On The Brink with Castle Island - Apolline Blandin (Cambridge Center for Alternative Finance) – Estimating Bitcoin's energy footprint (EP.85)

Episode Date: June 1, 2020

Apolline Blandin, cryptocurrency research lead at the Cambridge Center for Alternative Finance, joins the show. In this episode:  The Cambridge AltFin research agenda The motivation behind the Cambridge Bitcoin Electricity Consumption Index How they determine Bitcoin's electricity consumption How Bitcoin's energy usage can be understood in context The genesis of their new geographical mining estimates Why high granularity estimates of mining location are so important to determine its energy mix How and why Bitcoin hashpower migrates on a seasonal basis Why there is so much spare hydro capacity in Sichuan and Yunnan Surprising country-level findings The most defensible academic approaches Apolline has seen so far Why the per-transaction cost estimates of Bitcoin are misleading The prospects for wind, solar, and natural gas for bitcoin mining

Hello everyone. Welcome back to On the Brink with Castle Island. I'm Nick Carter. A seemingly never-ending topic around here is Bitcoin's energy consumption. I fan the fires of this debate recently with an op-ed in Coin Desk on the topic. I cited some new data coming from the Cambridge Center for Alternative Finance. Now, if you haven't heard of it, the Center produces some of the best research on cryptocurrency. They have this amazing benchmarking study, which comes out every year. And the next one is due in August, so stay tuned for that. And they've also devised this really fantastic index of Bitcoin's energy consumption, which is a credible and neutral source, unlike some sources out there. And just recently, they've released a mining map, which is an estimate of the actual

geographic location of miners, which is information that we never really had before. And what this allows you to do is to more precisely estimate the energy mix and get a better view of Bitcoin's carbon footprint, which is something that Akkadian. has gotten pretty badly wrong in the past. So to talk about how to estimate Bitcoin's energy consumption and its carbon externalities, I invited Apuline Blanda onto the show. She is the cryptocurrency lead at the Cambridge Center for Alternative Finance. So let's dive right into it.

Brought down by bad mortgage investments, Lehman, which has 25,000 employees, will be liquidated. The federal government loans American International Group, AIG, $85 billion. This is a different kind of market, and the Fed is asleep. The federal government is stepping it to stabilize Fannie Mae and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis. The Bank of England has pumped 75 billion pounds more into Britain's ailing economy with a new round of quantitative easing. You print a couple trillion dollars and all of a sudden people start to worry.

So out of this worry, we have something called the Bitcoin. Today we have a very special guest. We have Apoline Blondin, who is the cryptocurrency research lead at the key. Cambridge Center for Alternative Finance. Welcome to the show. Hi, Nick. Thanks for having me on. Yeah, I've been meaning to have you on for a while now. You guys are doing some really, really fantastic work. And I don't say that lightly. I was just on a call with a friend yesterday about market sizing for the crypto industry, and I referred them to your estimates that you came out in your benchmarking study last year, because I consider that to be the best.

and most reliable estimate of the number of crypto users worldwide. So I rely on your work a lot, actually. Well, thanks for the kind of words. I mean, yeah, that was definitely a figure that has been picked up quite a lot. Yeah, it's interesting. And we're a huge fan of your work as well. And yeah, regular listener of the podcast. All right.

That's great. You know, we don't have that many guests that are also fans of the podcast. So it's amazing. Yeah, you know, the thing is, I guess nobody ever before. you had thought to go and ask every exchange how many users they had or maybe the exchanges didn't answer them when they asked? So actually, yeah, the methodology behind this figure is a little we detail it in the report, but it's a little bit different. So we do ask exchanges. The problem is you have, you know, different terms. So exchanges might refer to like accounts, number of

wallets, number of identified users. So actually what we did is like collect public data by the number of accounts open and then we asked service providers the share of accounts that was actually KYCed users. And then we applied this ratio or like this percentage to the numbers of account we've been able to identify. Because generally, otherwise, yeah, they don't consistently provide this figure. So you have to do a certain amount of extrapolation, I guess. Yeah, we do. But yeah, we often found that actually asking for the share of active users and the share of KYC or identity verified user is usually more efficient. And then we collect manual data through other sources.

