Nick Hansen, CEO and Founder of Luxor Technology joins me to talk about the bitcoin mining scene, the narrative around bitcoin mining energy use and Proof of Work, Demand Response, ‘ESG’, China ban, Intel coming into the mining game, and hash rate derivatives. 

Links:

• Twitter: @hash_bender 
• Site: luxor.tech

Sponsors: 

• Swan Bitcoin
• Hodl Hodl Lend
• Compass Mining
• Braiins.com
• Unchained Capital (code LIVERA)
• CoinKite.com (code LIVERA)

Stephan Livera links:

• Show notes and website
• Follow me on Twitter @stephanlivera
• Subscribe to the podcast
• Patreon @stephanlivera 

Podcast Transcript:

Stephan Livera:

Nick, welcome to the show.

Nick Hansen:

Hey, how are you doing? Thanks for having me.

Stephan Livera:

So Nick, you are working with Luxor, you’re the CEO, obviously. Do you want to just give us a bit of a background on yourself and how you got into Bitcoin mining?

Nick Hansen:

Yeah, sure. So myself previous to building Luxor I was a lead member of technical staff at Salesforce working on Einstein, effectively—that’s their AI offering. So imagine something like Alexa but for Salesforce. Always been interested in crypto, was mining Bitcoin with GPUs back when you could do such a thing in my basement. Mostly as a hobby, just really interested in the tech, trying to understand what this thing was. This was around the Mt. Gox time and it was rife with scams and people didn’t really know what was going to happen. I think the entire market cap of all crypto in the world was something like $10 billion, if even that. It was just mostly playing around with different wallets and installing things. Got rekt and took a little siesta from all of it. And then really came back hard in 2016 and 2017 working on some projects that I was really interested in. Around that time I met one of the original founders of Luxor, Eddie. We were working on something completely unrelated to mining, or Bitcoin even. But we met on GitHub, found each other on Slack and then just started building what is Luxor today. We started building an altcoin pool, because it was a lot less scary to build than a Bitcoin pool. Bitcoin mining was quite large to us as two guys just working nights and weekends. And so we started building this altcoin pool and then over time just realized we kind of have a knack for this, started finding product market fit with some of our products and went on to go through the Luxor ecosystem, which is now much more than a mining pool. We do ASIC brokerage, we find hosting services, managed mining, all sorts of things that we do now outside of just the pool. So at this point we’re really considering the pool as just a feature of the broader ecosystem. And I think you can see that play out in a lot of other areas of mining as well, Foundry and all of those. They don’t even take any revenue from their pools. Anyway, that’s me, that’s how Luxor got started and that’s where we’re at. We are known for being a pool, but now we’re quite a lot more than that. We’re a team of 40 now.

Stephan Livera:

Excellent. And so maybe if you could just spell out some of the different service offerings there, because obviously as you’re known as the pool, you’ve got the Hashrate Index, you’ve got equipment procurement, you’ve got mining advisory. Could you just spell out some of those just for people?

Nick Hansen:

Yeah, certainly. So mining pool, pretty basic. I think everybody knows what a mining pool is. If you don’t, it’s basically the thing that coordinates work between all of the mining machines that are out in the world. You hear about Marathon and Riot and all these companies buying up machines—well they plug into a mining pool, and we manage those. The next step from there is procurement. We naturally interact with a lot of miners, interact with sellers of mining machines, and interact wwith buyers of mining machines, and so it was a natural fit for us to facilitate that trade, help people get machines from wherever they need them. Usually we’re brokering new machines, but sometimes we’re helping with used machines and getting that out there now as one of our new product lines. We’ve had great success with it. We launched it in Q4 of last year and helped procure almost 12,000 machines. So it’s really, really exciting to get that off the ground in the first quarter. And then mining advisory: so a lot of people are trying to deploy into mining right now, but have never mined. Usually it’s energy people, and we’ll get into energy shortly, but people that don’t really know how mining works but they know how energy works, we come in and we help partner with them to find ways for them to deploy that capital into mining, whether that’s through stranded capture, plugging into a grid area that they have access to. All sorts of different ways that we can come in and assist from that perspective. And that’s more than just procuring machines, that’s actually getting containers and transformers and plugs and all of the stuff that you need and then helping them get it set up. So those are some of the big ones that we’re working on. And then Hashrate Index is what we call our retail focus line, meaning we want people to come put their eyes on that and see the data that we have to share. Two of the most popular metrics that we have are the Hash Price Index and the Rig Price Index. Hash price is the value of a Terahash at any point in time. We really started pioneering this idea in 2020. And that’s what we built Hashrate Index for, was to start displaying this idea of hash price and get people familiar with it and see that it’s volatile and changes over time and it’s somewhat correlated with the price of Bitcoin, but not entirely. And then the other one was the rig price, which is the value of mining machines as they change over time. Our product officer Guzman has spent a ton of time finding data sources for that information, aggregating it, figuring out a very accurate way of calculating what that metric is. And then we bucket it by efficiency. So we’ve got effectively like an S19 bucket, an S17 bucket, an S9 bucket, and each of those have different prices throughout the day. And we draw really interesting insight from that. For example, the S9, the S9 series, or that efficiency bucket actually appreciated in price much more than the S19 did during the last big pump late last year. So it was really interesting to see that, and we think that was because that’s one that has come up over the breakeven curve in a lot of places. So those are some of the product lines that we’re working on. We’re also spearheading and on the leading edge of hashrate-based derivatives. So financialization of hashrate, taking hashrate and using it as collateral, using it as a way to fund your operation. That’s something that we’re going to be spending a lot more time on in 2022.

