Daniel Frumkin of Braiins joins me on the show to talk about calculating Bitcoin Mining profitability and a load of topics:

• Braiins insights dashboard
• Stratum v2 benefits
• Profitability calculator
• Cost to mine calculator
• CAPEX or OPEX costs for mining? 
• How to think about mining ROI
• Updates coming

Links:

• Twitter: @dfrumps 
• Site: insights.braiins.com

Sponsors: 

• Swan Bitcoin
• Hodl Hodl Lend
• 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:

Daniel, welcome to the show.

Daniel Frumkin:

Thanks, Stephan. Thanks for having me.

Stephan Livera:

So Daniel, you’re over at Braiins, and I wanted to chat with you about obviously mining and some of the insights that Braiins is sharing. And there’s this really cool insights dashboard that I want to chat about with you. So do you want to just give us a bit of a background on yourself? How you got into the world of Bitcoin mining?

Daniel Frumkin:

Sure. So I got into Bitcoin originally in 2016, and I was finishing up my university studies at that time. I realized I was much more passionate about Bitcoin than what I was studying, and pretty much anything else I was doing with my life. So I ended up pretty much getting my degree and then shortly after deciding not to use it, and to go travel the world and just try to figure out a way to make money location independently—so make money online. And the way that I figured out how to do that was to start doing copywriting and technical writing. So I was a freelance copywriter and technical writer in 2017. Thanks to the bull market, there was a ton of work in the crypto industry so I got to start writing about Bitcoin, but also unfortunately quite a lot of shitcoins. It was a really cool way to make a living—I was traveling the world and working online. By 2018, especially late 2018, it was starting to wear on me that I’m working on things that I don’t believe in, and I wanted to be working on Bitcoin. I did experiment with [alts], like I had Ethereum and a bunch of other stuff, and 2018 showed me the light, so to speak. So in 2019, I was in that place—early 2019. And I just happened to move to Prague because I had a friend moving there and I was hopping around and thought, Why not? And I was on the Prague subreddit—somebody posted about cryptocurrency meetups. I responded to it—Hey, I work in the cryptocurrency industry full-time. I just moved to Prague. I don’t know anything, but I’m down to grab beers with anybody. Somebody else responded to that: What do you do for work full-time in the crypto industry? I posted a couple author profiles, and then somebody from Braiins DMed me on Reddit and said, Hey, we’re looking for a technical writer. We’re a Bitcoin mining company. And then I looked up Braiins and I was like, Oh, SlushPool—I’ve seen them on the hashrate distribution charts forever. And then I looked into it more and I thought, Oh, this is really, really cool. Like, Yeah, I’m gonna go in and interview with these guys. So I went to the office and yeah, the rest is history, I guess. It’s been a little over three years now since I started at Braiins. And it’s pretty much the dream job for me. I really enjoy it.

Stephan Livera:

Fantastic. And I definitely keep up to date on what’s happening with the Braiin’s blog and the insights dashboard. And so that was partly why I wanted to get in touch with you and chat about it, because I think it’ll be interesting for listeners to just hear some analysis and insights around that. And so do you want to tell us a little bit about this insights dashboard and how it came about?

Daniel Frumkin:

Yeah, for sure. So in 2019, when I started with Braiins, I was just doing technical documentation, but it only took a couple of months for me to end up diverting most of my time to marketing. And so I quickly became the director of content. And so around the halving in 2020, I was producing a bunch of content for the Braiins blog and Twitter and whatnot, and I was using like btc.com and just a whole bunch of sites that—I didn’t want to put our competitors in my content. Miningpoolstats.stream, Clark Moody dashboard—which I don’t mind actually using—but it was just like, I’m using all kinds of external resources for content. And I love the Braiins design language. I love all our websites and user interfaces and everything. So I started thinking like, Why don’t we do this ourselves?—It couldn’t be that hard. And we were pretty lean at the time—the halving is a tough time for mining companies, so we weren’t like eagerly hiring a bunch of developers that I could just say, Hey guys, give me a developer and let’s do this thing. But I did get approval to at least build a proof of concept and see if it caught on. So I had a profitability calculator idea that was basically: I had Excel spreadsheets and I thought it would be cool to have visualizations for this because I change one number in one cell and the Bitcoin mined changes, the profit changes—everything changes. And it’s really cool to see that in a spreadsheet, but numbers changing is a lot different from having a chart where you can see the real real-time impact of different decisions you might make. So the first part of it was just like, Let me put this Excel sheet into a chart where people can visualize it and they don’t need to necessarily understand all the formulas going on behind it. And then there’s also around that time, we were watching the first version of BraiinsOS+ to do auto-tuning on Antminer S9s. And we were saying that, Oh, we can actually bring the efficiency of these S9s from like 88 joules per terahash down to 70-75, which completely changes the economics of whether or not you can run them at different electricity prices. So Jan and Tomas, Jan being the lead developer on BraiinsOS+, and Tomas is our head of product, they built this cool little cost to mine 1 BTC chart in Tableau, and you could just change the efficiency of the miner and see your breakeven electricity price change, see how much Bitcoin you mined, and all that stuff. So then I thought, Okay, this one is really cool, too. Let’s just put these two tools online as a proof of concept and let’s see if anybody uses it. So that was the first iteration, and due to the difficulties of the halving and whatnot, I ended up being the back end developer on that myself, and I don’t know how to code, so that was a very time consuming, slow and difficult process. And then we hired a freelancer to do the front end, and that turned into a mess. And basically like the first year or so of the project was just a slow, difficult learning process for me on working with developers and doing development work. But then we got a couple of part-time backend devs who I had known from previous projects and switched the front end to internal developers. And since then, it’s been awesome. And most of the stuff that we have online now is from that. So, yeah, it’s been really cool the last few months in particular working with those guys because we’re building stuff at a better rate.