And I guess you can also do some extrapolations based on web traffic, you know, assuming that there's a consistent ratio there. Yeah, that's something we've been, yeah, we've been considering as well. The problem is the different kind of web traffic providers are not have pretty pricey access. and some who are free actually don't necessarily integrate any app or like mobile access

visitors. So yeah, but we're trying to look for some metrics that could be supplementary of what we get from the survey. So your benchmarking report, you were telling me before you've

concluded the survey portion, you'll actually be publishing that later this year, right? Yeah, so we just closed the, so we have two surveys, one for the mining industry and one for service providers.

And so we closed the survey about 10 days ago. So now we're in the process of cleaning up the data, analyzing it. And yeah, we expect to have it published by end of August. Hopefully, yeah, that would work on. Well, I won't ask you to reveal any of the secrets just yet. I'm waiting excitedly for that headline number for the user accounts on exchanges, so no pressure. Yeah, actually, it's funny because after we had our first chat,

I came back to the team and told them, well, I think we will need to get these users figures again this year. So, yeah, we have some people starting to gain to this actively. Your tenure as research lead is, it's only recently you've taken charge the organization. Is that right? Yeah, so it's not taking charge of the organization. So we have several research teams within the organization. And so my team is looking at cryptocurrency, but then we have other teams looking at other alternative finance verticals. So we would have digital lending, crowdfunding, and then we have one team doing

research on regulation. But yeah, it's recently that I became the research lead for the cryptocurrency team. Before that, I was just kind of working with Michelle, so Michelle, who's the previous research lead, so kind of helping out with the whole research, managing of the survey, writing of the reports, etc. And can you tell us about your career trajectory to date? and how you came to be interested in cryptocurrency research. Yeah, sure. So I've been with the Center for Alternative Finance for about two years. So I joined early in 1718.

And before that, so I lived in China for my studies. And I was already kind of doing research about mobile finance and like financial inclusion or generally kind of witch out penetration in China and how people were using it. So I kind of came across Bitcoin mining at this time, so it was back in 2016, but didn't really pay attention because that wasn't really within the scope of my research. And then it came back to France and I worked with the French Association for Tech Startups and investors. And so that was set 2017. And there were a couple of companies, part of this association, among which like Coin House, which is the former Maisonub Bitcoin, which is one of the big exchange in France. and ledger. And so that was also the time of, you know, the ICU boom. And so there was a lot of, like,

talking on at the political level, it was in French and in France and at the European Parliament. So we were doing a lot of, like, education about this topic. But then I kind of came and realized that I missed the research aspect quite a lot. And, yeah, then applied to Cambridge and, yeah, get, got the job and, yeah, enjoying it since then. I know the objectives might be top down, but what's your objective? in terms of managing the research agenda, you know, on the cryptocurrency side of the center. So I think generally the mission of the center, and it's not that much stuff down, actually, like we have a lot to say on how we want to direct the research agenda.

And the general mission is to feel the knowledge gaps about these different industry that we're looking at that are emerging in on which there is not much data. So really like the crowdfunding and digital lending team at CCAF has been very like a real pioneer in terms of collecting trading volumes and data about the crowdfunding and digital lending sphere. And for cryptocurrency, I guess, I mean, the objective is to do the same, train to unearth data and an earth metric on this evolving industry and providing access to more reliable information. And the interesting thing is that the center is really at this crossroads of different audiences.

So we talk a lot with industry participants, but also with regulators, policymakers. And so we are really at the position. of like bridging the knowledge gaps and making these different audiences to talk to each other. So if the research reproduces useful to the industry but also relevant to the general public, other academics, then I guess that's what we will call a success. So we'll talk about one area where there certainly is a knowledge gap, I think. You've done some really good work here with the C-B-E-C-I. your Bitcoin electricity consumption index, a bit of a mouthful.

