Bitcoin, The Space Economy & Orbital Infrastructure w/ Dhruv Bansal & Aaron Burnett | Bitcoin 2026

All right. Thank you for coming, everybody. Especially those joining us from Orbit. My name is Dhruv. I'm the co-founder and CSO at Unchained. I have a minor reputation in the Bitcoin space for talking about Bitcoin in space. I think those who have read my work will be happy to know that we will be staying far away from some of the crazier speculations that I've written about. Um we'll be a little bit closer to Earth in in this session. I'm actually very excited because though I speculate about space, Aaron is actually an expert in the space industry. And so this is just going to be an opportunity for me to ask him a lot of questions that I'm curious about, but I hope many others in the audience will also be curious about. Aaron, introduction? Yeah, thanks for having me, Dhruv. I'm really excited to be here. I'm not sure I can think of myself as an expert as much as space can have experts, but really excited and happy to talk a little bit more about that. Okay, great. Well, I mean let's just get into it. Um I think the number one thing that I've always been curious about that a lot of people are probably aware of is space is far away. Or rather it's close, but you have to be very fast to to be in space. Launch costs are a huge aspect of the space industry and the space economy overall. Can you give us like a really quick historical overview of how launch costs have changed over the last, you know, 50 years that we've been doing space stuff? And how you maybe anticipate that changing in the future? Yeah, so launch cost is really the story of space for probably the last decade or so. And so what you had was in the era of shuttle roughly $50,000 per kilogram, which doesn't mean anything for people without context. But today that's about 50x less. It's about $1,000 per kilogram. And then what's And that's all driven by SpaceX almost exclusively by SpaceX. There are some others that would probably be mad that I didn't give them a shout out, but it's mainly driven by SpaceX and I think people would appreciate that. Um and then actually what's happening over the next little bit, let's call it 3 to 5 years, is a drop from 1,000 down to about $100 per kilogram. So another order of magnitude reduction. It's opening up all sorts of opportunities for commercial markets that just didn't exist. And probably one of the core reasons why SpaceX is looking at this IPO because things are really starting to take shape on the commercial side of things for them. Okay, interesting. So And so just kind of repeat that, it's incredible compression in cost. We're going like a 50x from maybe where we were decades ago to today. And in your view that there's another 10x waiting for us. The history of that, I feel like just as a layperson, a lot of that is reusable rockets, more computing power like for space launch and so forth. What are your predictions for what causes that next 10x in compression? Um you know, the next real drop that's coming, you're you're talking about the Launch per kilogram. Yeah, exactly. The next real drop that's coming is really driven primarily by reusability, but not just reusability. It's massive reusability. So right now they're launching rockets where half of it gets reused and they've reused them about 30 times. Um and in the future they're they have a very large rocket that's already been tested 12 times that's on the verge of being able to be reused dozens or more times. It's much larger. And so when you have a larger payload, you have more to take up, you know, your denominator is larger. So your cost for that goes down. And then on the other side you have it just being reused over and over and over again. You know, fundamentally we get you know, as as as laypeople, we get, you know, kind of accustomed to things being reusable. But in space until recently it's the equivalent of buying a car with a single tank of gas, driving it to a location, and then throwing it away. Mhm. Um and and that's effectively what's been done. So they're introducing So they're more and more reusable. Sort of moving from space as this like big rocket, never reuse thing to more like commercial air travel where the things are being used constantly all the time. Okay, that's That's awesome to hear. Let's I want to talk about power for next. That's the next thing that I think a lot of people think about as a big constraint on orbital economies. How do we generate power in space and how do we cool the equipment that uses that power? Well, I'm going to take your question and tweak it a little bit because it was just announced today that Meta signed an agreement with a company to beam power from space down to their data centers. That was just announced today. Maybe yesterday, but today. And basically what's going on in space is you have these very large solar panels. Just like you'd see a solar panel on Earth. They're a little bit optimized. But the difference is when they're up in space, it's somewhere between five and eight x the ability to convert sunlight. There's no, you know, atmosphere. We think of it as perfectly permeable, but you know, sunlight actually attenuates a little bit through that atmosphere. Um so you get a lot more power for the same thing up in space. And so in this example, you know, Overview Energy is the