Stephan Livera:

Yeah, for sure. I find it interesting just in terms of just keeping an eye on the industry. Just to see, Okay, this is where things are at. This is the hashrate, here are some of the interesting statistics. And I think probably what would be most educational for listeners is actually to talk through some of those statistics. What is it? How is it estimated? What does it mean for the network? So maybe you want to just tell us about some of the key stats, so maybe if you could just tell us—obviously the headline—if there’s one statistic you want to know about Bitcoin mining and the Bitcoin network, it’s the hashrate. So could you tell us a little bit about that in terms of how’s that estimated? And where is that sitting today?

Daniel Frumkin:

Yeah, for sure. So we have two stats for hashrate on the top of the dashboard on Mining Insights. One is the current hashrate and then the other one is a 30-day moving average. So both of those are estimated based on the block times and the network difficulty during the given time period. The current one right now, we estimate it based on 1-day block times, so it actually has huge variations. And that can be just a matter of luck. So basically whether or not there’s actually that amount of hashrate we don’t know, because probabilistically sometimes miners will find blocks faster even if they don’t have more hashrate, and sometimes it’ll be slower. Which is the same reason that mining pools don’t always find blocks. Like for example, you might have 5% market share as a mining pool and still go 16, 18, 20 hours without finding a block sometimes, and that’s just because that’s how the math works. There’s a low probability of that happening, but if enough time passes, those low probabilities eventually come to fruition. So yeah, we measure the hashrate based on: with the current network difficulty, are we finding blocks faster than 10 minutes on average? Or slower than 10 minutes on average? And if it’s faster then that means the hashrate has gone up relative to where it was in the previous difficulty epoch. And if it’s slower then the hashrate has gone down, and the difference between the actual average block time and the 10-minute expected block time is how we can estimate what the hashrate actually is. But at the end of the day, all of those hashrate estimates—regardless of the time period—it is just an estimate, because there’s no way to account for that luck and the probabilities. So one of the cool metrics that we don’t have yet but we will in the future that I’m really excited for is that we’re going to have a real-time hashrate estimate which aggregates the reported hashrates of all the different mining pools. And then if a mining pool is not reporting their hashrate via an API, then we will just use the estimate based on their block times. So that will enable us to have a third way of measuring hashrate, which should be more precise than the other two, but does require us to trust that the mining pools are reporting their hashrate accurately.

Stephan Livera:

Right. And maybe they have an incentive to overreport to say, Look how big we are, or how many people are using us. But anyway, that aside, I think the interesting one—at least as I’ve seen, and maybe you have something to share as well—is sometimes, in Bitcoin’s history, there have been more mainstream news sites or these propagandists kind of sites who basically say, Oh, look how big the drop in the hashrate was, when really they were operating off a bad estimate, basically. And the short version of it was that they were looking at some website that was just doing a really bad estimate and not looking at a long enough period of time to actually get a good estimate of whether Bitcoin’s hashrate had fallen and by how much, or whether it had risen and by how much. Is that something you’ve seen as well?

Daniel Frumkin:

Yeah, exactly. I think like everything in the mining industry, whenever non-mining entities try to analyze this stuff, if they don’t actually work with miners or work in the mining industry, there’s always stuff that’s wrong about it. So that was another one of the reasons that I wanted to build this ourselves is that I think we have a really good understanding of it and we can help share that with people. So having different metrics for hashrate, but also eventually I’m working on now a methodology page where we can explain all this stuff. I think that will help clear up a lot of those misconceptions. And then whenever that stuff gets circulating around on Twitter or wherever, we can post like, Okay, here’s our estimate and this is how we came to it. So, let’s see why there’s such a big discrepancy here.

Stephan Livera:

Right. And so you can provide a more robust estimate given by actual technical experts in the field, as opposed to sometimes when a journalist is looking at it and maybe they’re not a Bitcoin specialist and they’re just picking off a number from some random website they saw and it paints things in an inaccurate light. And I think the other interesting thing is: just as we’re speaking today—this is April 2022—the hashrate is something in that range of 200-207 exahash and we are just broadly where things have been after the events of the last year or so where we’ve seen this huge drop and then rise back up again in terms of Bitcoin’s security as a network.

Daniel Frumkin:

Yeah. And when I was in Miami, I talked to some guys from Bitmain who were saying that they think still 30% or more of the hashrate is located in China. I don’t agree with that, but we know for sure there’s still a good amount of mining going on in China. But I think the real story of the hashrate recovery is probably these very large North American miners scaling up very rapidly, particularly the public miners—Core Scientific, Riot Blockchain, Marathon, Bitfarms et cetera—like, all these mining companies are adding exahashes every quarter. So it’s been a really impressive recovery, but if you were looking at the PRs—the press releases—that those companies were putting out a year ago or even 18 months ago, it’s not all that surprising, because they were putting in massive, massive machine orders back then, and now we’re just starting to see a lot of those machines coming online.