What was the reasoning behind creating this model? Because other electricity consumption indices exist for Bitcoin. So why jump into the fray? So I think the idea was really to address this debate about Bitcoin mining that's going on. So on one side, you would have all the detractors and critics of Bitcoin who would argue that Bitcoin mining is burning the planet and boiling the ocean. And that will often come from publics that are less knowledgeable about the space. And then on the other hand, you would have a defensor of Bitcoin who would argue that Bitcoin mining is actually driving the green revolution. So, you know, surely there's some

nuances in this debate. So you're telling me there was a middle ground there. But yeah, so the idea was to provide like a bit more reliable figure on this because we're not really satisfied with what we saw in other papers. And also we wanted to provide like a light in. where people can actually refer to it on a regular basis. And how did you go about actually deriving this figure? So we took what we call like a bottom-up approach, and we actually adapted a methodology that was developed by Mark Beavent back in 2017, I think. And so the underlying idea of this index is that miners will only run equipment as long as they're profitable.

And so the first step, first of all, was to identify all, um, mining equipment available to the market and then calculate the profitability and then draw modelized kind of the profitability threshold. So that's that's that was the first step. Then, because obviously you can't get an accurate figure for how much electricity is consumed by the Bitcoin network. What we had to do is to draw like a lower bound and an upper bound. So like a minimum absolute minimum of electricity being consumed and an absolute maximum. And in between these two boundaries we draw our best estimate. So the lower bound then rely on the assumptions that all miners will run the most efficient hardware available to the market.

And they will have like a very efficient PUE, so very efficient data center. And then the upper bound is relying on the assumptions that miners will run the less least efficient but still profitable mining equipment. And that they will have like a PUE of 1.2. And between these boundary, then we drew our best estimate. which is based on the assumption that miners will use an equal weighting or equal weight basket of profitable hardware equipment. Yeah, that certainly makes a lot of sense. You know, what's interesting is that we now have estimates over the relative share of hardware being used to mine bacon, which is kind of a new thing.

Yeah, actually, that's totally. So before actually we saw the Coin Matrix report about S9 representing it. It's 23% of the networks based on the non-distribution. We did see like a lot, like some research done by people trying to identify the market share of the different hardware equipment. And so there was people from the Technical University of Munich who published this paper back in 1018 looking at IPL feelings and trying to get a more. accurate distribution of the different equipment.

But then we actually tested their distribution against our equal distribution. And then we also run like a simulation. And it seemed like the weighting didn't really affect the final output. So that's why we went with these very simple assumption. But yeah, I think as we get more data, then we'll be able to have a more accurate estimate. Yeah, I didn't even know that Coin Metrics was working on that, so I was surprised by that, which is kind of funny. So we brought on a new researcher who did some amazing stuff in terms of looking at the randomness of nonces, which is very revealing, which is amazing to me that Bitcoin has all these hidden signals in publicly available data. But yeah, it's a wide bound that you've set here.

But it makes sense because older equipment has to consume more electricity to create one, one tarahash, right? Mm-hmm. Mm-hmm. Yeah, I think there's been a bit of a transition lately as the S-9s are reaching the end of their tenure. They're not really profitable anymore. Yeah, definitely. And that's why, like, we kind of keep updating this list of hardware equipment because as well as, yeah, some hardware equipment are no longer profitable, but new are coming to the market. So we really need to keep it up to date.

So these numbers seem very big, but how do you contextualize them for people? I mean, you know, five, so your estimate right now, your best guess is 5.9 gigawatts, which corresponds to 31 terawatt hours on an annualized basis. I mean, what does this mean? Is there anything you can set it equivalent to? Yeah, so that's a good question. And that's why we kind of created this comparison. page. I guess it's poorly named, which would probably have, yeah, name it differently.

