← Inevitable
Inevitable
Crusoe CEO and Co-founder, Chase Lochmiller: Live Special at MCJ Summit
00:40:04
Chase Lochmiller is the CEO and co-founder of Crusoe. If you’re a regular listener, Crusoe isn’t new to the pod. This summer, Cody sat down with Chase’s Co-founder and COO, Cully Cavness, during our live event in Austin. This latest episode was recorded live at the inaugural MCJ Summit in San Francisco at the beautiful Autodesk Gallery. Cody and Chase dive into how Crusoe is building data centers at the intersection of AI and energy. Chase traces his path from MIT soccer captain and mountaineer to climate-focused entrepreneur, and how those experiences shaped Crusoe’s core values of preparation, curiosity, and speed. He shares the story behind the company’s 1.2-gigawatt Abilene, TX project, its energy-first approach to powering AI infrastructure, and his vision for an era of abundant energy and intelligence. The discussion also explores the future of AI labor, grid integration, and what digital abundance could mean for society at large. Special thanks to our MCJ Summit attendees...
Loading summary
Transcript88 lines
- [00:01]
A
From McJ. I'm Cody Sims, and this is inevitable. Climate change is inevitable. It's already here. But so are the solutions shaping our future. Join us every week to learn from experts and entrepreneurs about the transition of energy and industry. Today on Inevitable, we are live.
- [00:33]
B
At.
- [00:33]
A
The inaugural MCJ Summit at the beautiful Autodesk Gallery in downtown San Francisco. And our guest is Chase Lockmiller, CEO and co founder of Crusoe. Crusoe is a vertically integrated clean AI infrastructure company, and we have a ton to talk about. Chase, welcome.
- [00:50]
B
Thank you.
- [00:51]
A
So I wanted to start by really understanding you, Chase, and digging into how you ultimately built this business at Crusoe, but going in the wayback machine and maybe sharing a little bit about where you grew up and what kinds of things you were interested in as a kid. Because AI infrastructure businesses were not a thing when you were a kid. It's a new category.
- [01:14]
B
That's right. You know, data centers were scarcely a thing, honestly, when I was growing up. But, you know, I grew up in Denver. Denver, Colorado. Coley, my co founder, and I actually went to high school together in Denver at a school called Kent, Denver. I think growing up, I was always very interested in math and science. I was, like, very drawn to math and math competitions. I was very competitive kid. I played a lot of sports, loved soccer. We got the World cup coming here next year. Have very fond memories of the 94 World cup and just following that closely. And that helped inspire me to just fall in love with the game of soccer. Ended up playing soccer through college, played at mit and actually there's a number of. There's people from every chapter of my life that have been part of Crusoe. And there's a number of kids that were on the MIT soccer team that actually work at Crusoe, including our cto, Nithin, I think. Oh, no way.
- [01:58]
A
I didn't know that.
- [01:59]
B
That's neat. Nathan was our sweeper. He's a great.
- [02:01]
A
What was your role?
- [02:02]
B
I was a center midfielder, of course. You were CEO of the soccer team. I originally went to mit. I was convinced I was going to become a theoretical physicist and help unlock the mysteries of the universe and expand the human knowledge set. I spent a lot of time doing various research things and did some research at mit, did some research at Los Alamos National Lab. And I think one of my realizations during that era was that basic science research just moves pretty slowly and I just had a ton of energy. Maybe today that's changing with AI, but, like, certainly at that point, it just Felt like a very slow, drawn out path. I felt like I wanted to do something faster paced.
- [02:42]
A
Did you think originally you wanted to be in academia?
- [02:45]
B
Was that the path you were on? I think that was the path. Academia or a lab or something. Exactly. I mean, you know, ITER had just been announced so, you know, this big international collaboration to build this giant tokamak to, you know, hopefully have a big breakthrough in fusion. You know, I remember kind of my first day working at Los Alamos National Lab and my principal investigator I was working under, he told me, chase, this is really exciting time for fusion and plasma physics and hopefully in 30 years we're actually going to have fusion. And I was like, wow, that's pretty cool. That's like, cool. And he's like, yeah. You know, they told me that When I started 30 years ago, here we are and you know, we're still 30 years away. That was always the joke in fusion was that it was 30 years.
- [03:24]
A
Hopefully we're soon. Hopefully we're.
- [03:25]
B
Feels a lot cooler today, right? Exactly, exactly. When I realized I wanted to do something faster paced, I felt like in certain ways I had lost a purpose. I was like, oh man, what am I gonna do? I had my whole life set on being a theoretical physicist.