Stephan Livera:

Yeah. And so I’m also curious as well: do you see it like the large miners are maybe more amenable to state capture? We’ve seen this argument, right? That, Oh see, the small miners are going to be more able to support the true values of Bitcoin censorship resistance. But maybe on the other hand you could argue, Well, okay—but at least they’re helping raise the amount of hashrate and in many different countries around the world, so it’s decentralized in that way. So I’m curious if you have any insight or anything you’d want to elaborate on there?

Daniel Frumkin:

Yeah, for sure. I do worry about having large miners and particularly public mining companies controlling too much of the hashrate. I think there’s pros and cons to it. The obvious pro is that some of those companies are investing in lobbying in Washington DC and elsewhere, which can ultimately help out everybody if they educate Congress and educate other legislatures around the world about what Bitcoin mining is and the benefits of it and job creation and economic ability and for energy grids and all the different things that Bitcoin mining can bring—I think that those public mining companies are the best entities to do that. But I also do definitely worry about the flip side of that being regulatory capture. And for that reason, that’s one of the reasons I’ve spent a good amount of time in South America recently trying to learn about the different mining that’s going on here. There’s obviously a huge mining community in Venezuela. There’s a lot of mining going on in Mexico. There’s a lot of mining going on in Paragua,y and Paraguay in particular I think can be a really good hotspot for mining in the future because of how much hydro power they have. And basically there’s the second largest hydro dam in the world called Itaipu Dam in Paraguay—it’s about 14 gigawatts and Paraguay owns half of that. And currently they sell almost all of it back to Brazil because the Paraguayan population isn’t that big and there’s nothing else they can really do with it. So when I visited Paraguay in the last few months, I talked to people actually who work with the energy companies there about like, Keep a little bit more of that energy in Paraguay, sell it to Bitcoin miners, and then they can work with you on demand response programs so that you can ensure that your citizens are never facing energy blackouts. And at the same time, you’re keeping more of the economic value of that energy within your country instead of selling it to Brazil for pennies on a dollar. The thing that’s always kept me optimistic about the future of Bitcoin mining and it being decentralized is that ultimately, very cheap energy is distributed across the world. It’s the energy—the waste natural gas—that’s being vented and flared. There’s a ton of that in the US, but there’s also a ton of it in the Middle East, in Africa, in Russia, et cetera. Obviously there’s still mining going on in China, but if China were to actually make it legal again in the future—which I don’t think is totally out of the question—then we have some real competition between US miners and Chinese miners, which I think would be great for the network. And then there’s a lot of energy abundance in Scandinavia in particular, Iceland, Norway, Sweden, around there, Canada, and all over South America. So I’m definitely keeping a close eye on it. I don’t want to see the market share of public mining companies get greater than 50% of the total network—and it is definitely trending in that direction. But for now, I see a lot of reasons to be optimistic about decentralization. And then another point there is that we’re working on Stratum V2, we’ve actually handed that off to some independent devs that are funded by Spiral, formerly Square Crypto. And that’s making some nice progress recently. So if we can get Stratum V2 implemented on multiple mining pools—particularly the north American ones that are more likely to face that regulation, and get North American mining entities using it as well and constructing their own block templates, which isn’t currently possible but eventually will be once that is merged into Bitcoin Core—then I think we’re in a good spot.

Stephan Livera:

Yeah. And just for listeners who aren’t as familiar, could you just give a bit of a high level for them? What are some of the key benefits there for Stratum V2, if we were to be able to get more mining pools and people on board with Stratum V2?

Daniel Frumkin:

Yeah. The main benefits for miners without considering the decentralization aspect: one, there’s encrypted communication channels between the miners and the mining pools. So the way it works right now with Stratum V1—the original Stratum protocol—miners and mining pools are sending back and forth JSON human-readable data packets. And for example, your internet service provider can see what those data transfers contain. An eavesdropper who just gets into your network can see what those data packets contain. And not only can they see it, but if somebody is really sophisticated and understands what it is, then they can actually steal some of that hashrate without you even knowing it because they would just change the pool account to their own, and the pool has no way of knowing that that’s happened because, as far as the pool can see, we’re just sending payouts to the account that mined all of those shares. And as far as the miner can see, they’re just submitting all of their shares to the pool. But because there can be that intermediary that nobody knows about that interjects in those communication channels, that’s a really serious security risk for miners. And Stratum V2 fixes that by having encryption on the communication channels. And then of course the encryption adds some weight to the data transfers—it makes them larger. But the nice thing with Stratum V2 as well is that the data transfers are no longer JSON format—they’re switch to binary, which is significantly lighter. So even though you’re adding encryption and that makes it heavier, the data transfers themselves are still about 50% lighter compared to Stratum V1 because of that switch from JSON to binary. So data transfer efficiency improves, which would be significant in the case of latency at all, but particularly for miners that maybe don’t have quite as stable or as fast of Internet connections. Those are the two biggest benefits in terms of the performance aspects of mining. The part that doesn’t exist quite yet or isn’t merged into Bitcoin Core quite yet and is being developed right now is the block template proposals. So what that basically is, is: right now, mining pools are the ones that are determining which transactions go into the blocks. So they’re running the full nodes and they’re constructing the block templates and they send those block templates to the miners, the miners hash them and send them back as shares to the mining pools. And what Stratum V2 will ultimately enable is for the miners themselves to run full nodes to determine which transactions go into the blocks that they hash, and then submitting those same shares to the mining pools without disrupting the economics of the pool model at all, because they’re still gonna be paid the same way as they were before.