But the idea was to provide other figures for people to control and textualize this tower hour of these estimates of the Bitcoin electricity consumption. And so there is obviously no perfect comparison that you can draw because there is nothing similar to Bitcoin. So instead, we just took like figures that are familiar for people. So like we look at the total electricity the productions globally, but also like we look at this country comparison, which again, is not a fair comparison, but it's just to enable people to put this figure into perspective. And we also provided kind of a few fun facts because if you argue, for instance, that, you know, Bitcoin is wasteful, then you would look at the amount of electricity that is consumed by always on, but

inactive home devices in the US. And then you realize that these could power. the Bitcoin network for like 4.6 years. So yeah, we actually try to come up with fair comparisons in looking at systems or elements that would have similar attributes to Bitcoin. But interestingly enough, there is no reliable research or study or papers published on how much electricity is consumed by these different comparable elements. So like we look at the banking system, or at gold mining, and there was very little research done at this level. Yeah, I mean, so Bitcoin is the 51st most energy-hungary nation on Earth. But you're right.

I mean, it doesn't mean much, you know, like I don't have an intuitive sense of how much energy as Pakistan consumes. You know, I guess ideally you would have some industry ranking and you could see how it's stacked up against other heavy industries. Yeah, I fully agree. And generally, we've also been looking for such papers. But what seems to happen is that there is no study based on the electricity consume, but there is a lot more studies. There are a lot more studies down about the carbon footprint of these industries. So what we're hoping for is that as we progress through this research,

we would actually be able to compare the carbon footprint of Bitcoin with the carbon footprint of other industries. So yeah, so on that topic, you've done another really fascinating and edifying piece of research. which is of course the mining map, which is totally new to me. I mean, I'd seen some estimates before. I know coin shares had done some estimates. They went out and talked to miners. But you've now managed to muster up real-time estimates of the relative share of hash power on a global basis, which is just staggering.

I mean, I was so pumped when I first saw this. So can you tell us about how this project began? how you managed to pull it off. Yeah. So actually, we've been tracking kind of the geographic distribution of mining activities since the beginning of our research since then 17. But our methodology back then was to actually identify individual mining facilities through public sources.

And I believe that's kind of the same methodology that's contrary to using. Now, the problem with this is like through these public sources, you'll identify facilities that are willing to advertise their activities. And so there was a natural bias towards, I would say, like, North American or Northern Europe facilities. And it was much harder to identify facilities in Sichuan or Tsingjiang in China. And then what happened is we came across to that paper by Stoll and his colleagues at the Technical University of Munich. And they used the

servers of, the location of pool servers to identify, to have a rough estimate of how mining activities was distributed across the globe. Now, the problem is obviously they didn't have enough granularity in that data because pool servers locations are disclosed at the continent or like the region level. So you would know for like roughly for North America, Europe and China. But then we thought truly like pools have access to these small, granular data that we need because they know the IP addresses of miners connecting their pools. So we started approaching pools early September to 19. And I have to say, like, I guess we

contacted like, yeah, eight or nine of them. And generally, the responses was very positive. I mean, only a few kind of completely refused to talk to us. But generally, they were very, very positive. But some of them kind of highlighted that actually they're not collecting that data. They were not processing that data already. And so for them to provide us with the data, they would need to change some of their internal processes. So that's why it took kind of a bit of time to talk to them. And then what really helped as well was that we attended, we presented some of our research in a summit in Changdu in Sichuan early October. So then we met with a lot of these pool operators.

And so, yeah, so that was, you know, small steps, but then we talked a bit more and then we kind of set up the whole agreements and the way they will share the data, the type of data we need. And so we have three pools who contributed to this first version of the map, but we're hoping to increase the number of posts participating in future iterations of the map. Yeah, and I'll post the link, but it's really, really, really fantastic. You know, for the first time, I think we have really good numbers on where Bitcoin is actually mined. And no surprise, it's mostly in China.

