannibalize your managed inference business?
Speaker 1Look, I, I think there are multiple layers across the stack that Crusoe competes with.
Speaker 2What do they have that you don't by being specialized?
Speaker 1I think there are certain workloads that, you know, Fireworks, you know, does an incredibly great job serving. I think, you know, the user experience, I think is something that is unique across all of these different inference platforms that, you know, people may like more on, on one than the other. There's a suite of different tools that, you know, people are, you know, there, there's a whole bunch of different features that people basically provide across, you know, the inference ecosystem.
Speaker 2But there's nothing like you're, ah, I wish I was them because then we could do this.
Speaker 1I sort of come back to this notion of, you know, the oil and gas analogy, where the, if you look at the value chain, there are different places where you can compete. And there are different places where you can collaborate. When Exxon is pumping wells at the wellhead, they may compete with another upstream oil and gas company to, to, you know, most efficiently drill and operate wells, but they may sell to those same customers. If they're building a midstream pipeline to transport oil from the wellhead to the refinery, they may provide capacity in that pipeline to other folks that they're competing with. And so, again, I think it comes down to this perspective that like, I just view there are so many more ways for us to work with folks across the AI infrastructure stack than compete with them.
Speaker 2Can I just ask you, when you look at like open versus closed, given the purview that you have from the managed inference perspective, what do you see about distribution of dollars of tokens across open versus closed, given that unique perspective you have?
Speaker 1I do think that people are spending more money on closed source, on frontier models than they are on open source, but they are generating more tokens on open source than closed source. We think open source is actually going to be very important, very valuable. I think it's important for, from a data sovereignty standpoint, folks wanting to own their own models, own their own intelligence. I think there's far more private data that's untapped that can actually lead to massive performance gains.
Speaker 2Do you think we will exist in a world where many companies, both mid-market and large enterprise will have their own models with their own data, and that will actually cannibalize a lot of frontier model business?
Speaker 1I think it's possible, but I do think there's always going to be demand for the frontier. Like the frontier is- I'm not
Speaker 2saying there's not, but you're your Harveys and your Rams and your, you name it. And we're an investor in McCall, we're an investor in Fireworks. And both of those enable that. They propagate that as a marketing message for sure. And I'm just wondering, is that true?
Speaker 1I do think that the model capabilities today, I think there's massive room for improvement by incorporating more private repositories of data. You can do that with the frontier models that are closed source. You can also do that with open source and own the model yourself. I think it's going to be some combination of those things in the future. I don't think it's going to be like one approach is going to be the dominant approach. I know there's a lot of exciting stuff happening between folks like Cognition post-training their own model with a lot of their own data. Harvey kind of maybe trying to take a similar approach from a legal perspective. There are a lot of these amazing domain specific model approaches to things. The frontier labs are also doing that as well. So they're also looking at that and they're trying to basically increase the surface of knowledge.
Speaker 2What does the managed inference market look like in five years?
Speaker 1The key aspect is that you want to provide an easy way for companies to access intelligence. It could be their own intelligence. It is hosting these custom bespoke models that are very domain specific for their specific use case. And it may be post-trained on a lot of their own private data. But the key aspect to managing inference is that you want to provide an easy way for companies to access intelligence. So the key aspect of managed inference is like it abstracts away a lot of the infrastructure complexity of managing the service across a lot of different GPUs, a lot of different data repositories, a lot of different models in terms of like how the queries get routed and which models they use.
Speaker 2I would love to do a quick fire round with you because there's many questions that I want to ask that I'm also aware that you have to get on with your life. Okay. So if we start with one, parenting. I heard from so many of your friends that you're a parent and you're a parent of a parent. You're an unbelievable dad as well. If you were to advise me on how do I crush it at work and be a great dad? Very important. What are the must do's? What are the must not do's? It's really about prioritizing it.
Speaker 1Spending time with my kids is like one of my greatest joys in life. I get it, dude, but you can't.
Speaker 2You're raising $3.9 billion.
Speaker 1It's such a cliche, but having kids is the best thing in the entire world. And I have a very busy schedule. I have a very challenging life. I have a lot of commitments at work. I'm on the road a lot, but it's about prioritizing things. Oftentimes what that means is I end up taking worse flights. I take a lot of red eyes just so I don't miss a bedtime with my kids. I'm home almost every weekend. Those weekends are very committed to coaching soccer or helping with swim or whatever activity that my kids are very passionate about. And then I think carving out very specific times that people at the company know I'm very unreachable. Right. So if I'm home in the Bay Area, everybody knows that I'm unreachable from 7 to 8 a.m. It's like I'm not taking any meetings at that point because I'm making my kids breakfast. I'm getting them ready for school, getting them pumped up for the day. And I'm just unreachable in days when
Speaker 2I'm home. What do you believe that many people disagree with you on and makes you unpopular?
Speaker 1I guess right now that data centers are a good asset for communities and they should be celebrated in the communities where we build them. And again, this comes down to all of these benefits that we feel like we can provide, whether it's economic, whether it's energy, whether it's the investments we're making in communities from a school perspective, from a- Do you worry about the income inequality and the wealth inequality that we're seeing? No, I don't worry about income inequality or wealth inequality because I think there's something inherently human about like, hey, I have this and that other person has that. And why don't I have that? There is a bit of like a jealous neighbor type thing. But I do think that it's natural when you have these huge things that are driving economic progress, where everybody's actually benefiting far more than if we didn't have this investment happening.
Speaker 2They are, but three IPOs are the cumulative value of 45 years of IPOs. The concentration of value is unprecedented.