- [03:40]
A
Or was entrepreneurship a thing that was around you as a kid? Was there any influence there for you?
- [03:46]
B
I mean, my dad was kind of an entrepreneur. Like in certain ways.
- [03:49]
A
I rode the bus with your dad going to the Redwood site in Reno and it was amazing. He's a character.
- [03:56]
B
He's a total character. He's a one of a kind. He's certainly had a big influence in my life. Entrepreneur in a very different sense. And what's funny, actually, I think there's this aspect where sometimes if you have a parent that has a lot of success in one vertical, you tend to be like, look, I want to make my own path, my own sense of success and something completely different. And I felt like I was doing that for many years and building things in technology and I've worked in quant finance for a long time and just doing things that my dad never would have done. Like, it was just like never. His career path. He's been building in real estate. He's a multifamily and commercial real estate landlord, owner, operator. The funny part is this whole big boom in AI and all these big AI data centers that are being built. It feels like I'm coming full circle where it's like, okay, all of these things that I've learned in technology that have inspired how we're building out all of this AI infrastructure to power intelligence in the future economy. And yet I own a building and I'm a landlord, just like my dad.
- [04:54]
A
Just like your dad. It's amazing. And you personally have done some pretty crazy stuff too, like back to your Colorado roots, your mountain climbing adventures. Maybe share a little bit about that and what inspired you there.
- [05:05]
B
Sure. In middle school, I had this teacher and her husband was this blind guy named Eric Weinmayer. He was attempting to climb the Seven Summits, the highest peak on each continent. And he was blind. And at the time, you know, I was in seventh grade and he was training to climb Mount Everest. David Brashear's Everest documentary had come out Into Thin Air. The Jon Krakauer book had come out just about the account of the disaster that had happened. And I was like, man, that's so cool. It's the first time I'd heard of this concept of the seven summits. You know, we have a lot of 14,000 foot peaks in Colorado. It's like a very Colorado thing to talk about climbing 14ers. But I think I was just kind of inspired to like climb big mountains and, you know, spend a lot of time outside, and grew up going skiing a lot and spending a lot of time outdoors. And so after I graduated college and I started, you know, sort of stopped playing sports very competitively, I was like, all right, what else can I do? And I became sort of a alpine mountaineering kind of hobbyist, for lack of a better word. Right. And so I did a bunch of different expeditions sort of all over the world at this point. You know, I've climbed five of the seven summits, including Mount Everest. I had a couple different expeditions to Mount Everest. You know, once in 2014, and there was a very large avalanche that happened that year in the Khumbu ice fall. Tragically, 16 Sherpas were killed in that accident. But for some reason I felt compelled to go back and ended up summiting in 2018. So had a very successful summit that I learned a lot from sort of the first time. Maybe eventually I'll get to the final two summits, but it's just kind of.
- [06:32]
A
You're a little busy right now, then deprioritize any of them. That really stood out as the hardest challenge of any of the climbs you did.
- [06:39]
B
You know, Everest was like this challenge because it was this crazy expedition and there was so much preparation that went into it and there's so much emotion as a result of just like shifts and things and, you know, it's just kind of like the world is so focused on what's happening in Everest. There's, like, media coverage for, like, every little thing that's going on in Mount Everest, every climbing season. So I think kind of coming out of that, we actually, you know, when we started Crusoe, one of our core company values that we created is to think like a mountaineer. Coley spent a lot of time mountaineering as well. We actually were climbing a mountain when we sort of were ideating on the origins of Crusoe. The sense there is really about trying to have a lot of the same principles that you have in mountaineering applied to your life and your work at Crusoe. And it's not like everybody has to go climb mountains and go be a big alpine enthusiast. It's more about the sense of incredible amount of preparation, like, thinking through, okay, here's my plan in terms of how I'm going to get to the summit, and then I'm going to get down safely. And here are the different things that could go wrong. And if those things go wrong, here's my plan for them. Here's my plan B. Here's my plan C. Here's what I'm going to do if the weather changes. Here's what I'm going to do if I have this piece of gear fill. Here's what I'm going to do if my partner gets sick. Here's what I'm going to do if.
- [07:49]
A
Hope for the best, plan for the worst.
- [07:50]
B
Exactly, exactly. And, you know, there's this notion of mountaineering that getting up is optional, getting down is mandatory. And, you know, we actually try to instill a lot of the core safety practices in mountaineering. Just really thinking through safety is a critical aspect. And in our business where we're dealing with both hardware and software, you know, there's millions of man hours that have been worked in Abilene. We have to be thoughtful about sort of safety culture so that we can build these things at scale, at speed. But also, I was going to say.