Stephan Livera:

That’s great. And I think one thing that came out from what you were saying as well, is the difficulty of actually getting good numbers on some of this. And of course, some of this comes into the energy conversation that we’re having around Bitcoin, and that comes up around, Oh, look, how much energy Bitcoin mining is using. And what’s the carbon emissions of that? And should we be using proof of work? And of course we should be using proof of work, but the challenge then becomes, How do you educate people who aren’t familiar with this? So what’s your view, Nick, of the Bitcoin energy debate as it were, and the current state of play of education of the broader world out there—obviously outside of our Bitcoin world?

Nick Hansen:

Yeah. So there’s obviously two classes of people that we have. We have folks that are really into Bitcoin and understand this, and then people that just see what the media feeds them around—CNN or whatever is going to say that Bitcoin uses a ton of energy. These folks, they don’t really want to dig in and understand a lot of the intricacies there. So we can do a little bit of simple math, very basic rudimentary math. I can’t tell you exactly what the amount of energy that the Bitcoin network uses, but I can tell you the absolute minimum amount and the absolute maximum amount that it uses. So the way that we can do that is we can use the lowest efficiency machine that’s usable on the network today, which is the S9, and we can basically work backwards from the network hashrate and determine that it’s around 18 gigawatts. If every machine on the Bitcoin network right now was an S9, we’d use about 18 gigawatts, which is not that much. And then if every machine were the latest gen, meaning like an S19j Pro, or the most efficient miner that exists today, it would use about 6 gigawatts. So somewhere between 6-18 gigawatts. I don’t know exactly what the number is, but that’s a pretty tight range given that ERCOT itself generates something like 90 gigawatts itself. That’s the entire [Bitcoin] network. That’s just Texas. We’re just a drop in the bucket. But I don’t like that argument, because we need to make that drop much larger. We need to become a much larger portion of the bucket to continue to secure this network. And so the next step in the argument is saying, Well, okay, so you don’t really use that much energy, but you want to—so what does that mean? Right now, the cheapest generation sources use a certain amount of this energy today, and we necessarily need to use more. So the natural next argument would be, Okay, maybe you don’t emit as much carbon as we think you do today, but you will eventually. But right now, the cheapest sources of energy are generally the cleanest. Using flare gas capture, your cost is effectively zero. And in some cases you get paid to take off that extra gas. And that using that energy that would’ve otherwise just been wasted. Using the sources in ERCOT—ERCOT is very, very clean: natural gas, wind, and solar make up the large proportion of its energy production, and by providing load to that grid, the Bitcoin network is able to make that grid much more robust, and able to make it survive some of these events like we’re seeing now, which is a good segue into what’s happening today in Texas.

Stephan Livera:

Yeah, of course. And so this is coming up today as well because of what happened last year with some outages. So just for listeners who aren’t familiar with what happened last year, could you just give the background on that?

Nick Hansen:

Yeah. So last year, Texas had an unprecedented cold snap, meaning it got much colder than it generally gets, and their grid was not able to support it. So people were without power, and in a place like Texas, that means a lot of people that aren’t used to cold weather are not prepared for being subject to a very, very cold winter. It was actually a couple of weeks. And so basically what was happening is the grid was not able to support that demand for energy. So this demand for energy was coming, which was to heat homes and to keep lights on and do all the things that people do when they’re cold—the grid was not able to [handle it]. So fast forward a year, they’re getting ready to go through another very big cold streak here, but we’ve had a whole year of some of the biggest mining companies in the world building out on the ERCOT grid. And the story that they’ve been telling for a whole year is that we’re going to use a bunch of extra energy, and the idea is that you’ll be able to produce a bunch more. So let’s say Houston on any given day needs 50 megawatts. Well during this cold snap, they need 150. Well you can’t build out 150 megawatts for a single city because you don’t have anywhere for that extra 100 to go when the grid is not needing it, so it’s effectively wasted. And so what Bitcoin miners will come in and do is they will provide that demand, that base load—like Bitcoin mining, this type of base load has never existed in the world until today. This economically-incentivized base load that has effectively zero turn off cost other than the opportunity cost, and it’s also instantaneously rampable so you can turn it up and down effectively instantly. A few seconds here and there. But this idea that when Houston or some city in Texas gets very cold and needs an extra 100 megawatts—well now Riot and Lancium and all of these great mining companies that have been building out in Texas will just turn off their Bitcoin mine and let that energy flow through to the city and bolster the grid. There could be times in the summer when it’s hot and people need all the air conditioners turned on that this works just as well. So the idea of building these demand response centers really bolsters the grid and gives Bitcoin mining a phenomenal story to tell to the market.

Stephan Livera:

And so it comes down to how much of that is a factor, right? Because I could understand where maybe somebody might be listening, they might be a bit more skeptical and saying, Well, how much is that? How important is that, really?