Stephan Livera:

Yeah. And that’s interesting because there’s so many different complicated aspects of it, because: why do people use the pools? It’s to smooth out that variation so that a small miner can not be sitting there waiting for 20 years to find a block—because they wanna be part of the pool. But then because they want to be part of the pool, they’re currently at the whim of the pool in terms of censorship. And I guess that’s where some of that argument is, that: hypothetically, if censorship or regulation came down to say, No, you must blacklist or whitelist certain coins, et cetera, then this would represent a way that miners could reject that, in a way, because they are forming their own template. And I guess that’s the key argument, although to be fair even today, the argument is also that miners can simply switch pools. So if there’s one pool that’s doing censoring—it’s not costless to do do this, there is some work required—but they can switch pools to one that’s not censoring. And to your point earlier, you were saying as well, is that: because there are many different places on Earth that have cheap electricity, it’s an incentive there to go and set up Bitcoin mining there. And there’s different jurisdictions on the Earth, and it’s harder to try to make this argument that somehow you’re gonna create a successful censoring cartel across all these different countries and different jurisdictions, all of which with different rules—some of these countries are fighting each other—why would they coordinate and collaborate on what the blacklist or the whitelist should be to censor? And can you even make a coherent singular blacklist or whitelist globally? I mean, even that is not really an easy—it’s not a trivial thing to do so. Yeah, but it’s just interesting as well to see these different dynamics play out, and see where the Bitcoin mining ecosystem is going. Because arguably, if we looked years ago, people would say, Ah, look how much hashrate is in China. Now, that criticism is no longer as much of a criticism now.

Daniel Frumkin:

Yeah. Now I want more hashrate back in China—now I want to see that competition between the US and China grow, because that would be great for decentralization. Because, like you said, China and the US are never gonna agree on, Hey, let’s coordinate to stop these OFAC addresses from sending their coins—it’s never gonna happen. So yeah, anytime we see hashrate growing in competing jurisdictions, that’s really good for Bitcoin.

Stephan Livera:

Fantastic. And so from the perspective of somebody who’s looking at this industry and they’re thinking, Okay, maybe I want to start mining myself—and they’re trying to learn a little bit. Some interesting statistics that we can talk about: we’ve got here hashvalue. So that’s satoshis per terahash per day. So could you tell us: what is hashvalue?

Daniel Frumkin:

Yeah, it’s basically just a tool for analyzing the revenue that you’ll produce from mining. We don’t factor in things like the electricity price or anything like that yet—there’s no OpEx yet. It’s just pure, This is how much Bitcoin you’ll mine per terahash of hashrate that you have per day. So if you’re looking at purchasing an Antminer S9 and you want to know how many sats can I mine with this, then you would look at the hashvalue and multiply that by the terahash that you would get from your S9—if you’re running BraiinsOS+, hopefully it’s like 15-16 terahash per second. And then that tells you your sats per day, and that should give you an idea at least of whether or not it makes sense to even purchase that machine based on how many sats that machine costs.

Stephan Livera:

Fantastic. So it can inform our decision around what miners we buy, how many we’re buying. And then also interesting is the hashprice statistic. So what’s hashprice?

Daniel Frumkin:

Yeah. This one gets people confused a lot because price sounds like you’re paying for something, not that you’re earning something. But this is just hashvalue multiplied by the current Bitcoin price. So it’s actually another statistic for revenue but in fiat terms. And we use dollars—you could do hashprice in any fiat currency if you multiply the hashvalue by whatever the Bitcoin price is in that fiat currency. But the thing that people often get it confused with is dollars per terahash, as opposed to dollars per terahash per second per day. And the dollars per terahash is what we’re referring to it as rig price. And that’s the dollars that you spend per terahash of compute power when you’re purchasing a machine. So if you’re buying a machine that has 100 terahash and you’re spending $1,000, then you would be spending $10 per terahash to purchase that machine—and that’s your CapEx. So that’s why hashprice is really, really often confusing for people. And I understand it: it makes more sense for something with the name price in it to be referring to a cost as opposed to a revenue, but that’s how it was set up, so—harder to change it now than to come up with something new.

Stephan Livera:

Right. And also on the dashboard, you’ve got here: fees as a percentage of block reward. So as we speak today, that’s about 1%. What are your thoughts on where that number is today? Do you believe that number has to rise over time? Or it’s not a big deal?

Daniel Frumkin:

I’m not particularly concerned about it right now. It’s definitely not nice to look at that chart and see there were times where it was 5% or 10% of the total block reward. And that looks like, Okay, we’re moving in the right direction for mining to be sustainable 10 or 20 years from now, for transaction fees to be a larger and larger portion of mining revenue. I think it was really when blockchain.com started using SegWit and then the size of their transactions went down—but it could have been many other factors—but ever since around last summer, it’s been between like 0.5%-1.5% for the most part. And that would be concerning to me if it continues through a bull market, but as far as right now, I’m not too concerned about it just because it’s been pretty much a bear market during that time. More like sideways—but the mainstream is not interested in Bitcoin right now and hasn’t been since $60k and since Elon was talking about it and whatnot. So we typically see transaction fees rise and transaction volumes rise a lot during bull markets, when retail investors are getting more involved and starting to use it. And so if we don’t see that happen in the next bullish part of the cycle—whenever that is—then I would start to be a little bit more concerned about it. But even so, daily mining revenue for miners right now is upwards of $40 million per day—maybe a little lower than that, but it’s been consistently above $40 million per day for the last six months to a year. I think that that’s enough to incentivize adequate security. Like, I think we have much more hashrate than we need right now, and that’s something that people need to consider as well is like: transaction fees becoming a larger part of the block reward overall is important super long-term, but how much of the block subsidy right now is funding hashrate that we don’t actually need? How much of it is funding hashrate that like—yes, it’s disincentivizing anybody from even thinking about attempting a 51% attack, so in some sense, the more the better. But I think if we had 150 exahash right now instead of 200 or 207, I wouldn’t be overly concerned about having a 51% attack still—I think that would be adequate level of security. So that’s how I’m looking at it long-term. Like, for now, no reason to worry, but let’s see in a few years or even in the next part of the bull cycle: like, what happens with fees then?