- [08:17]
A
How do you do it and still be the fast cowboys of the space?
- [08:20]
B
There's an element of mountaineering is actually an interesting thing to study. You have kind of like these, like, legacy styles of mountaineering that are like these big expeditions where you bring, like, everything. You almost, like, build a whole city, and, like, you sort of, like, are very, very intensely prepared. And over time, what people found to be, like, the most successful ways of, like, climbing big peaks and sort of breaking records is this Alpine style of light and fast, right? Where you're basically moving quickly, you have everything you need and nothing more. And you have this robustness in your planning and you're thinking about things safely. But oftentimes speed is actually a mechanism to be safe in mountains, especially when you're sort of.
- [08:58]
A
What is it? Fast is slow. Slow is fast, right? Is that a yeah?
- [09:01]
B
Yeah, maybe.
- [09:02]
A
Let's fast forward to today and where Crusoe is. Crusoe's had, from where I sit, one of the most unbelievable evolution stories. And for folks who really want to dive into that, we recorded an episode with Chase's co founder Cully a couple months ago in Austin. And we really dive into the path that the company has taken to get to where you are today. But with you, I really want to focus on where are you today and where are we going. So maybe describe what Crusoe is now and the business that you are running today.
- [09:32]
B
So Crusoe is a vertically integrated AI infrastructure business. Our goal is to help make energy and intelligence more abundant. Just really accelerate this abundance of both energy and intelligence, which we think are going to be critical to uplifting humanity in the future and driving progress for everybody around the world. Now, how we got here is pretty interesting. I think we have a culture of curiosity. We just sort of are constantly asking questions about how we would do X.
- [10:00]
A
And just quickly sense of scale, right? What 800ish employees.
- [10:03]
B
We actually just celebrated surpassing 1000 employees.
- [10:06]
A
A thousand employees.
- [10:07]
B
Amazing thousand employees. And then we have numerous sites where we have more than 1000 contractors. In Abilene, we have 7000 contractors that are working there every day. 7000 people on site every day, but 1000 full time employees across a couple of offices.
- [10:23]
A
So you're saying this culture of curiosity.
- [10:25]
B
We'Ve taken this very vertically integrated approach to AI infrastructure. And so what that means is we're doing the hardware aspects of building the infrastructure needed to power AI, needed to run these large scale intelligent workloads. So there's everything from land and power development, data center design, construction, cooling, mechanical, electrical, like the whole stack of the physical infrastructure to stand up AI workloads. And then we're also in the business of deploying these large clusters of GPUs. This is our Crusoe cloud business where we build, operate and manage these large scale clusters of GPUs and integrate them with critical services like storage and high performance networking. And then we also have sort of these managed services to extract intelligence from that infrastructure. So things like managed inference product, managed autoclusters product things that basically help people create intelligent results from these massive investments that they're making in infrastructure. We cover a lot of stuff from like hardware to software. I like to talk about our businesses creating these AI factories, these factories that produce intelligence, and doing the hardware of building the AI factories and then actually the software of actually how you utilize the AI factory to produce intelligence. And the reason we believe that it's the right moment to build a vertically integrated business is what we're seeing unfold is actually this massive new category of infrastructure. And whenever you have kind of this paradigm shift, being able to move quickly, being vertically integrated enables you to move quickly into sort of a new vertical. And this infrastructure for AI, you know, what I call like the infrastructure of intelligence, it's dramatically different than anything that we've ever seen before. It's different than traditional data center infrastructure that serves the Internet and traditional cloud computing. It's different than traditional power systems that have served us scaling to date. And a lot of that stems from the actual workload, the design of the systems, ranging from the way the clusters are built out, the way they're utilized, and the way kind of users interface with the services, it just merits a completely different design. And it's at a massive scale. So it sort of warrants this investment in vertical integration.
- [12:34]
A
Can you describe a little bit, just for folks who maybe aren't as familiar with AI infrastructure, the difference between training workloads and inference workloads, and are those setups substantially different for you both on the infrastructure side and on the cloud business side?
- [12:48]
B
We are seeing a convergence and I'll get to that in a second. So training is basically where you're setting up the model that you're going to be utilizing. It's basically like you're making the investment and you're amassing all of this data, all this training data, this historical data into your data center. And then you're setting up this large scale model. It's going to be like tens of billions, hundreds of billions, trillion parameter, large language model or other foundational model depending on your application. And then the training process is basically trying to fit a large scale nonlinear statistical model to that data so that it would be able to, with new inputs, be able to produce results that are intelligent, so to speak, and sort of learn from that data.