Nick Hansen:

The demand response part? Well, we’re going to find out. So you’re right. It’s somewhat of a double-edged sword for the critic here, because they’re saying, Well, Bitcoin uses a lot of energy, but in this case, what they would say is that Bitcoin doesn’t use enough energy to even make a difference on the grid. And we’re going to find out maybe this year—we’ve only had a year in Bitcoin mining. Bitcoin mining may be one of the slower moving areas of the sector, because it takes a long time to pour concrete, build buildings, plug in machines, get them from China—all that stuff. So we only had a year to build out. Hard to say exactly how much capacity they built there—maybe a gigawatt over the year? But we’re going to find out soon whether this is going to work out or not. There have already been press releases from Rhodium and Riot about how they’re spinning down their capacity and letting that energy flow through to the grid and to people’s homes. We’ll see if it’s enough this year.

Stephan Livera:

Yeah. And it also comes down to what is the cost of energy at this incredible peak demand time. And that’s part of this whole argument, right? Because the argument is that during these times of incredible demand, the price of electricity is going to shoot up so much that these Bitcoin miners have an incentive then—it’s rational for them to actually turn off and say, Look, I’m just going to take that income for turning off, rather than leaving my mining going.

Nick Hansen:

Right, exactly. That’s actually a really good point. So they get paid to curtail. In some cases, they’re contracted to take a certain amount of energy. But also that means that sometimes there are times when it is beneficial for them to shut off because they’ll get paid to curtail. The problem that we’ve had in the last year is that Bitcoin mining has had such thick margins that they’re willing to pay really, really high cost for that energy. As we reach more of an equilibrium in the price of Bitcoin, in the marginal cost to produce Bitcoin, I believe that that will become a much narrower band. But right now I think Bitcoin miners could be profitable up to 40 or 50 cents. I don’t know that off the top of my head, but some number that is much, much higher than the average. Here in Seattle I pay 12 cents a kilowatt hour. I know that I could mine an S19 very profitably just here in my house with current market conditions. But I know that won’t be the case forever, because there are places that are purpose-built in doing this type of demand response consumption that will eventually beat out my miner. And in that case, then we start getting to a point where the value of the hashrate and the value of the energy start to become a little bit more in equilibrium. Then it becomes a lot closer to the shut off point. So say for Riot, the price of energy goes from probably 4-5 cents, then it goes up to 20 cents—then they’re going to shut off, but I’m very happy to pay 20 cents to keep my household warm. My incentives are much more external than purely profit-driven.

Stephan Livera:

Yeah, that’s a good point there. And some of that also comes into this question of cost to mine. Now of course there’s so many variables that go into that, and there are profitability calculators. You can look at, Okay, what machinery am I using? How efficient is it? What’s my cost of electricity? And then what’s my break off point? Or so on. And I’ve seen others comment that just because of various conditions today, given the energy prices, given the availability of the mining equipment, the current network hashrate, that a lot of the even publicly listed miners have actually a cost to produce Bitcoin that’s quite a lot lower than the current market price of Bitcoin, which is call it $37,000 as we speak today. Some of these miners, apparently their cost to produce is something like $10,000 or $15,000 per coin—in that range, right?

Nick Hansen:

Yes, 100%. So you can go look at the public filings and see exactly how much these guys are mining for. But on average, I would say that some of the best miners in the country, at least in the US, are probably somewhere in that $7,000-$8,000 range that they’re able to mine, but they do have to amortize the cost of capital over their lifetime. So that’s how much it costs them to actually mine the Bitcoin, but then that revenue margin there between the $8,000-$9,000 or whatever it happens to be, let’s just say they make $30,000 a coin in revenue. They also end up having to pay for their machines and for all their infrastructure. So there is a lot of amortization of CapEx that a lot of people miss, but at its very base, yeah, there’s a lot of thick revenue to be made there.

Stephan Livera:

And I’m curious, and I imagine listeners are also curious, what might be the reason for that? Could it be—so again, there’s been this big chip shortage and it’s been difficult to create as many mining rigs as the market demand is, as much to basically serve that market demand. Is that part of the reason why that there’s that disconnect?

Nick Hansen:

Yeah. There’s a lot of reasons. So one, yeah, it’s definitely been harder to get machines than before. Nobody would argue that. The number of machines getting plugged in now is vastly lower than the demand. And that’s because they just can’t produce enough. But—and I’m going to get blown out in the comments for this—hashrate follows price. Hashrate follows price—I don’t care, we can argue about it in the comments all you want, but hashrate follows price. And that’s what we see here. So price generally moves a lot faster in one direction or another than hashrate, and when we see these big pumps in price—let’s just say we went up effectively 4x-5x from the 2017-2018 high—hashrate has not followed that trend yet. We’re getting there. We’re definitely catching back up. The China ban threw a little bit of a wrench in things, but hashrate is still catching up. So it’s definitely lagging price, and we’re going to get there where hashrate and the marginal price to produce Bitcoin is going to come up, and either the price of Bitcoin goes down or the marginal price just continues to rise up to that level where it becomes only profitable to mine for the most efficient miners. Which is what we saw Summer of 2020—that was the record low for mining profitability. I think that may have been the first time, at least in my history, that we saw the marginal price of mining a Bitcoin bump off of what I call the difficulty cap. So the difficulty actually dropped a couple of times because miners were dropping off the network—it was no longer profitable to mine. And that was one of the first times we ever saw that.