We all already knew that. But there's also a bunch of other places with material fractions of hash rate. USA is number two. Let's go. Russia, Kazakhstan, Malaysia, Iran has a surprisingly high. fraction according to your estimates. And I guess the other thing is that you will need to fill out the sample of mining pools. So I guess you're hoping to grow that sample so you can characterize more and more hash rate, right?

Yeah, precisely. So we have those three pools, so BTC.com, pulling and via BTC who provided us with data. Then we keep talking to other pools and hopefully we can get your second version by, yeah, maybe. by maybe September, October this year. Because really what we wanted is to have a full year of data so we can actually see all these migration patterns happening. So yeah, really hoping to get more pools contributing. Yeah, so I guess the way this works is the pools

take the geolocation of the IP addresses of contributing miners and they give you that data in anonymized form and then you aggregate it. publish it here. So they, yeah, they provide us with a country level and province level data and then we, yeah, aggregate for all the pools together. But province level is pretty good because in terms of devising the energy mix, that's much more informative than just using the country level energy mix, right?

Yeah, fully agree. And so we're hoping to get that level of granularity for other larger countries like Russia United States, Canada. So let's see what we can get. Because right now you just do the administrative unit, the sub-state administrative unit divisions for China, right? Yeah, exactly. But so it's quite informative because you can actually see

these seasonal migrations occurring, which is something that, again, we knew was the case, but we didn't really have hard data on it. So can you tell me about what you found in terms of the migrations of Hachpower? Yeah, I think there are a couple of interesting findings in this map generally. So, yeah, the migration patterns is definitely one. I mean, we knew that was happening because of the rainy seasons.

The rainy season in Sichuan and Yunnan runs from kind of May to end of October. And so if you look at the share in September in October on the map, then you see that Sichuan is about like 37%. But then as the rainy season ends, a lot of miners are relocated. the operation in other regions like Sintiang or Inno-Mongolia. And so then you see that Sichuan is losing a certain share, so it moves from 37% to 13% in November, whereas, like, Xinjiang is increasing.

So it's like 15% in roughly in September and October, and then it increases to 34%. And the reason why these miners are kind of moving during the rainy season is because there is a lot of hydropower And there is a lot of excess being produced in Sichuan during the rainy season where the electricity cost is like below three cents per kilowatt hour. So that's very profitable for miners to mine there. And so people know that Sichuan has lots of hydro and lots of otherwise curtailed hydro potentially. Is that also the case with Yunnan?

So Yunnan, yes. But then provinces like Singtang or like Inomongolia would mostly be coal-based. So knowing the relative share of Hastery on a per province basis is really critical because you have this process of going from energy consumption to carbon footprint and to devise a reliable estimate of the carbon footprint. You need to know how the energy is being produced in the first place. Yeah. And so that's the approach, for instance, taken by Coinscher.

So they look at the renewable. in the different provinces based on data provided by Morgan Stanley. And so that's interesting because you actually get information or you get data at the provincial level. Another approach which has been taken by the people at Technical University of Munich, they directly take the carbon emission factor of the different countries. But the problem is that data, this carbon emission factor, is only provided at the country level. And so what they do for China is they kind of adjust it to take into account the fact that there is a lot of hydropower coming from Sichuan. But it's not very quite transparent of how they adjusted it. And unfortunately,

there is no provincial level data on the carbon emission factor. So we're actively looking because the data they use comes from the international energy agency, which only provides country level data. So we're trying to look whether the Chinese government is also providing some more like provincial level information about the carbon emission factor. So for for Sichuan, can you explain a little bit about why there is this abundance of apparently cheap hydropower? What is the local factor that explains that? So China is very much obsessed with its energy security. So they developed a lot of hydropower infrastructure in Sichuan, notably, because there is like the three gorges.