Stephan Livera:

Yeah. Right. And also, as the price rises—which, we’re still in an early phase of Bitcoin, right?—so price is going to rise, and that is also gonna drive up the mining revenue number. Now, okay—maybe the argument could be longer, longer-term, maybe it’s not just gonna keep doubling every four years forever. I mean, who knows? But also, I mean, my friend Pierre Rochard has also made the argument that he believes it’s actually more about: the function of the mining and of the block reward and block subsidy, particularly, is more about distribution as opposed to the security aspect of it. And so he’s almost got like a different argument as well where he’s saying, No, it’s actually better that it’s low fees, that we should be—it’s a sign of scalability and that it’s accessible. All these arguments, there are actually good arguments either side there, so we’ll have to just see where that goes. But I’d also like to talk a little bit about your Bitcoin mining profitability calculator, because I think this one is probably gonna be a very relevant one for anyone who is thinking about getting into Bitcoin mining. So can you tell us a little bit about this calculator? And how to use it?

Daniel Frumkin:

Yeah. It does scare some people when they first look at it—there’s a lot of inputs. The reason that I built it in the first place is because, before this existed, I think the only way to really calculate mining projections—economic projections in a realistic way, long-term—was with an Excel sheet, because most of the mining profitability calculators out there don’t have things like increasing Bitcoin price or decreasing Bitcoin price over time, increasing or decreasing the network difficulty over time, including the upfront cost of your machine, including the residual value of your machine or machines over time. So your hardware inventory—all those different factors. So I wanted to make this calculator include as many different inputs that are actually important as possible so that—yeah, it’s gonna be a little bit harder to understand, but once you do understand it, it’ll be a lot more useful. So we have things there—I could just go through the inputs and talk about how I think about setting them and different things. So we have obviously Bitcoin price and network difficulty—that’s what determines the fiat revenue of your mining operation. Your hashrate and power consumption determines how much you mine and how much you spend on the electricity. And then your electricity price is gonna be the biggest portion of your operating expenses—your OpEx. And then we have inputs for the block subsidy and the pool fee. The block subsidy obviously is just 6.25 BTC right now—that’s the portion of the total block reward which is new coin issuance. And we have it built such that if you do a 3 or 4-year time analysis, for example, you’ll see the halving automatically cut that block subsidy, and we calculate when that halving is supposed to occur based on 10-minute average block times and the current block height and the block height of the halving. So just building in realistic things like that so that you’re not projecting your profitability 4 years out without accounting for the revenue getting cut in half suddenly there. The pool fee, that’s just an inherent part of mining now, that if you want to have stable income streams, then you need to mine with a mining pool. The typical mining pool fee is between like probably 1%-3%, but larger miners can see lower pool fees and potentially even as low as 0%. Average transaction fees in BTC, that’s the transaction fees per block, which adds to the mining revenue. So, the second part of the block reward: you have the block subsidy of new coin issuance, and then the miners are also earning the transaction fees for all the transactions that they include in their blocks. So you can estimate what that will be over the duration of the time that you analyze. I typically put like 0.1-0.2 BTC there because it’s been low recently, like we were talking about. So I don’t want to overestimate that and make the picture too rosy. Other fees is a percentage on revenue—it’s actually calculated right now the same way as the pool fee: it’s just taking out from the initial revenue. So some examples of other fees could be dev fees on firmware like BraiinsOS+, maybe management and hosting fees that you pay to somebody if you’re hosting your miners with a third party. In a future update I’m gonna actually change that so that we have one fee which takes out from revenue and a separate fee which takes out from profit, because we’re seeing a much more common business model nowadays that hosting providers will actually offer a lower electricity price to entice miners to put their machines with them, but they will also take a small profit share from what’s leftover of the Bitcoins after paying for the electricity. So that’s something that I plan to add in the near future, is: other fees on profit, as opposed to only on revenue as it is right now. And then the last two basic inputs are the difficulty increment and the price increment that I mentioned. And that just allows you to change the Bitcoin price and the Bitcoin network difficulty over time to more realistically analyze what your profitability is going to be long-term, because, for example, the network difficulty since 2016 has increased by close to 100% per year. I think before the China mining ban, network difficulty was increasing by well over 100% per year. So that means that with your hashrate being constant, you would be mining half as much—or even less—BTC per day as you were when you initially started, after a year has passed, because of difficulty increasing. So if you don’t factor in a difficulty increase and you’re analyzing your profitability long-term, then you’re definitely going to overestimate how much Bitcoins you end up mining, because difficulty goes up over time.

Stephan Livera:

It’s just not accurate, yeah.

Daniel Frumkin:

Yeah. And then price increment is the same.