Stephan Livera:

Yeah. So it’s a range of reasons. As you were saying—that’s one of the arguments, right? The argument is, Does price track hashrate or does hashrate track price? And I’m actually with you. I think it’s more like hashrate is following price, but it’s like a lagging indicator. And there are times where maybe it wants to catch up, but it can’t. But then times when let’s say—

Nick Hansen:

Yeah, we’re certainly right there where right now we’re actually mostly capacity-constrained, meaning we don’t have enough gigawatts in the world available to plug in. Maybe we segue into the mining ban in China, because there was a lot of energy in China that came offline, and those plugs are no longer available to plug in. So we have to either rebuild them here or elsewhere to get that amount of hashrate back online, and then continue on the path towards that equilibrium.

Stephan Livera:

And so put in other words is that, at the time of the China band—call that halfway through last year around July, August or so of 2021—we were talking about literally hundreds of thousands of Bitcoin mining machines that were being unplugged in China, sent overseas, and plugged back in. But as you were saying, the constraint in that case was also around what we might call rack space, that there’s not enough buildings and racks and everything lined up where you can actually physically go and plug in all of those machines. So maybe that was where one of the scarcities was, at least at that point in time. Is that what you’re getting at?

Nick Hansen:

Yeah. So we did the math. We found that, at a minimum, there were 600,000 machines that came offline during that time. Because you could see the hashrate drop and then you could do a little bit more estimation and figure out how much actual power that is to build out. And I’m very surprised at the human ingenuity that was put into getting those machines back online. I didn’t think we would’ve broken 180 EH/s in 2021. And we eclipsed that pretty easily. So exactly like you said, all those machines came offline, and the problem wasn’t that people didn’t want those machines, it’s just that they had nowhere to plug them in, and it takes a long time to build out these facilities.

Stephan Livera:

And what of the question of guerilla mining? So I see there were a couple news articles about this of people who are still in China, and potentially they are using some means of masking or hiding their use of the power and using say a VPN to VPN to other countries, and then that hashrate looks like it’s actually coming from another country, but actually it’s still back in China.

Nick Hansen:

100%. Yeah. So we learned about this and I think that was one of the biggest contributing factors to getting back to the hashrate that we are today, is that people were figuring out ways—our Chinese counterparts in Bitcoin mining in China, they’re clever guys. They’re clever folks. They’re going to figure out some way to get those miners plugged back in. They’re either going to work with their local municipality to shield them from the federal government—I heard some stories of folks that had their Bitcoin miners in effectively semi-trucks, and they would hear that, Oh, there’s going to be a federal inspection of the grid now, so you’ve got to get out of here. And so they would like move and then come back when it was safe. I’ve heard of that, and then like you said guerilla mining where they’re so far off the beaten path that they’re not worried about inspections or anything like that coming through and either seizing their machines or fining them or whatever it is that they’re doing. So yeah 100%, there is guerilla mining going on in China, and we applaud them—keep those rigs on. We love to see it. And I really hope that they can continue to do so. Also heard a little bit of news through the grapevine that they may be—I don’t want to say un-banning mining—or maybe they’d be un-banning it but adding an additional tax, which, not a huge fan of adding a tax, but if we can get down the path of getting China back online, I do think it’s good for the network to have an equal distribution of hashrate around the world.

Stephan Livera:

Why do you believe that? Why is that a good thing?

Nick Hansen:

So whether we like it or not—Texas is maybe one of the freest places in the world, and they’re going to stand up to the federal government if the federal government in the United States told Texas to turn off—they would just give them the finger and continue to do so. But that doesn’t mean that it’s not a possibility. So what that means is it doesn’t really matter where the machines are located. I don’t think that they are all going to forever be insulated from some sort of external power having influence over them. So I’d like to see 10% in Texas, 10% in the rest of the United States, 5% in Canada, 3% in Central America, 10% in South America, 15% in China, 5% in Kazakhstan, et cetera, all around the world, so that way it doesn’t really matter. Kazakhstan shuts down, 5% comes offline, just gets moved somewhere else and no big deal. And we’ve seen that for sure happening now, but I am a little concerned with how much we are getting here in the US. We always marketed ourselves as North America’s first mining pool and the greatest North American mining pool. And I always said that North American mining is going to dominate the conversation over the next decade, but I didn’t want it to be like this. I wanted it to be because we could out-compete our Chinese counterparts and our global counterparts, but not because their governments shut them down.

Stephan Livera:

And on that topic of inter-country competition as well, because there were different views that I saw—some people were saying, Yeah, look at how great this is that all this hashrate is coming to the USA. And so perhaps in some sense, the patriotic element of it was there, but perhaps from the overall point of view of Bitcoin as the decentralized project, distributed project, that’s where this argument is of, Actually it’s better if it’s distributed out around the world.

Nick Hansen:

Yeah. I was as happy as anybody to hear—my company directly profited, and a lot of American companies directly profited from this happening. But my personal view on Bitcoin as a project is that we want this to be as decentralized as possible. The more machines we can get the better, and then don’t put them all in the same place.

Stephan Livera:

Yeah. And I agree with you there. With the aspect around regulation and the public perception of that, what do you believe are the best ways forward? Do you see it as education? Do you see it as Bitcoiners have to get involved in the political game even if they don’t want to?