- [13:29]
A
And then most of the big giant projects we're seeing today are building these training centers for the most part.
- [13:35]
B
And this is where I was saying it's changing and it's converging. So there's this notion of pre training which is basically trying to create structure from all of this unstructured data. And that's like what's historically been producing these original GPT models. So GPT1, GPT2, GPT3 where we're seeing massive compute scaling is actually in the post training and sort of test time compute scaling where you're actually having, you have this foundational model that has been pre trained and then when you're serving a workload, OpenAI sort of released this 01 model about a year ago and there's since been a lot of innovation on these different chain of thought reasoning models that actually produce results and then take those results and feed them back into the model. And what it ends up doing is it's a bit like training on new data and it's sort of training at inference time. You know, it's called test time compute scaling and it's where you're basically running a lot of passes through your neural network when you're actually trying to produce an inference result.
- [14:34]
A
Could that in theory start happening at the edge more then it can, but.
- [14:39]
B
It also needs a ton of compute. And the benefit of doing it at the edge is you actually are, you have lower latency to like the end consumer. If you're actually doing a significant amount of test time compute scaling, you're actually thinking about this for a long period of time. So if you're thinking about it for even a second, oftentimes you're thinking about it for many seconds or minutes or if you've tried like deep research or some of these other things that they just require a lot of time to produce a result because they're thinking, they're thinking about what the result's going to be.
- [15:08]
A
So they're training themselves in real time as they're inferring them.
- [15:11]
B
Exactly. But if you're at the edge it actually doesn't matter, it doesn't benefit you. So you might as well be in these big centralized locations. Now that's for like one school of, not one school but you know, one category of different results. Right. There are definitely things that can be done very quickly at the edge. And we sort of believe in sort of this bimodal distribution of infrastructure that's going to require both these like massive scale centralized facilities like, you know, what we're doing in Abilene and some of these other campuses that we're building as well as widely dispersed inference infrastructure everywhere in the world.
- [15:45]
A
I think of inference as basically being the edge serving layer of the queries Is that a very basic way of thinking about it?
- [15:52]
B
You know, I think like one way of thinking about a large language model is it's kind of like a database. It's like a statistical database in certain ways where it's like instead of looking up, you know, a row and a column in a database, you're actually looking up all of the information would map to in a row that may not exist. And so you're sort of statistically inferring in that database. So in a lot of ways inference is like looking up something in a database, but it's a statistical database.
- [16:16]
A
So maybe just for clarifying what you have today, describe the Lancium project now and where it sits and ultimately what that is going to be.
- [16:25]
B
We've always sort of taken this energy first approach to developing compute infrastructure from when we had a bitcoin mining business to, to the early days of scaling our AI infrastructure business to today where we're building and operating gigawatts of capacity to support these large scale global workloads. A bit of the history of Lancium is that I've known the CEO of Lancium, this guy, Michael McNamara for five, six years. He initially started building this project as a bitcoin mining site. And it's funny to track the arc of bitcoin mining infrastructure and compare that to the arc of AI infrastructure because I think they're following kind of similar paths. They're going to end up in different places, but it is an interesting comparison to make. But bitcoin is like, you know, has been notoriously famous for consuming a tremendous amount of energy, which led a lot of bitcoin miners to going to areas where they could access abundant low cost energy. So Michael had started working through the Abilene site partially because there was an abundant amount of energy there. And when he started telling me about it, it was something that had always been on my radar as like very interesting market to pursue to access, you know, large scale clean energy solutions. And what had happened there is on the back of production tax credits, you know, a lot of wind developers had built out these large scale wind projects in that region. It's a naturally very windy region. So you get very, very high capacity factors. For a lot of the wind farms that are in that region now, there didn't exist the transmission infrastructure or the load to basically consume all of the power gen that's been installed there. And so what's resulted is power prices are frequently negatively priced. There's just too much power being generated and people are Basically getting these incentives, these production tax credits to the point where they're willing to sell power at a negative price because the production tax credit is bigger than the negative price they're paying for power. Once the production tax credit incentives roll off after 10 years, they're merchant and they're in a position where they're actually having to curtail. So they're basically, they could be producing power, but they're not because there's no marginal buyer for the power.
- [18:24]
A
So when you see the windmills out in the field just standing.
- [18:26]
B
Yeah. And you're like, wow, it's windy today. Why isn't this thing spinning? It's like it's because it's.
- [18:30]
A
There's no money on the other end.