Stephan Livera:

Yeah. And so if you had to give an average of that difficulty increase per year, what kind of ballpark are we looking at just historically over the last couple years in Bitcoin?

Daniel Frumkin:

Historically, I think it’s probably just above 90% right now in the modern mining era, which I would say started with the Antminer S9 in late 2016 or early 2017. So, typically I would tell people: if you want to do a conservative case, set difficulty increment to 100% per year, and then if your economics still look good with that 100% per year difficulty increment, then you can rest a little bit easier because—although that has been close to the historical rate—now that hashrate is so large and bringing new hashrate online requires massive, massive infrastructure projects, I don’t see it continuing to grow at that rate or higher than 90%-100% per year, meaning that: I don’t see the hashrate doubling every year going forward the way that it has in the past few years. But what I tell people is more of a, perhaps, the realistic case. Not the best case scenario, but something where, Yeah, maybe this is the most likely to happen—would be somewhere between 65%-75% per year. And that’s based on press releases from public mining companies, just a general understanding of how much hardware is being produced by the major hardware manufacturers, and the rate at which infrastructure is being built and machines are being deployed. That seems like probably in that 70% per year range is going to be pretty realistic, at least for the next year or two.

Stephan Livera:

Excellent. And I think that makes a lot of sense, because in the early years there was so much fast technological development, and now it’s almost like we’re starting to hit almost like certain bottlenecks that are slowing down that increase in the difficulty rate. So obviously, the network is still growing very quickly, but it’s just like that percentage rate is coming down over time—and it’s similar even with price. Like, if you looked at the price for the first 10 years of Bitcoin, it might have been like 200% per year, and now it’s tapering down and maybe now the expectation might be more something like 60% per year or something like that, instead of this crazy-lofty 140% or 200%.

Daniel Frumkin:

Yeah, exactly. And because of that—like, that’s such a huge unknown—that’s the part of mining that’s both scary and exciting if you’re investing a lot of money into it, because those two inputs drastically change the profitability of your potential mining operation. And we really don’t know. Like, difficulty, we have pretty good idea because, at least for the next year or so, we know how many machines have been bought and we know that the average cost of mining 1 Bitcoin for miners today is probably pretty far below the actual Bitcoin price. So as long as price doesn’t totally crash, then miners are gonna keep bringing hashrate online, even if price is sideways. But with price, it’s really almost impossible, still, to say, What is the Bitcoin price gonna be a year from now? Or two years from now? So that’s why I don’t recommend doing one of these profitability calculations and calling it good. I recommend doing like four or five: one where you say, Okay, difficulty is gonna go up by 100% per year, but part of the reason for that is that price is going up by 50% per year. So this is a somewhat bearish case for mining economics, because difficulty’s going up faster than price, but it’s still realistic because price is going up enough that miners will continue to close that gap between their cost of production and the actual Bitcoin price. But then let’s also analyze one where Bitcoin price does really, really well. Let’s say like Bitcoin price goes up 110% per year. And that’s like your high upside case where, even if difficulty goes up 70% per year, your fiat profitability is actually improving over time. And then do the opposite: do the one where price maybe stays flat for an entire year, or maybe it only goes up 10% per year over 3 or 4 years, which is pretty close to a worst case scenario, I think. And meanwhile, difficulty goes up let’s say 30% or 40% per year—how does your economics look then? So that’s what I recommend to people: test out all those different scenarios, do bearish cases, bullish cases, sideways cases. And if your economics look good in most or all of those, then buy the machines and start mining.

Stephan Livera:

Right, and the other aspects of it are maybe even calculating in some probability about machine failure or even like your machine going offline. Or maybe let’s say you’re setting up a facility or you’re setting up an area to do it and you have a delay in how long it takes you to get that set up. So you have to even put in a little bit of conservatism about those aspects of it too. But all those things considered, if you’re careful and conservative about it and you accept, Okay, there’s certain risks—it can still be quite profitable. And I think it’s just useful to think about those aspects of it. Have some different scenarios, think about what are some of the risks. And the risk is also, if you’re trying to get a lot of miners on, then it’s also about, Can you make sure you’ve got enough rack space? Because sometimes that can be the bottleneck. Like, there were times where even during this last year—especially with this whole China mining situation—from what I was hearing is that obviously there was a lot of mining ASIC machines out there available back at that time, but the rack space was the bottleneck. So that can be the other concern.

Daniel Frumkin:

You were hearing correctly, for sure. It was interesting—like, during the halving and immediately following the halving, the big story was, Oh, semiconductor chip shortages, there’s gonna be ASIC supply chain bottlenecks, and nobody’s gonna be able to bring new hashrate online quickly enough. And that totally flipped after the China mining ban. And I think it’s still the case today that the major bottleneck is on infrastructure. And the supply chain issues are present in that case as well. Like, there’s supply chain issues with transformers, fuses, all kinds of different components that go into mining infrastructure that are hard to get a hold of right now. And especially for miners that are building multiple megawatts of infrastructure, it’s not an easy thing to do. And you’ve got a lot of competition because all the miners around the world are trying to do the same thing—everybody’s trying to maintain their market share. So yeah, that infrastructure bottleneck is definitely something to consider.