Nick Hansen:

Yeah. I think we’re going to see a Bitcoin-centric party form in the US. Whether we like it or not, American politics crafts a lot of the global conversation. And so I think we 100% need to have a Bitcoin party. Whether that’s one of the parties that exist today—I prefer it not be. I’d prefer it be a new party, or somebody that’s able to come in and really disrupt what’s happening here. I do know that if we do get a Bitcoin-denominated party over the long term, it’ll be the best-funded party that exists. And so maybe that allows them to win more, kind of like Peter McCormack—his football team—I think they’ll be one of the best funded football teams over the long term because they’re denominated in Bitcoin. I think that this is the path forward for us. This is how we craft the conversation. We get a political party that is Bitcoin-denominated, and then start having those conversations, bring them to the White House, bring them to Congress, even your local governments—it all starts local and then filters up. And I think we need to be having those conversations, and we need to be talking to people about this. So it’s education and political activism combined.

Stephan Livera:

Right, because the way we all see it, we’re all bullish on Bitcoin. And as we’ve agreed, price is going to go up, and if hashrate follows price, then necessarily there’s going to be a lot more Bitcoin mining machines that get plugged in. And this conversation is only going to become more and more important and larger in the broader political world, whether we like it or not. And so that also brings that question then of whether Bitcoiners should be trying to make the argument—and I’m curious your view on this—should Bitcoiners be trying to make the argument of, Oh, look, it’s not that much carbon emissions, or it’s lots of renewables, or do you see it more like it should just be made on the principled argument of, Even if it was 100% fossil-fueled mining, it would still be worth it?

Nick Hansen:

I think we could try to make that argument. I don’t know if it would land. And being pragmatist, I would prefer to just take the easiest path to acceptance. And if that happens to include an ESG narrative, I’m happy to go down that road. And I do agree with you that even if we’re all coal powered, 100% dirty, I think that the value that Bitcoin brings to humanity and the globe is worth the cost, because I don’t think that there’s been an innovation in the last decade that will change the planet more than Bitcoin. What it brings to people from an inequality perspective, from an equity perspective, allows you to actually own something without anybody’s intervention—that has no [price]. So I do agree with you that there’s no price that you can put on this thing, but I do think that it’s a lot easier to tell the story that Bitcoin doesn’t use that much energy and is green in all these different ways. So that’s the way that I would like to tell the story, just because I think it’ll be an easier one, but I think long-term we can tell the story that it doesn’t matter the price.

Stephan Livera:

And so the hope also is that those of us who see Bitcoin as this money of enemies or money for anyone to use—we also don’t want to see it get regulated. And I suppose that’s one of the arguments is that if we try to kowtow to the ESG narrative, that potentially there might be more censorship coming down the line, whether that be whitelisting, blacklisting, other kinds of, You may not participate unless you are X, Y, Z, or X% level of renewable, right?

Nick Hansen:

Yeah, definitely. It is a slippery slope, but I think that every slope that we live on is a slippery one, and so we have to figure out where it is we’re going to stake our foot and stop the slipping. I think kowtowing to an ESG narrative in the interim is okay. Obviously not ideal, but we live in a world of compromise where not everybody’s going to agree on everything. And so if you can get a majority to agree with you on one thing, try to use that majority to your advantage. And I do think that over the long-term, people will come around to this idea, especially as they start to own Bitcoin, and Bitcoin just becomes more ubiquitous. I think adding things like Strike Pay is going to definitely change the way people think about Bitcoin, because it will allow them to interact with a global monetary system almost frictionless. So in that regard, I think it’s just about getting a critical mass of people interested and involved, and from there they’ll all view the value of Bitcoin. The price you have to pay for Bitcoin to exist, the value that you get from it, is vastly more—it’s orders of magnitude greater.

Stephan Livera:

So thinking now about Bitcoin mining hardware, I’m sure you would’ve seen the recent news around Intel getting into the Bitcoin mining rig production creation game. So I’m wondering if you have any thoughts on that or any ideas on where that’s going? Will they be competitive with the current competitive Bitcoin mining manufacturers?

Nick Hansen:

Yeah. So we have heard a little bit about those machines. So I think that the consensus is right now that they’ll probably fall somewhere between—so the S19 XP is the latest version of the Bitmain miner. It’s their pinnacle machine which is supposed to come out in Q3, I’m hearing that it’s probably going to slip to Q4 of this year. And then their current model is the S19j Pro, which is their most efficient one. It seems like the Intel’s going to fall somewhere in between those two, so it’s going to be a really competitive option. And really now all they have to be able to do is compete on price of the machine. If they can produce a machine that is lower cost per Terahash than the S19 XP or the S19j Pro, then they certainly will be competitive. People will buy the machine. I mean, people still buy S9s today, even though it’s a much lower dollar per Terahash number, but the S9s are from 2016-2017. So I definitely think the Intel machine will be competitive. A couple of their barriers right now will be (1) producing at scale. So I think Intel, if there’s anybody out there that can produce at scale, it’s probably Intel. So they may have that one covered. We’ve seen some other popups come around that have not been able to produce any meaningful number of machines. They produced a machine that’s marginally competitive, maybe competitive with the previous gen, but they haven’t been able to produce enough of those machines. So Intel I think we’ll be able to. And then (2) the next step that they have to do is be able to get chip space, which I think Intel should be able to do no problem. They should be able to get enough capacity in the foundries to produce enough chips. So I think that they’ll definitely bring a machine to market. It’s going to be competitive. And I think if the first run is even close to competitive, that means the second one is going to be really good. So this is great for the space. It’s a huge validator for everybody that’s been around. Intel, the biggest chip-maker in the world, is going to come to market and try to compete here. I wouldn’t want to be in Bitmains’ and Whatsminers’ shoes for sure.