- [18:32]
B
Exactly. And this seemed like an incredible opportunity for us. The early, early end of this like energy hungry demand coming from AI. And due to the fact that AI, even on the inference side is more tolerant of latency than traditional web applications, we could really build it anywhere. That led us to Abilene as this great market that had low cost, abundant clean energy. And today we're building a 1.2 gigawatt campus that's supporting some of the largest, most important workloads in the world for Oracle and OpenAI. And you know, we've been energizing that incrementally and been standing it up at sort of a record pace. We first broke ground on the facility in June of 2024. Quick story, there was originally like an RFP that went out for the site and the fastest anybody had committed to build 100 megawatts of capacity was two and a half years. I was talking to someone and they were like, could you do this in like 12 months? And I was like, yeah, for sure could definitely do that. You know, having done no research on it, but was thinking about it with my team and you know, we brought together.
- [19:33]
A
Going up is optional, going down is required. Right there. Requirements go.
- [19:38]
B
But you know, we ended up delivering the first 200 megawatt buildings in 11 months. And this comes back to the sense of curiosity that we built at Crusoe where you know, we just challenged every single way of traditionally building data centers and said, wait, why do we have to do that? Why is it designed this way? Why aren't we designing it this way? It's a giant cluster of GPUs, it should be designed this way. You know, why do we have to power it that way? Why can't we energize it this way? And you know, just asking those questions of why, why, why, why, why? And then when you get to the answer, it's like, oh well that's sort of the way the industry's always done it. You're like, okay, well that's stupid, let's do it this other way. This other way makes more sense. And we were able to save an incredible amount of time and really able to accelerate this infrastructure at a record pace in Abilene.
- [20:17]
A
Not to not dwell on the size and scale and amazingness of that particular project because it's been relatively all consuming for you all for a while now. But what's next? Seems like there are many more of these being contemplated or signed with ink and starting to be developed.
- [20:33]
B
You know, we have a couple of other projects that were under development, some that we can talk about more publicly, others that we're sort of just less public about. They all have a similar philosophy though, which is taking this energy first approach to developing the compute infrastructure, going to places that we know we can access this low cost, abundant energy. So there's a couple of projects we're working directly with ipps where we're actually a behind the meter offtake for them. You know, in the case of like a wind farm, where we actually have a behind the meter power purchasing agreement with the on site wind generation, we'll actually use the same interconnection, the same substation that they're using to sell power into the grid. We use that to buy power from them as well as buy power from the grid when the wind's not blowing and the sun's not shining. Our perspective in all this is that, you know, if you look at energy infrastructure today, if you look at data center infrastructure today, they're largely saturated in terms of like production and demand. And if you look at the ambitions of AI and you look at the ambitions of what people want to build and how they want to scale it, how they want to utilize it, it's going to require just like all net new infrastructure. And as a result that requires us to think about how do we actually bring online the new power infrastructure both on the generation, storage, storage, distribution side and then how do we bring online the net new data center infrastructure to support these mega clusters of advanced AI accelerators and GPUs. You know, we have a campus where we're planning to build wind, solar batteries, gas and have a grid interconnection. And all of those things combined are actually more generation capacity than what we're using in the data center on a day to day basis. But we have to sort of engineer it for overall peak demand. This is where I get back to to this notion of like creating this abundance of energy. We can create an abundance of energy by virtue of the fact that we need to like almost oversize our infrastructure demand and that extra can lead to lower cost power for everyone.
- [22:27]
C
Hey everyone, I'm Yin, a partner at mcj here to take a quick minute to tell you about the MCJ Collective membership. Globally, startups are rewriting industries to be cleaner, more profitable and more secure. And at MCJ we recognize that a rapidly changing business landscape requires a workforce that can adapt. MCJ Collective is a vetted member network for tech and industry leaders who are building, working for or advising on solutions that can address the transition of energy and industry. MCJ Collective connects members with one another with MCJ's portfolio and our broader network. We do this through a powerful member hub, timely introductions, curated events and a unique talent matchmaking system and opportunities to learn from peers and podcast guests. We started in 2019 and have grown thousands of members globally. If you want to learn more, head over to MCJ VC and click the membership tab at the top. Thanks and enjoy the rest of the show.
- [23:28]
A
Do you see a world where, where you know, you were talking about behind the meter and off grid where the data center power demand in this world maybe decades from now becomes the primary power consumption source in whether it's the United States or in the world. And you move from a world where utilities are the gatekeepers of power to the data centers actually being the drivers of power production and distributing it.