Speaker 1Right. But think about all the value that's being created for the world. The market cap of the company is one thing. But actually, if you look at the value that's being created for the world, it's not just the value that's being created for the world. It's not just the value that's being created by abundant intelligence that is going to drive massive productivity gains for the global economy. It's many fold of the market cap that's being
Speaker 2created. What have you changed your mind on in the last 12 months?
Speaker 1I think maybe something that I've changed my mind on is what creates ingrained long-term competitive advantages. I think a lot of, we try to think of, VCs love to ask about what's remote, like what's your, you know, what's your long-term moat? What's going to create, you know, sustained above-limits? I think a lot of, you know, we try to think of VCs love to ask about what's remote, for you. I think what I've changed my mind on is like most moats are a illusion. Most of them are ephemeral. And especially during a moment of accelerating technological progress and capabilities of these models, I think a lot of it just comes down to being able to move quickly and adapt efficiently to the ever-changing chessboard that's in front of us.
Speaker 2I think everything starts off as a moatless wrapper and it's the value that you build on top of it. In segments. Yeah. Will Crusoe be public by the end of 2028?
Speaker 1I'm not sure. The company itself is, requires a lot of capital to build data centers, to build AI factories, to deploy large-scale clusters of GPUs, to build out that infrastructure layer of intelligence upon which, you know, so much global economic value is going to be created. That requires a lot of capital. Being public, there's a lot of advantages to being public and being able to access, you know, scaled capital resources. So we do think that ultimately the company is probably better off in the public markets. It's just a matter of like, when does that make sense for us?
Speaker 2Who do you not have on your board that you would love to have on your board?
Speaker 1Good question. I would love to have Sark. Yeah. I would say Michael Dell. He's not going to join my board. I just have tremendous admiration for Michael just as a human being. I think he's an incredible person. He's built an incredible business. He's very, he's got incredible business instincts. He's a great family man. I just really look up to him a lot.
Speaker 2I spoke to Zach before this.
Speaker 1Zach's an awesome guy.
Speaker 2You told me you're a douche. Oh, bugger. Sorry, I was anonymous. I didn't see that. That's so funny. Totally get that. Final one. What would the chase of 2018 find most unbelievable about the Crusoe of today?
Speaker 1I think the chase of 2018 would find the fact that this all worked just so unbelievable because it was big, bold, and ambitious, you know, to build out this AI cloud platform. The energy first narrative would come to fruition and energy truly would become the bottleneck to scaling the layer of intelligence. That it would have happened at this scale, I think it would have been almost unbelievable eight years ago. And when I think back to that era, you know, there were certain people that like gave me great advice, like during that period. And I had this, a dear friend that, you know, was an entrepreneur. And, you know, I think his company at the time was, was about 150 people. And, you know, the, when you're first starting a company, the idea of having 150 people working for you is like an almost unthinkable number of people. Like it's like a, how, how am I ever going to have 150 people work for me? And, you know, today Crusoe is approaching 2000 people, which, which is just like, you know, even when I say it, it sounds like an insane number of people that have bet these ultra productive years of their career on Crusoe. And, you know, they're, they're with us building in the trenches. Every single day. I just don't take it for granted, but people have invested their time with us. I think it's like such a important investment that these people have made in sort of making the bet that this is an important thing that I want to spend these years working on.
Speaker 2I've so loved having you on. As I said, it was such a joy stalking you and hearing the many stories that I did hear about you. So thank you so much for joining me. And this has been fantastic. Thank you so much for having me. It's great. But before we leave you today, founders and investors, sleep on eight sleep. It is the intelligent sleep system trusted by Mr. Mark Zuckerberg, Dr. Andrew Huberman, Charles Leclerc, and the Aston Martin F1 team. The pod is a smart mattress cover that fits on any bed. It cools each side to 55 degrees, heats to 110 degrees. It even tracks your heart rate, HRV and sleep stages without a wearable. Thank God. The new pod six has a hub 50% smaller than the last one and starts at just $1,999. Visit eightsleep.com slash 20VC. That's E-I-G-H-T-S-L-E-E-P.com slash 20VC. While eight sleep helps you sleep better, MongoDB helps you build better. Every serious AI company right now has the same problem. Agents need accurate context fast. Most infrastructure though, it wasn't built for it. Well, MongoDB is. 11 Labs runs 40 million agents on MongoDB. 75% of the agents are from MongoDB. And MongoDB is the only agency that the Fortune 100 run their most critical apps on MongoDB. Trillions of dollars moving through MongoDB every single day. Instead of stitching together 10 different tools, you get one JSON native database, vector search, and Voyage AI embeddings in the same system. And it runs anywhere. No lock-in. If you're building an AI, MongoDB for startups helps you move faster with Atlas and Voyage AI credits and technical support to help you ship. Build agents that answer once and forget. Don't build agents that answer once and that remember and learn your real-time data go to mongodb.com slash agents that's m-o-n-g-o-d-b dot com slash agents while mongodb scales the product framer scales the story if your team wants a website that looks and feels handcrafted but is still fast to ship framer is built for exactly that framer is the ai website builder that helps creators teams and businesses ship production ready sites faster than ever while getting every detail right prompt inspect edit and publish in one place at a whole new pace agents and humans work in tandem agents bring speed and scale you bring taste judgment and control the work lands on the canvas and stays editable build custom code components manage cms content optimize seo and audit for issues all in one place enterprise grade hosting security and 99.99 uptime slas trusted by leading brands like perplexity and miro learn how you can get more out of your site from a framer specialist or get started building for free today at framer.com slash 20 vc for 30 off a framer pro annual plan that's framer.com slash 20 vc rules and restrictions may apply
PodcastTranscript AI