Stephan Livera:

And how will this change the game of hardware procurement? Because as I understand today, sometimes you have to go through a partner, and I know obviously with Luxor you have hardware equipment procurement and advisory services. So what does the current state of play look like for somebody who’s trying to purchase mining equipment generally? And then maybe we’ll talk a little bit about what it looks like if the market gets bigger and bigger.

Nick Hansen:

Yeah. So buying machines right now is definitely an interesting space. So in general, there’s a couple ways you can buy machines: you can buy futures orders, meaning say, I want to buy a delivery of machines in Q3 of 2022. You need to buy a minimum amount and they get delivered over each of the months during that quarter. Let’s say I buy 1,500 machines. Usually they’ll come in batches of 500 over the course of that month. That can change. All of this is very flexible, so if anybody in the comments is like, Eh, that’s not what I did, well it’s because it’s very, very different for every order. And what the Luxor service is here to provide is clarity on how many different levers there are in this equation. So you buy these futures orders, you buy the machines, well you have to put up 60% of the order price. So right now you’re looking at around $15 million if you wanted to buy 1,500 machines, probably gone down a little bit since the price of Bitcoin has gone down, but still—you have to put up 60% of that. And you just wire that to a Chinese entity. And if you’re not working with a broker, there have been cases where that money just goes away. So Luxor comes in, provides you a level of security, and makes sure that that money is actually going somewhere and is actually going to result in machines. So once you put down your deposit, then within 30 days of your order it’s starting to be completed. So it would be within one month of the start of your quarter, you’ll put down another 20% and then 20% to get them shipped. So that’s how a futures order generally works. That’s not always the case, but that’s pretty much your prototypical order. Now spot orders, meaning I want to buy machines and I want them shipped now, generally are 100% upfront. So if I wanted to buy 1,500 spot machines, wire my $15 million to somebody, and they start shipping them within 10 days. Again, this is all very unorthodox when you’re procuring infrastructure equipment. I’m certain that Dell and Intel, when they go buy their new servers for their data centers, have a little bit different process than that. But that’s the world we live in—the mining manufacturers right now have all the power in the world because they have effectively money printing machines that people want to buy, so we have to play by their rules. So Luxor comes in, helps you a bit in that regard. And that’s all for new machines. Buying used machines just opens up a whole new can of worms. It can be all over the place, especially if you’re buying previous gens like S9 or S17, you end up losing a lot of machines because they get damaged in shipping or they’re not good from the start. So Luxor provides a barrier there, or a little layer of assurance that you’re actually going to get close to the number of machines that you attempted to order.

Stephan Livera:

Yeah, that’s interesting because it seems like if you are not already well connected in that world, and potentially also if you are doing smaller orders, that’s where maybe the value of having somebody to broker that for you helps more so. Whereas if you are a bigger player or you’ve already built those relationships, then you might be in a better position to just go it solo.

Nick Hansen:

Yeah, certainly. Like Marathon probably doesn’t need a broker. They go out, they buy 100,000 machines. They go direct to Bitmain, probably get the CEO on the phone and talk to him about it. It doesn’t even include their secondary buyers in China that go out and buy machines straight. Because they have the relationship with the manufacturer, they go, they buy say a 100,000 orders and then they resell them for a profit. And that’s usually where most of the machines are coming from is from these resellers. And so it’s a very disparate world, very difficult to navigate, and so really what the brokers are doing, Luxor included, is just trying to provide you that layer of assurance the machines you actually get and pay for are going to show up.

Stephan Livera:

Right. And so I’m curious then if you have any thoughts on how things change as the Bitcoin market and the mining market grows. Does it then become a bit more like standard technology vendor relationships that exist today?

Nick Hansen:

Yeah. 100%. I think we’re headed in that direction. We’re maturing as a market for sure. And I think that we’re going to continue to mature. I think we’re headed in a direction where it becomes a lot more egalitarian and a lot more transparent as well.

Stephan Livera:

Interesting. And I’m also curious as well if you have any thoughts around hardware life cycles, because this is an interesting one from years ago, because—my understanding and you tell me if I’ve got that right—so in the earlier years of Bitcoin, because there was so much tech advancement happening so rapidly, the life cycles were much shorter. And then what’s happening now it seems is that the life cycles are stretching out more, that you are potentially having more and more years that you can use that machine. And so there are these circumstances where if you have very cheap power, running on an S9 from 2016 or 2017 as you were saying, it’s still quite profitable for people. But if you’re operating on expensive power rates, then you really need the more efficient, newest, latest, and greatest. So I’m curious, where do you see that going with hardware life cycles? Is it going to lengthen further? Obviously there’s all these variables out there, but just if you had to speculate?