Stephan Livera:

Yeah. And so let’s just summarize a little bit of that. So maybe if we can put that into some scenarios: let’s say you’re a retail stacker. You’re thinking about, Hey, maybe I want to just do one or two ASICs at home—I’m just gonna run some profitability there. So as an example, you might take this calculator, you might say, Okay, I’m gonna get two Bitmain machines or two Whatsminer machines or something and think about, Okay, what’s the hashrate going to be for those two machines? Okay, punch that in. Okay, What do I estimate the consumption is going to be based on those machines? And then crunch that in, and then crunch your scenarios, and then pull together a number of, Okay, do I believe this is gonna be profitable? And also interesting if you have any insights to share or elaborate around assessing your profitability both in fiat terms and then also in Bitcoin terms?

Daniel Frumkin:

Yeah. The next version of the calculator is going to really highlight doing it in Bitcoin terms, because it is a much different calculation, and ultimately I think a much bleaker calculation in the majority of cases. So what I’m talking about is: if you invest, let’s say, $500 into an Antminer S9, then you have a CapEx breakeven period which is the amount of time it takes you to mine $500 worth of Bitcoin, or to mine $500 worth of profit after you pay for your electricity and other expenses. So that’s where, Okay, now I’ve mined enough—my profit has been enough to offset the initial cost of the hardware. If you do that in Bitcoin terms, it’s a different calculation. So you divide your hardware price by the Bitcoin price, and that tells you the amount of Bitcoin that you’re spending on that hardware. And you can basically think of that as like, Should I buy Bitcoin with this money? Or should I buy an ASIC with this money and then earn Bitcoin over time? So that’s how you set the CapEx in Bitcoin terms. And then the question becomes, How long does it take you to mine that same amount of Bitcoin as you could have just bought in the first place if you had bought Bitcoin rather than buying the ASIC? And that’s your Bitcoin CapEx breakeven. And if you’re a Bitcoiner and you’re long on Bitcoin, then that’s the one that you should really care about. It should be: Can I ever earn as much Bitcoin from mining as I could have just bought? And: Is that the case even after I pay for my electricity bills? One of the nice things, though, with being a home miner or a small-scale retail miner is you can manage your treasury however you want—you’re not accountable to anybody besides yourself. So maybe one of your incentives for mining, for example, might be that you want KYC-free sats. You don’t want to buy it from an exchange and sacrifice your privacy—you would rather mine it yourself, get it from a mining pool. And as long as you take proper measures to ensure your privacy there, then those coins are far more private than they would be if you bought them from an exchange. So maybe that’s one of your main motivating factors to mine at home. And then you would say, Okay, I’m just going to foot the bill for this—the electricity expenses—out of my pocket with fiat, and I’m gonna keep all of the Bitcoin. And that’s another thing I’m adding to the calculator is: one of the inputs now is called the HODL ratio, which is the percentage of the profit which you keep in Bitcoin. So it’s a really interesting exercise to change the HODL ratio if you have a non-zero price increment for BTC. So if the Bitcoin price is going up over time in your analysis, and you’re holding some portion of the Bitcoin, then that will drastically change your end profitability in fiat terms. So the next version of the calculator, I’m going to make it possible to do a HODL ratio on revenue, not just on profit, so that you can say like, I’m going to pay for my electricity with fiat and I’m gonna HODL every single satoshi that I mine, which I think is what most home miners should be doing if they’re mining at a relatively small scale and they can afford those electricity bills without selling. Then, like, Why mine in first place, if you want to sell? HODL it all.

Stephan Livera:

Right. And so you might think of it like, Hey, this is just my little non-KYC premium on one side, and two, you might also think of it like, Hey, I’m contributing to that ideological censorship resistance aspect of the network in a warm and fuzzy feeling sense, let’s say. And I think there’s another complicating—or maybe not just complicating, but opportunity, potentially—using fiat credit, right? Like, you might actually be in a situation where you might not be able to get a fiat loan or fiat credit in order to buy Bitcoin, but if you’re able to get a fiat credit loan in order to do some mining, then maybe that also changes your calculus too, because then it’s like you’re not spending your own money—you’re using the bank’s money or the lender’s money to do this mining operation. Now of course there’s risk—there’s always a risk, but I think it could also change your decision because now you’re just thinking more about the financing cost. And in fairness, this applies more for the medium and the larger miners, but potentially even for a smaller miner, it might be part of the calculation here.

Daniel Frumkin:

Yeah. I’m glad you mentioned that. That’s the most exciting part of the next update, is: we’re actually gonna build in the ability to do debt financing. And I think that that’s something people should think about when looking at these really large miners as they’re scaling, is that being a public company, one of the biggest benefits of that is cheap access to capital. And that’s what all of these public miners are doing, granted in different ways—Marathon’s doing it by just buying as many machines as they can. Riot and Bitfarms are doing it by also building out infrastructure and owning their infrastructure as well. Core Scientific’s doing it with both hosting and self-mining and buying a lot of the machines, but then selling a portion of them to hosting clients. So, looking at all those different factors, it really does make sense for these public miners, although I think Marathon did just raise money and use it to buy Bitcoin. But I think in the majority of cases it’s true that most of these public miners are able to acquire Bitcoin much cheaper by mining it than they could by just purchasing it, because it’s harder to justify a loan for $500 million if you’re just gonna use it to buy Bitcoin. Michael Saylor has done it pretty well, but probably it would be difficult to replicate that business model. Whereas if you were to say like, Hey, we have really cheap electricity, we’re really good at building infrastructure, we have good relationships with hardware manufacturers and other mining service providers, and we can do this really competitively long-term—that’s an easier value proposition to go to investors with.