company. They're taking these huge solar panels, turning it into infrared light, which you can't see, sending it down at night to the solar panels on the ground, and they'll start actually, you know, converting just as if they would during the day, but you can't see that. It's just they're kind of being baked a little bit in the light IR. Um and then other companies like what SpaceX is doing, they want to take tons and tons and tons of solar panels and then turn that into put chips on orbit. So they're just keeping the energy up there. Um but fundamentally the Earth only gets so much of the sun's energy. It's It's a small dot in a very large solar system. The sun is throwing out, think of it as a giant fusion reactor in the sky, and solar panels that are out there are kind of like little collectors. So that's why there's so much opportunity. >> Kind of a trade-off. Expensive to get them up there, but once they're up there, they're able to produce power more consistently and at a higher wattage and so forth. What about the cooling side of that picture? Yeah, on the cooling side, um basically cooling is not that big of an issue depending on what you're doing. There's We're going through this big change in the space ecosystem where people are talking about, you know, orbital data centers and putting things in space. They need a lot more power than they currently do, but this is a, you know, energy can be neither created nor destroyed, right? If it comes in, it needs to go out somewhere. So it can come in as solar capture, turn to electrons, run a chip. That need That produces heat and that heat needs to go somewhere. Um so there are ways to to do that, but in space primarily it's radiative cooling versus convection. We think of like AC in here is moving air around. And radiative cooling is just sits out there in a in a vacuum and it slowly radiates out. And it's, you know, it's a maybe a little bit slower. You need surface area. The more surface area you have, the more you can cool. >> And so I I think sometimes in the discourse when people talk about data centers in space or Bitcoin mining in space, which is a topic I want to turn to, I think people acknowledge that yeah, there is power in space available for us to draw upon, but then they counter that with or their skepticism I think is anchored in the idea that um the waste getting rid of the waste heat from all that power, however you however you use it, is going to be a challenge. In your view, is it not that much of a challenge that we have sufficiently good radiative technologies that they can always counterbalance whatever workload that we're using up there? Yeah, it I mean it's it's a challenge. It's an engineering challenge. This isn't a physics problem. I think a lot of people look at something that might be relatively difficult to solve engineering and do a hand wave and say it's impossible. Um they're confusing impractical for impossible. So in reality on Earth, like something about 40% of all the power that makes it to a data center is used to cool the chips before it ever goes to the chips. Um so we're already putting up wasting a lot of that power. So once you start from an economics side, once you start to realize that, you realize, okay, once we create these nice radiators, they have to be double-sided. There's some innovation there, but it we're not talking about new materials or anything like that. We're talking [snorts] about putting many engineers in a room and just solving the problem. Um you know, effectively you end up being able to put more power in into the chip instead of wasting it on cooling internally. Okay, you heard it here first, folks. Radiators can into space. So we talked a lot about like sort of like the the really physical aspects of the problem [clears throat] of orbital computing and infrastructure. Let's talk a little bit more about the computing aspects. So things like computing equipment, CPUs, GPUs, data storage, and in particular bandwidth. Right? These are the other things that we need if we're going to build serious computing applications in the sky. Can you walk us through the constraints that we're under right now? For example, in bandwidth and how little bandwidth is actually available in orbit and in deep space and how that may change over the coming years? Yeah, there you know, when you're thinking about constraints in space, it's just physics first. So there's there's plenty of things to think about. One is like you said, radiation of heat. You have to get heat out, so you need surface area. Another one might be radiation from from the sun. You're actually getting, you know, we get sunburned. It's UV light, but you know, effectively it's the same kind of thing. If you bake a chip with those things, it can create bit flips and create problems. So you do have to take the radiation into account. It's again one of these solvable problems. Um and then there's I mean there's several other things, but one of the core things is that I think people underestimate just how much space there is in space. It's It's quite big and massive. So, the further you get away, the kind of the human mind has a hard time really fathoming how many miles that is. So, at some point you do start to actually run into a speed of light thing where this light can only go so fast. We're used to mega you know, millisecond delays on Earth and it those little you know, that data is moving from all over the place and it's going through fiber and all this stuff has gone many many many miles. But then if you need to go even if though it's a straight line, you need to go really far, it takes a while. So, there's a there's a true component of latency that you need to take into account. Normally that means on pretty large scales. To put that into context, it's like a couple of seconds from here to moon, it's measured in seconds of a delay. And then going from here to Mars, it's measured in minutes. 