And so during the rainy season, that produced a lot of like oversupply and excess of energy. And actually what you've seen recently, it was like I think a couple of weeks ago, we saw a city in Sichuan who was actually calling for miners to come and install some of the facilities there to tap into the excess of energy. So there's been this theory that a lot of the mining in China has been done on a kind of a black market basis, you know, with the federal government not necessarily aware that there's mining occurring. Is that kind of consistent with your experience or does it seem to be kind of an aboveboard activity? I don't know, like I don't have really like any evidence on that.

I've heard similar, yeah, similar suspicion. But then, like, for instance, like, yeah, as I mentioned recently, where we actually saw, like, the official authority calling for miners to come in and settle in Sichuan to use the extra amount of energy available. Then you would, yeah, it would be quite surprising that, yeah, they don't know at all about what's happening. They definitely know to a certain extent. And, yeah, they seem to start leveraging.

So I guess China has gone back and forth on the legality of mining and their kind of stance towards the Bitcoin industry. Right now, what is the atmosphere like between miners and the government? Yeah, I haven't followed that very closely, but what it seems like is that, yeah, mining is being promoted differently than any other activities around Bitcoin. so like, you know, exchanging your Bitcoin for yuan is still banned. But mining seems to, yeah, benefit from a different treatment. I guess it could be a point of pride that China hosts such a large percentage of the network. So in terms of expanding this in order to obtain really reliable estimates of the energy mix, what fraction of hash rate would you like to characterize?

I guess if we can, you know, if you look at the remit, remaining pools that haven't contribute yet. I mean, there are still a lot of like large pools that could contribute and that will easily increase the sample. So like if you think of the F2 pools or like slush pool and pool, so we can easily get, you know, to like 70% of the of the hash rate if they are willing to contribute. So I guess if we reach that, that will already provide us with a more reliable picture of the distribution of mining activities. Yeah, and some people were saying that you had a bit of selection bias because you were relying on Chinese pools and maybe certain Western miners wouldn't connect to those pools necessarily.

Would you agree with that? I mean, are there countries that are maybe being underweighted in your sample because you don't have some of these more Western-facing pools? Yeah, so definitely I think what came as a surprise was the fact that Canada is so small. It's like below 1% I think. And that can come from a sample bias. Now the fact is that some of these pools, I mean, among the three pools that contributed, I mean, some of them are like quite global.

They do have minors from all around the globe. And because no matter what, you know, miners stays, you compete a lot on fees. So China's both still have a more profitable fee. Now, it's true that some public companies, in particular in North America, would be not, from what I've heard, are not necessarily willing to, you know,

worked with Chinese mining pools for a transparency reason. So they will go with European and Western polls instead. So definitely if we can get other Western polls to contribute, we would probably have a more reliable view. But I wouldn't expect this to change, you know, tremendously. So were there any, you mentioned Canada being underweighted relative to your expectations. Were there any other kind of country level findings that surprised you when you put this together? Yeah, there are actually two other findings that are interesting.

So one at the country level that you can't really see with the share. So the amount of hash rate has increased, obviously. But when you look at the amount of hash rate coming from China, this has remained constant. So that would mean that the growth that we would have ordinarily expect from China is taking place elsewhere. So like in other region, like Kazakhstan, Iran, Russia. So that was quite interesting.

And the other surprise came from the share of Malaysia. I think we really didn't expect that much of Hushrek coming from there. So some people have speculated on why that is the case. Some people have said, for instance, that some of the large hardware manufacturers have warehouses in that country, and so that they would run their machines before shipping them. So that could be one potential explanation. Yeah, you know, in my mental model, I'd never thought of Malaysia as a hotbed of Bitcoin mining, but according to your figures, they're doing over 4% of the Bitcoin network.

Exactly, yeah. So in terms of the papers which discuss Bitcoin's carbon footprint and energy consumption, there have been some pretty good ones and there have also been some downright crazy ones. I'm thinking of the University of Hawaii paper, which is one of the craziest things I've ever read. But so, you know, given that you're so close to the academia on this, do you have certain papers that you actually think are fairly reasonable, all things considered? Yeah, I think I agree with you. I mean, I think the first paper was published back in 2016 by scholars at the University of Ireland. And then since then, I think you've seen more over like a tens of like 12 or 15 papers being published on that topic.