- [23:53]
B
Yeah, I mean the silver lining in all this is the people making these investments are the companies with the biggest balance sheets and the best balance sheets and greatest cash flows in the history of business. They're very well capitalized to make these investments. You know, I think you're already seeing it today. I think Northern Virginia, I think data centers, I don't know, I heard some statistic, you'd have to fact check me on this, but I think it's like 40 some percent of power in that region is being consumed by data centers. Not unreasonable to think that will manifest in many other markets where data centers are going to be built to support large scale intelligent workloads. And I think it's very natural.
- [24:27]
A
I mean does it become at the beginning a shadow grid almost of this off grid power sources, but eventually becomes the feeder to how everybody else accesses power? You think it's that macro?
- [24:38]
B
Maybe there's one way of thinking about it, like data centers can be co located with large generation resources and be kind of like at the, I don't know the origin of the power production, right. It's like in order to bring all this infrastructure online, we need more power. So we're just going to do it and then like whatever excess is available can be, you know, I think everybody's commercial, right? It's like how are we going to monetize that? How are we going to bring down the overall cost for the data center infrastructure? We're going to sell it in some downstream capacity, we're going to get it to some other market where we can do something useful with it. So I think that's like definitely like a trickle down effect of these huge investments that are being made in the infrastructure. One other thing I would just point out is that because of the scale of these projects, both in terms of megawatts gigawatts as well as capital investment and the types of capital being invested, I think it is one of the greatest opportunities ever for advancing next gen energy solutions, climate technologies, new battery solutions, new energy production solutions. Another project I'll talk about is in Wyoming. We've tried to think very first, principled and thoughtfully about our energy production, but we're looking at pioneering what would be, I think the largest post combustion carbon capture and sequestration system in the world. It's a carbon hub that's been developed by our partner Tallgrass. We're planning on bringing a lot of gas infrastructure online to support initially a 1.8 gigawatt facility for AI workloads. But one of the benefits is there are four Class 6 wells that have already been permitted. There's existing carbon infrastructure where you can actually capture the post combustion CO2 and permanently sequester it underground. Benefit from the incentives in 45Q. But I think a lot of these things have been discussed or done at sort of a pilot scale, but never at a massive scale. And part of it is because there hasn't necessarily been the capital invested in terms of making a reality what AI infrastructure.
- [26:24]
A
And it maybe hasn't penciled for power that's just flowing into the grid. Right. But you have potentially energy buyers who are willing to pay a different price for fast access to power and fax access to clean power.
- [26:37]
B
Exactly. And I think what a number of these customers, they're looking at this saying well okay, if I produce power from gas, that's great, I can get my power quickly. Our sustainability goals are over here and if we do that, we're going to increase our carbon emissions, we're going to have to go buy carbon removal credits. What's the price of those? Okay, that's pretty expensive to buy that amount of carbon removal credits. What if we just paid for post combustion carbon capture and CP sequestration and got the 45Q credits? Is that cheaper? And I think in a lot of cases the answer is yes.
- [27:06]
A
Talking about the speed at which all of this needs to get built, there are a lot of companies, not just infrastructure companies, but the biggest AI companies in the world, who have now raised so much money at these skyscraper valuations. And there's going to be pressure for them to start delivering from a revenue perspective, both on their own valuations, but also their old ultimately the patron of these data centers that you're building out. And so to some extent, the economics there need to start penciling for them too, in terms of the cost and the amount of infrastructure they're building out to support their businesses. If that doesn't happen or happens more slowly than they're expecting or their investors are expecting, do things start to cascade quickly in terms of the public markets?
- [27:49]
B
Definitely a possibility. I don't want to rule anything out, but I think a lot of the big AI labs are seeing just incredible revenue growth and adoption in these services that are being built out. Kind of the optimist in me believes that we are going to hit these growth trajectories because the models are getting so good. If anybody hasn't played with them, I encourage them to use a lot of these new services that are being built. I mean, they're incredible what they can accomplish. And if you compare it to what you would be paying a software engineer or something like a call center, what you're spending on a lot of these human labor aspects, it's a massive cost saver. So that's like the optimist lens in all this. If people don't hit their growth targets on revenue and how are we going to pay for all this infrastructure? In any major bubble, there's always kind of like what leads to the unwind or what leads to the crash. Oftentimes it is like unresponsible amounts of leverage. And we are seeing kind of signs of some of that today. It's not like everywhere, and not all leverage is bad, but we are seeing a lot of debt capital pouring into the space because how else are you going to pay for a lot of this when you're talking about spending hundreds of billions or trillions of dollars to build out this infrastructure? This is something that we are very focused on at Crusoe. In terms of understanding counterparty risk and understanding what are the real risks we're taking and real downside scenario, we try to think through really bad things happening. That's part of thinking like a mountaineer, thinking about the robustness in the business. I think a lot of the bigger higher leverage situations that we're undertaking, we typically have investment grade counterparty risk for very long duration. Even in an economic downturn or even if AI isn't turn out to be everything that's cracked up to be. Is Oracle still going to be around? I believe they will be. The financial markets still believe they will be. They have a great enterprise software business and database business like still a great moneymaker for them. Same could be said for Microsoft, for Amazon, for Google, for Meta, for all these companies that are still producing incredible free cash flows independent of anything that's happening in AI. So you know, we try to underwrite the risk through that lens, shifting a.