Nick Hansen:

Yeah, certainly. You hit the nail on the head. The early, early years of Bitcoin mining, we went from CPU to GPU to ASIC mining really, really fast. Like by the time I got my GPU rig up and running, Butterfly Labs was producing ASIC mining machines. I went and bought one, I bought a Jalapeno, the little cube. And by the time that thing showed up, it was obsolete. And so we were just plowing through this series. And so what was actually happening underneath the hood is they were probably producing chips that were—it’s called a process—so they were probably 55 nanometer chips, which at the time, almost anybody could print those. It really didn’t take a genius to do. But as the chip process gets smaller, it requires more and more expertise on how to produce those chips. And that’s where we’re seeing—we’re bumping up against that loss. So now we’re producing chips that are like 5 nanometer and there’s like a theoretical limit. They think that you can only go down to 3, maybe 2 nanometer chips. So we’re bumping up against that hard limit on the chip process. So as we were going from 55 to 5, there was massive innovation that was happening very, very quickly. And so that’s why we burned through those first generation machines in no time at all. The S9 came out, landed and like you said, those are still being used in production today. From 2016-2017 those machines were produced. And then the S17, there’s public companies that are basing their entire business model around the S17 series. The S17 series did have its troubles. They used a different manufacturing process, which caused it to be very delicate and breakable, but there are people that built business models around fixing those machines—getting them at a discount and then fixing them and using them. And then the S19, I could see the S19 generation lasting for 5-7 years easily. That series as a whole will probably last 5-7 years, just because of the marginal increase in efficiency is reducing to a point where all I need to do is take my S19s and find a little bit cheaper power. I need to only find 5%-10% cheaper power to keep them on, where before I would need to go find power that is 80% cheaper to continue running my old machine. So 100%, you hit the nail on the head with machine series and how they’re progressing. It’s slowing down, certainly.

Stephan Livera:

And just that comment around the series, because there are many, I guess S9s and S17s and S19s, could you just explain that for listeners? Why is there a difference between these?

Nick Hansen:

Yeah, that’s a good point. So in what we call the 19 series, there’s the S19, the S19j, S19j Pro, S19 Pro there’s a T19—there’s all these different models within that series. They all have a little bit different price, a little bit different efficiency, and really that comes down to the chips that they put on each board. Like I’m using a MacBook here: the chip in this MacBook and the chip in an identical MacBook are going to have different performance specs, but it’s going to be very small. And they keep them all within a range, but that doesn’t mean they’re gonna be exactly the same. And so that’s why they have these different models within a series. So the 17 series, there’s a T17, an S17, an S17 T, like there’s all these different [models] within the S17, and same with the the S9. So that’s why I call them series. Generally there’s a bucket that the efficiency of the 19 series is somewhere between 36 and 42 joules per Terahash. I don’t remember what the S17 is off the top of my head, but yeah they fall within an efficiency range and that’s why we call them a series.

Stephan Livera:

Excellent. And so when it comes to actually procuring them, I guess that’s why they’re bucketed in that way. How about the difference in reliability? How are you seeing that? Are there some manufacturers for which their equipment is more reliable and it can last longer, or it’s maybe easier to repair? Let’s say the Toyota Camry or the Honda Accord of the Bitcoin mining world?

Nick Hansen:

Yeah. So the S9 is the Honda Accord—not the fastest, not the prettiest, but man that thing is going to run forever. Same with the Toyota Camry. So the S17 series I alluded to earlier had an issue with its manufacturing. So they used a different type of glue on the heat sink that was not up to par. And so what would happen is during shipment, the heat sink would rattle loose. And the heat sink is a little thing that attaches to the chip and gets the heat out, and it would rattle loose. And then the chip would just melt, right? The same thing would happen to your Macbook or any other machine that has a chip in it, it has a heat sink on it—if that heat sink rattles loose, the chip’s going to melt. So that was happening to the S17s. And they have a failure rate very, very high. It’s not because the chips or anything about the chip board themselves are bad, it’s that the heat sink was not produced correctly. So like you said, there’s definitely variance within series. The S9 was a really good series. There was not a lot of manufacturing defects with the S9s, and we’re seeing the S19s are also really, really good. And now across manufacturers—so the S series, those are all Bitmain machines. There’s also a brand called Whatsminer. And those Whatsminer machines we’ve seen have generally been very high quality. It doesn’t seem like they’ve really had a bad series like the 17 series. They’ve had the M30 series that is equivalent to the S19, and it seems like the failure rates on those are really, really good. It’s not really possible to put an exact failure rate on these things. It’s all anecdotal. So it’d be like me asking my friends, or talking to people about what kind of machines they use and how many failures they have. And there’s so much nuance there because the machines running in an immersed building versus machines running in an air-cooled facility out in the middle of the desert are going to have vastly different lifetimes. And it’s really hard to pinpoint exactly what the lifetime of a machine is.

Stephan Livera:

And on this question of lifetime, I think that’s the other argument that I’ve heard as well around this demand response. So bringing it back to the demand response question, as I understand, if you’re running a car on the highway and it’s just on all the time, that’s efficient, but if you’re in the city and you stop/starting all the time, it’s not as efficient, it’s going take more fuel. Is there a similar impact there for that demand response miner, for example, even in Texas, are they going to see reduced lifetimes on their mining equipment? And is that going to be an issue?

Nick Hansen:

Yeah. So we don’t know first, first of all. Anybody that says they know for a fact doesn’t, because we haven’t done this at scale for a very long time yet, so we’re not 100% certain. If we look at just chip technology as a whole, there shouldn’t really be any degradation of lifetime or performance. Basically, if a chip is using energy, as long as it stays cool, it should last just as long. So in the case of lifetime, it’s very difficult to say for a fact, but most likely not—it probably will not affect the lifetime of the machines. As for energy efficiency, no, I don’t believe there’s any [degradation], because it’s actually more like a negative battery. If you go around the Bitcoin mining facility, say you’ve got a big hose and it’s going through the facility and all the energy that’s flowing through it is going into Houston. Well, if you route the hose around the facility, the energy just still goes into Houston, the facility powers down for a bit, and then when it’s time to turn back on the hose just flips back over. I don’t think there’s any major performance impacts to the efficiency of the mining facility or anything like that. Also the curtailment periods are generally infrequent. It’s not happening like multiple times a day. It usually is something that happens for a period of time and then stops. You don’t get it very frequently. So that would be the answer to that question. It’s a great question, and I know that we’re going to have to get more data on this. And so I’d urge anybody that is doing demand response to try to keep as much data as you can and try to produce a good report that we use down the line.