And there is, so again, I think the one published by scholars at the Technical University of Munich is quite interesting because they really try different approaches. And they've been very transparent in their methodology compared to other papers who do not necessarily kind of highlight the limitations. But generally, these paper would suffer from a couple of things. So one is not having access to reliable data. So, you know, about the geographic location of mining activities. A lot of them will also kind of have this misconception, on which it's difficult to have a very exact figure. But a lot of them would assume that Bitcoin mining is being powered by electricity from the grid. But what we know is that a...

certain amount of mining is coming from off-grid activities. So like, for instance, when mining is stepping into surplus of hydropower energy. So I think that's another limitation. And generally also, like, some disregard, this energy mix that is quite important when you want to assess the carbon footprint or at least, yeah, the carbon footprint of Bitcoin. Something we've also seen in the academia is this kind of these strange approach where they they back out the like per transaction cost in energy terms. And they take that really literally. And then they try and scale it up to like a visa scale.

So and it leads to really wacky estimates. So basically, you know, you say, well, Bitcoin consumed X amount of energy on on Y day. And it did, you know, 300,000 transactions. And so, you know, the pro rata energy share of each of those transactions is, you know, is, you know, a certain figure. And then, you know, the methodology is to say, well, if Bitcoin wants to scale up to be a global payments network, you know, it's a linear function in terms of the energy consumption. So then it would consume like three earth's worth of electricity or something, which is, I remember seeing that in a paper and just chuckling at how. kind of naive that was. Yeah, and I actually think that half of the paper I reviewed had this comparison of like the energy

transaction, which is, yeah, as you say, very misleading because we know that the energy cost does not increase with each marginal transaction, right? And furthermore, like you could have like one big coin transactions might actually kind of contains hundreds of payments of individual addresses or like being a settlement, like, being a settlement layer for second layer payments or like being useful time stamping data. So there is much more, there's a lot of hidden semantics, I guess, in this Bitcoin transaction, but a lot of people not familiar with the nature of the system are not able to kind of, yeah, see and capture.

Yeah. And the other thing is that, you know, Bitcoin's energy ally is much more a function of the subsidy as opposed to the per transaction fees right now. Precisely. Although that's changing. But the good news, I guess, for the people that are worried about energy, is the subsidy drops every four years. So, you know, assuming the price doesn't double every four years, which I'm not rolling out,

that should be net, not good in terms of the energy footprint. Yeah, exactly. So, you know, in like the final calculus, we've done a lot on methodology, but what do these numbers mean to you? You know, would there be a point where you found, that Bitcoin was consuming so much energy and creating such a large carbon footprint that you would decide that it's not worth it for the network to exist? They're like a threshold you've set in your mind where it becomes the outlay is too costly,

too significant. That's an interesting question because, yeah, definitely. I mean, I think as a researcher, I don't really care. You know, I'm just studying this thing and kind of trying to provide more data and more transparency and better understanding of how this works. But it's interesting because generally I think that really is the crucial challenge to a lot of this discussion around Bitcoin energy consumptions. What I kind of encounter is when you talk to people about this topic,

they often disregard your research because they have a preconceived idea. They already have an opinion about Bitcoin and whether it's bad or good, and then they will receive your research and your data as a way to serve their narrative. So I think it's, yeah, it's a big challenge, I would say. But it seems to, a lot of these debate seems to come down to that question of how do you see it? Do you think it's worth it or not? And I wish people were, you know, openly, openly kind of engaging with the research to be able to have a fair opinion on it. It's quite unique as an industry because Bitcoin is so transparent by virtue of the fact that you can literally see the hash rate on hundreds of different websites.

and then from that hash rate, you can back out an energy estimate. That's not the case for any other industry I can think of. You know, there's no central data repositories telling you how much energy fishing consumes or forestry or aluminum smelting even. Yeah, precisely. And that's why this, again, going back to this earlier point about doing a fair comparison with Bitcoin is not really available because you don't have as much transparency and as much research being done on that topic.