- [29:54]
A
Little bit into where you see AI going. Let's start with the energy markets and then go bigger than that. How far away are we from AI as an actual power orchestrator in terms of controlling the. The knowing when to buy curtailed wind and powering a data center with it, or not knowing if a data center doesn't have to have 99.999% uptime when it shouldn't need to be up. And what you can do with that excess power in terms of sending it back to the grid, for example. How far away is that from reality?
- [30:24]
B
Depending on who you ask, both closer and further than one might think. The tricky part in adoption here is trying to get all of the players to coordinate on solutions. We sort of know the answer to a lot of these solutions. I think we have the technology to do it. I mean there's great things happening across the space. You look at companies like Emerald AI, that's like trying to build a demand response essentially for large scale datacenter loads and working with the utilities to help curtail during some of the most challenging peak hours. You look at the infrastructure and how we build smart infrastructure that can help with the transmission and distribution of power. I look at a company like Heron Power started by Drew Baglino. That's like reinventing a lot of the, the critical electrical stack from high voltage to low voltage and doing that with power electronics. There's a lot of stagnation in the space where like not much innovation has been done for a century plus. You know, there's a lot of interesting new ways that we can use power electronics to actually do this smarter, faster, cheaper. I think we like have the tools and I think also like, you know, thinking through things like two way power systems where people have generation, people have load and people have storage in their home. You have cars driving around everywhere that have giant batteries on them. They can be these giant sources for power generation in moments of peak demand. But that requires this massive orchestration effort and requires buy in from utilities and a lot of the other counterparties throughout the entire ecosystem to overall make it happen. So I think that's the bottleneck to adoption is really like getting utility and regulatory buy in. Not like do we.
- [31:58]
A
So it sounds like it's much more.
- [32:00]
B
Do we know how to do it?
- [32:01]
A
Yeah, and it sounds like it's much likely to be more of a gradual move toward it until we're ready for all at once as opposed to oh, the software's ready. So it's an all at once switch.
- [32:10]
B
But I mean you're seeing like signs of progress here, right? I mean like, you know, looking at things like VPPA adoption with like Tesla powerwalls, like, you know, they've built out this whole infrastructure of you know, distributed power walls in people's homes. And I mean it's been helpful in terms of serving peak events here in California. You know, you wouldn't think of California when you think like fast nimble government agencies, but you know, you're seeing similar things that base power's doing in Texas. Like love all the work that they're doing in terms of trying to create both distributed storage, distributed generation, distributed mechanisms to basically support grid resilience and utility.
- [32:44]
A
If we take a big step back and everything that Crusoe is building and helping to enable, this is more than just a technology revolution. This is ultimately societal, cultural. As we think about a world where we actually are moving toward artificial general intelligence or the like, your business is very much at the forefront of enabling that to happen. And so I'm sure you've thought a lot about what does the world look like in 20 years, 30 years? What does the job market look like? What does misinformation, trust, media look like? What kind of life are kids going to have in 20 years? How are you thinking about the future of AI, both the pros and the potential negatives of it as it starts to become pervasive in our lives.
- [33:31]
B
Maybe I'll talk about my perspective on a lot of the growth and impact of AI. And then I'm a dad too, so I have three young kids and how I'm thinking about like setting them up for success for this, like, next generation chapter of abundant intelligence. I think number one is people call things like AGI, asi, like Artificial Superintelligence. There's all these acronyms. And ultimately, the way I view it is like, okay, can we do something with a computer that exceeds or has some cost optimization compared to what we would do with a human? And in a lot of ways, I think about that as like, digital labor. And so the key thing that I sort of think about is this boom going to have real legs and sort of transforming human prosperity. And the human experience is going to be good for people, is really around. Is this going to lead to significant GDP growth per capita? That's the main thing that I think about. And if you go back to your macroeconomic 101 class, the growth in GDP, there's three main things that lead to the growth in GDP. It's change in capital, like in the economy, like all sorts of capital, change in labor in the economy, and then change in technology. Those are like the three main things that cause increases in gdp. You know, historically, technology has been like a hard thing to measure. So it's kind of like you measure these other two things and then you have this like, technological observable, unobservable. You sort of measure it as the difference. But what's fascinating about this boom is that with AI, it's created a new way for us to create more labor. That change in labor can be way bigger than it's ever been in the history of humanity because we can make digital labor, we make labor in silicon, and that's going to lead to incredible outcomes. And so my belief is that this is going to lead to an era of increased leisure time for people. People are going to be able to do far more with less than they've ever been stretched to do. And for those that are incredibly high agency, people are going to be able to do more, more with their fingertips. You're gonna have access to a workforce of thousands of people if you have the agency to just go engage with it and be curious about how you actually use this technology to, you know, will your dreams into existence.