Stephan Livera:

Yeah. And because I can appreciate where some Bitcoiners and some listeners are thinking, You know, we don’t want the fiat system. But on the other hand, you can think of it also like this is almost using the fiat system in a judo move against itself. So in some ways you could see it like you’re leveraging this system that we think is unjust in some ways—obviously, many of us Bitcoiners think Bitcoin is the answer—we want to Fix the Money, Fix the World, as my friend Marty Bent says. And so of course, with all of these things, you have to think about the risk of it, because you could get liquidated—you might take out a loan and something goes wrong with the mining operation and now you’re out in both ways. So obviously it’s not risk-free, but for people who are able to play the fiat debt game well, they are able to do it in a way that they can stack more sats and use the old system to build the new system.

Daniel Frumkin:

Yeah. If you think Bitcoin is gonna go up 50%-100% per year or even more, then it makes a whole lot of sense to take out a cheap loan and use it to mine. If you can’t use it to just buy Bitcoin, then using it to mine is a great option.

Stephan Livera:

Yeah. So I think we are going to see more financialization and products and companies helping around this. And so it’s just an interesting area, and there’s just so many different areas—I mean, it’s just like Bitcoin, right? There’s all these different aspects to it: there’s economics, there’s cryptography, there’s computer science, there’s networking, there’s all these different elements. But even here, we’re thinking, Okay, people might be thinking about accounting, hardware, electricity costs, tax consequences, interest rates, all these different elements that, when combined, can create a very interesting thing both from an ideological point of view and from just a business and monetary point of view. So really interesting stuff. One other topic I wanted to hit with you—I know we normally go about an hour, but last question: one area that I’ve seen some discussion on Twitter, I’ve seen some people really fighting hard back and forth and I’d love to get your thoughts on this, because people are talking about cost-to-mine calculations. And what we see is sometimes people say, Ah, that’s not being accurate because cost-to-mine oftentimes is not accounting for the hardware cost of buying that ASIC miner. It’s seen like, No, that’s just the energy cost. But I can also hear the other argument, is: Well, it’s just like in accounting standards—they have certain metrics that are done from a comparability point of view. So as an example, EBITDA is done in a certain way: it’s earnings before interest, tax, depreciation, et cetera, because people have different tax rates. So to make things comparable, you need EBITDA, right? So in a similar way, maybe cost-to-mine is kind of like that. So what’s your thought on that?

Daniel Frumkin:

Yeah. We’ve got something coming for that too—glad you mentioned that as well. So the team at Galaxy Digital, the mining team there, put out a report and an Excel model. I’m not sure when it was—it might have been even in late 2021. And they had two different versions of the cost to mine 1 BTC calculation: one was marginal and one was total. And marginal was like, Just subtract your OpEx. Or it’s just your electricity price and maybe your waiver and other small factors like that—you subtract your pool fee and that gives you your cost to mine. But the total one was also including hardware depreciation. And I thought that having both of those is really important because hardware costs, at least for now—like, hardware is not yet commoditized to the extent that you can not think about it—you have to consider the hardware costs as a really, really significant portion of whether or not it makes sense to mine in the first place. So including depreciation of your hardware in your cost to mine gives you a much more realistic estimate of how much it’s actually costing you. So we’re gonna have both of those: those will be calculated from the profitability calculator, and we’re gonna update our cost to mine 1 BTC tool eventually. I’m not sure when that’s gonna be, but that will include both marginal and total cost to mine. And I think either one of those in isolation is not nearly as good as having both of them, because the marginal cost to mine does still tell you, like, Is my electricity price competitive? Like, What’s my breakeven point on electricity price? What’s my breakeven point on Bitcoin price? And those different metrics that, in day-to-day mining, are really useful to know. But the total cost to mine including hardware depreciation is more of like, Okay, what is the Bitcoin price at which I’m screwed if I make this investment into mining and then price goes down or the price doesn’t go up enough? If you’re analyzing maybe over a 2 or 3-year time period then you’ll see your cost to mine is going up, but as long as Bitcoin price is also going up then it’s okay.

Stephan Livera:

Yeah, really interesting stuff. I’d love to keep chatting, but I want to keep the episodes accessible and piecemeal size, so we’ll have to get you back on soon, Daniel. So let’s finish up: if you’ve got any closing thoughts for listeners, maybe any tips for them if they’re thinking about how to assess profitability? And finally, of course, where can people find you online, and Braiins online?

Daniel Frumkin:

Sure. My biggest tip is definitely start thinking about mining profitability more in Bitcoin terms. Hopefully we’ll be able to help out with that pretty soon with the next profitability calculator update and the full mining insights update, although that’s probably a month or two or three away—hard to say. But yeah, definitely start thinking in terms of Bitcoin CapEx breakeven, 100% HODL ratio on your revenue, not just your profit, things like that. Where to find me: I’m @dfrumps on Twitter and basically everywhere. And then also we have two Twitter accounts right now—it’s kind of confusing. Braiins is the umbrella company, the parent company, and then products are BraiinsOS+, SlushPool, Stratum V2. We have a farm proxy coming out and some other different projects that we’re working on. But the Braiins Twitter account is @braiins_systems. And then the SlushPool Twitter account—which I think is the larger account by a good margin—is @slush_pool.

Stephan Livera:

Fantastic. Well listeners, that’ll all be there in the show notes. So make sure you follow Daniel and you check out the braiins.com website. Daniel, thank you for joining me. And I look forward to chatting again soon.

Daniel Frumkin:

Thanks for having me, Stephan. Likewise.