Stephan Livera:

One other topic I was keen to touch on, as you mentioned earlier, around hashrate derivatives and financial products being built around Bitcoin mining, and you mentioned this is a priority for Luxor as well. So maybe if you could just touch on some of that, like what kind of products do you see coming in the industry for this?

Nick Hansen:

Yeah, so I think the very first one is going to be a physically delivered forward swap. What that means is I have a bunch of machines that I want to sell forward some hashrate for. So I go and I sell forward let’s say 100 petahash. That’ll be on the order of a couple million bucks. I bring that forward and I get to reinvest it immediately and start buying more machines, buying more infrastructure, et cetera. And then the buyer is incentivized to purchase that hashrate because they think they can get Bitcoin at a discount by doing so. Then we think that as enough volume of these physically delivered forward swaps gets put on and occurs, then eventually you’ll need to hedge that risk. So if I’ve sold some of my hashrate forward and the value of that hashrate goes up, I’ve effectively shorted my hashrate. I can go and purchase what we call a cash settled future to hedge that. So if it starts moving away from me, I can go and purchase a cash settled future that, when it settles, it will help me hedge out the value of my underlying production. We think those are two of the most important primitives for this space. I think there’s going to be a lot of different primitives that come down the line. There’s going to be a forward future, meaning I sell forward hashrate that doesn’t exist yet. So say I’ll sell you 100 petahash of hashrate that comes online May 1st, and I will sell it to you for three months. Your incentivization for taking that risk is that you’ll get a discount on the value of the underlying. And then again, you can start building in really long futures for this and that sort of thing. So I think these are some of the very basic primitives that we’ll need. And you can see this happen in pretty much every other commodities market: oil, energy, corn, pig bellies, all have futures markets, and they also have physically delivered cash settled—all different types of products. And we’re really starting to hone in on what the market demands for these products. It does require quite a large tech stack to do, so actually settling these forward contracts requires a pretty robust tech stack, as well as the ability to do advanced financial reporting and the ability to produce some cogent model at the end that demonstrates, Yes, this amount of hashrate was actually delivered, and here’s the result of that hashrate in terms of Bitcoin or dollars

Stephan Livera:

Really interesting to think about. And maybe as an example, just to make it real: let’s say I’m a Bitcoin miner and I want to do my forward planning, and I need some certainty in my business planning. I might go to you and say, Hey, Nick, I want to sell you my future hashrate. And let’s say you are a trader. You’re taking the other side of that trade. And you’re saying, Hey Stephan, okay, give me a discounted rate on that future hashrate. And I’ll pay you some money upfront for that privilege. And then I take that money and I’m going away and using that to build my mining farm, and that then is the basis for this. And so that’s that first product you said.

Nick Hansen:

Exactly, yeah. And that’s what we call a physically delivered forward swap. And the idea there, like you said, is the buyers incentivize the trader. Speculators are incentivized to do so, because they think they can get it at a discount. And then the seller gets to bring forward a bunch of revenue and reinvest it in infrastructure or buying new machines or unlocking a new project. I think this is going to be one of the most exciting ways that miners can fund their projects in the future. Once we’ve got a more robust system, somebody says, Hey, I’ve got the ability to produce X amount of hashrate. I need this level of capital. Right now they have to go get a very expensive loan, or figure out how to raise capital for that. They could just sell it to the open market.

Stephan Livera:

Really fascinating to see. And I’m curious to see where that goes, and what kind of markets evolve for this. Like, will we have some kind of centralized exchange provider for that, or will it be a more decentralized model where there’s different Bitcoin mining companies who are offering this as a product or some kind of service to match between the miner and the trader.

Nick Hansen:

So the first one, we’ve seen this a bit already. There’s some folks offering a bespoke OTC product around this. Basically it’s a one-off type of contract that’s originated between two parties using a third party intermediary. Someone like BitOoda is doing this. We’ve seen that. I think right now the hold up that’s causing them from breaking through to that next level is the ability to service the contracts, which requires a pretty large tech stack which Luxor fortunately has the ability and the expertise to build. So we’re on the track building these things and really looking forward to bringing these to market over the next—probably won’t be 12 months because it’s also a very regulated type of endeavor—but over the next I would say 18 months, to start seeing something like this happen.

Stephan Livera:

Well, I’m excited to see where that goes. I think we’ve pretty much hit the time for this episode. So Nick, where can people find you online and follow what you’re doing?

Nick Hansen:

So you can find Luxor at http://www.luxor.tech. From there, pretty much all of our product pages are available. You want to yell at me for saying hashrate follows price or any of the other dumb things that I’ve said today, you can find me at @hash_bender on Twitter, or if you want to follow along, we have @luxortechteam and and @hashrateindex. So come give us a follow, let’s chat about this. This is one of my favorite things to talk about. So if you have something interesting, I’m certain that we can find a topic to discuss. So really appreciate it. Thank you so much for having me on.

Stephan Livera:

Thank you, Nick.