8 to 20 is the rough estimation how long you would have like one photon going from Earth to Mars and then the other back the same way. So, um these things start to play a significant role. Yeah, that idea of latency is actually very very interesting to me and I want to return to that. But a random question maybe to try to quantify the idea of how data is already moving out in orbit and in space. I'm putting you on the spot. Maybe you don't know the answer to this, but it's an interesting question. Is there more data being transferred I mean obviously data is transferred is being transferred from Earth to space and in the opposite direction. But which direction has more data transferred today? Would you estimate? Um right now there's a lot of so, the best kind of metric for this is Starlink cuz there's 10,000 of them in space right now which is 10,000 Starlinks out of I want to say like 13,000 total satellites. And so, Starlink's capacity right now across the whole network is approaching a thousand petabytes. Yeah, I'm I'm trying to remember all my zeros there. A thousand petabytes per second is is what it's approaching. So, that's a lot. Anytime you're you know, kind of know how down link is more important than up link or at least from a speed perspective often times. So, there's probably a lot coming down that way. Um from a limitation perspective when you go outside of just the consumer satellite internet space and you look at the industry, there is a down link problem because they you're taking a lot of pictures and observations that I was just a ton of them taken. We we've all seen Google Earth and Google Maps and has nice pictures. I mean, you do a lot of that over time. It's pretty hefty files and so, what you're seeing is it's it's kind of a bottleneck to get all that raw data down because you do have to hit down links somewhere. So, with a Starlink satellite, you have a dish and it comes down. These bigger satellites with taking big pictures are doing the same thing. Interesting. So, there there is a bottleneck if you're doing a lot of data down link. So, I guess let's turn to some of those the application dimension of this. We've talked a little bit at you know, about launch costs, bandwidth, power, computing elements. Um what when people talk about putting data centers in space, what applications are they envisioning running in those data centers? Yeah, I mean primarily the the current you know, these are pitches. They're still being developed. So, we have to you know, take things with a grain of salt. But primarily the applications for AI data centers in space are are primarily AI inference. And the reason for that is the what feels like effectively an infinite demand. Um you know, we use I use AI in our team uses AI tokens every single day. Like we feel it inherently. You've probably felt it as well. Um and so, it just feels like there really is no limit to that. Um and then the inference side of that is where people are people much smarter than me, Jensen Huang and others are expecting it to be you know, the biggest portion of the demand. So, I think it's really driven by that and then the costs you know, that the margin is what people are really going after. So, yeah we're not going to be running you know, SaaS applications for the finance team or whatever in in orbit. We're going to be running AI training and inference workloads. >> That's the orbital data center you know, kind of drive currently. But today like there are people actively doing Earth observation processing. What I mentioned that whole down link right? You take all these pictures, you shoot it over to a relatively small computer runs machine learning or even kind of LLM type of stuff on it and then you shoot back a response and it's the other response is much smaller than the raw data. And then they then they can down link a you know, a specific response or you know, a queue. Hey, there was something we should go look at that one picture not the hundreds of pictures you took before it. So, I'm hearing that is sort of co-locating more compute in space helps solve the down link bandwidth problem because we can send smaller amounts of data back that are more post processed. >> Yeah. Okay, that's very interesting. >> That's probably phase one. Yeah, okay. And and of course I think your your claim that AI inference and workloads are another major driver is probably also driven by as you said, expectations of revenue and demand. And I want to turn this obviously cuz we're at a Bitcoin conference not necessarily an AI or space conference. Though sometimes it's hard to tell. I want to turn a little bit to that question of Bitcoin mining in space. Like it's a it's a specific kind of computing application. Is it reasonable that we would see Bitcoin mining in space be successful today? And if so, on