- [35:40]
A
How does it impact those who don't have as much access to it? Do you think they get more left behind in that economy?
- [35:45]
B
I think it's like the greatest equalizer, right? It's like, how easy is it to access, you know, if you have a phone, if you have an access to an Internet connection, you know, the cost of.
- [35:54]
A
I guess, just like anybody can become a star on YouTube today from anywhere.
- [35:57]
B
In the world, the cost of intelligence is going to converge to the cost of energy, right? It's like the cost of compute is going to come down over the course of time and you know that compute is the equivalent of intelligence. And people are going to be able to access these things very freely, you know, not just in developing world economies, but, you know, all over the world. I think having that access at your fingertips, certainly there's going to be places that are going to adopt it faster, but I think it's like the greatest equalizer ever in the history. Like you don't need to have an incredible teacher that happens to be born in your town to educate you on some specific topic or inspire you to do something. It's like literally the AI tutoring can be unbelievably effective. I mean, you're seeing this with.
- [36:40]
A
I mean, it also means anyone in the world could become good at driving deep fakes and videos and propaganda and like controlling the narrative too, right? And so do you think an infrastructure company like Crusoe should have some say in what are the guardrails we put around AI?
- [36:56]
B
Like, how does that. We don't really view our role as being like the hall monitors of what should happen in the space. We really view it as like more intelligence in the hands of more people. We think it's just going to benefit humanity. And we're really just kind of like we rooting for the whole space, right? We're not trying to say like, you should do this, you shouldn't do this. You know, I think most of the applications that we're seeing are like just generally positive things for people. I know there's been some criticisms around like, you know, some of the AI. The AI generated video content. It's like, okay, this seems crazy.
- [37:26]
A
It's insane. Has anyone played with Sora 2? It's insane. It's amazing.
- [37:31]
B
It is amazing. You know, from my perspective, I think it's going to completely transform storytelling, right? In terms of like, you know, what it would take to build a feature length film just from a budget standpoint, from a creativity standpoint, you're going to end up with more ideas out there that get built because the barrier to entry for those big ideas is much smaller. I'm very optimistic about the overall video generation content platform. My own kids, I'm really trying to hone in on developing a sense of high agency and high autonomy, high curiosity. We send our kids to a Montessori school, so it's like very much like creating this sense of independent thought. And the reason for that is I sort of talked about this earlier just in terms of in a future where every person has access to 100,000 person workforce at their fingertips. That's great. If you have the agency to.
- [38:22]
A
Can we all be generalists? Like can the world work if we're all generalists?
- [38:25]
B
Do you think you're a general of your own domain? And I think like there's going to be different layers to this and different ways that humans are curious in terms of using that to their benefit. And you know, it creates unbelievable opportunity for people to build incredible things that benefit others.
- [38:41]
A
I think we're probably wrapping up here. I guess the last question I would ask is how do you want Crusoe to be remembered 30 years from now? You are absolutely at the core of building out this wave of massive change that is hitting humanity right now. You know, whether humanity wants it or not, it is inevitable. Right. And how do you want your company and yourself to be remembered as part of being there for it?
- [39:05]
B
What I'd like Crusoe to be remembered for is really being one of the platforms that help orchestrate intelligence for the economy and really sort of standing up, you know, the infrastructure, the energy solutions, the data center solutions, the computing solutions that really enabled intelligence to scale.
- [39:23]
A
Chase. Thank you so much. Yeah, amazing.
- [39:24]
B
Thank you. Appreciate it.
- [39:31]
A
Inevitable is an MCJ podcast. At mcj, we back founders driving the transition of energy and industry and solving the inevitable impacts of climate change. If you'd like to learn more more about mcj, visit us at MCJ VC and subscribe to our weekly newsletter at newsletter MCJ vc. Thanks and see you next episode.