what scale? What would be the challenges? If not, how do you see that changing over time? Yeah, I would say like if you if you really you know, constrain me in a box that said within the next couple of years, then it's very hard to see that. To be fair, you know, orbital data centers are pretty young and a lot of the tests that are going up are over the next 2 years. So, they're still kind of testing all that. So, you know, given what I was mentioning earlier about kind of bandwidth constraints and all the opportunities that exist in in space, really what you're looking at are things that can have the highest possible margin at first cuz it's still costly to get up. As those costs come down, you know, into the 5 10-year window, maybe maybe it becomes a lot different. But right now if you just base it on a gigawatt of of computing power, you ask yourself what would make you the most money right now. Is that AI or is it Bitcoin mining? That's interesting. Yeah, I mean so, I I I have some numbers here from ChatGPT and other forms of research. I think part of what we've been feeling in the Bitcoin world for the last couple years is the displacement in certain for example Bitcoin mining facilities and other places of Bitcoin mining workloads in favor of AI workloads. That AI even for Bitcoin miners on Earth in some places is a more profitable workload to operate. And we're seeing mining facilities and companies turn to of course accepting that workload because it is more profitable. Um I'm not enough of an expert to to predict over the coming years how how does the demand for Bitcoin and the demand for AI training sort of interest intersect and overlap and so forth. But it is interesting that we're already seeing that effect on Earth and I think what you're saying is you're going to see that in space as well. That before we get to Bitcoin mining in space, we're going to be pursuing the more profitable avenues of things like AI inference and training. Um ChatGPT estimates that you know, if you keep all the numbers the same and you ask what Bitcoin price today you know, given launch costs and ASIC costs and the requirement to replace ASICs over time and so forth, Bitcoin needs to be a hundred times more valuable in order to make Bitcoin mining the dominant application that we might want to run in space. Conversely, you can sort of ask since the dominant cost here is still launch. What would launch costs have to fall to all other things being the same you know, $75,000 Bitcoin in order for Bitcoin mining to be profitable in space and and the answer is you know, on the order of tens of dollars a kilogram, right? So, again about two orders of magnitude gap right now. At least that's what seems to be out there. That might sound discouraging, but I think I'm actually rather encouraged by some of what I've heard you say today. That you're already predicting something like a 10x compression in launch costs over the you know, the coming I don't know, whatever time frame. I don't want to pin you down on it. And we've already heard today innumerable times about people telling us how Bitcoin has you know, maybe not infinite, but very high demand and that of course drives Bitcoin price and adoption here on Earth. This intersection of an increasing Bitcoin price, a decreasing set of launch costs and truthfully that as Bitcoin mining as an industry itself matures, the requirement to replace ASICs constantly over time with the next hardware generation becomes a little bit lessened. And I think that's actually very encouraging. My own hope or estimate or prediction as a non-expert here on the space side certainly be something like you know, optimistically 10 to 15 years for these numbers to kind of converge and maybe less optimistically you know, longer than that. But I want to turn in the last minute we have to probably the most speculative you know, part of the of the conversation. Something I found very interesting when we chatted is this notion that you've described a couple times even today around latency being the one thing you can't defeat. All right? That as we just discussed, launch costs will get cheaper, compute will get better, radiative cooling will get better. But we can't transmit data faster than the speed of light. And we're always going to be bound by latency no matter how sophisticated our technology becomes. And I think you had described it as sort of an onion shell of different applications that will surround the Earth with low latency things near the middle and high latency things far away. I think that's a very interesting parallel to some of my own writing around Bitcoin mining in far away locations like Mars and other places ultimately being limited by latency. And it's a kind of the same idea but in a vertical kind of Bitcoin domain. I think as my closing thought in this call, I feel like that idea that applications in space are ultimately always segregated by latency, deserves a name and I propose the Burnett rule. I'll take it. >> So, we coined it here. All right. Thank you all. Thank you. >> [music] >> Every year this community comes together to celebrate, to debate, to build what comes next. >> [music] >> And every year the stage gets bigger. >> [music] >> Sound money center stage. So, where do you go to celebrate the next chapter in Bitcoin history? [music] You come home. Nashville, July [music] 2027.