It's the curse of transparency, I think, because there's such a glut of data around Bitcoin that people can get very alarmed about it. But we don't even really know what other industries consume. So one other interesting topic that Bitcoiners like to talk about is potentially bringing online more exotic sources of stranded energy, not just hydro, for instance. So the big,

topic of conversation that comes up a lot is that the vented methane at oil wells, which is either vented or flared very rarely these days is it actually captured and compressed and put into pipelines because it's expensive to build that pipeline infrastructure. And so in theory, that's kind of a pretty neutral impact, all things considered, because something had to be done with that gas anyway. So has this come up in your research at all? Have you found that this has been moving the needle, or is it still, you know, very de minimis? Well, I mean, it definitely got a lot of visibility recently, and that's a question we cover in the survey.

So we do ask miners about their energy mix and the main source of power. And it seems pretty minimal for now. I mean, a lot of, yeah, I mean, the main source of power is like between hydropower and, like, I would say as far as I can remember from the data. And hopefully we can provide more insights into this when we release the report. So yeah, I think it's definitely an interesting topic, but there is still a question about how scalable this is and what amount of mining can actually be powered by this. Yeah, alternative source of power.

And what of the possibility of wind and solar to power? Bitcoin mining. I mean, the capacity factor is low. Is that prohibitive? Yeah, similarly, so when we look at renewables powering mining, it's, again, mainly hydropower. We do have some people who said that, yeah, they're using solar or wind energy. I'm not an energy expert in any way, but I mean, these source of power are much more intermittent. And so we've heard like some anecdotes or like some cases where someone was building a solar plant in Texas and they were kind of incentivizing miners to come and install their equipment there. But at the same time, they had to guarantee that they could provide

the miners with access to electricity when the, well, during that nighttime, essentially. So, yeah, I'd be curious to see how this evolves. Yeah, I guess that's the problem is that A6 have this kind of fixed lifetime where they're economically relevant, because, because hash rate is generally increasing, the technology is getting better and more efficient. And so it's almost not worth it if you're only going to be running your A6 part time. You want to be running them, you know, 90% of the time or more. And if the wind's only blowing half the time, then you're kind of out of luck. You need to be pulling in grid energy, right?

Yeah, exactly. And another point is that, you know, one could argue that Bitcoin mining has emerged in a era of like energy abundance where there's been a lot of like excess of energy coming from hydropower potentially like other renewable sources but because of you know global warming and green energy has been put so high on the political agenda what we will see is like governments and other actors will invest a lot of resources in developing like you know grid level storage solutions and also like improving the transmission lines between the difference the the site of production in the site of consumption.

So that's what you see, for instance, in China is that they've deployed a huge networks of ultra-high voltage transmission lines between Singtyang and the greater Shanghai. And so as this happens, I guess, you would have less surplus of energy available to mine. Which is, I guess, what does that mean for Bitcoin? There'll be less opportunity to mine with these otherwise un-economical sources of energy. Yeah, I guess one, yeah, you could argue that then they will turn to like fossil fuels. Because the same kind of consciousness of the Chinese government to secure their energy position has also led them to open more cold plants, right? And similarly to these hydropower plants where a lot of energies produce in excess, you also see that as see that in the cold plants as well.

So mine is because I'd well just go and tap into this. excess of coal energy. So just saying that what we're seeing now with the energy mix might actually change in the future. Well, I wish you the best of luck with continuing to characterize it. I think it's really vital work. How should people follow you and follow your work? I guess the most efficient way would be to just through Twitter, so at Apolline Blondon. And otherwise, you have the cbc.i.org when you can reach out to us, we have a contact page. I'll link to everything in the show notes for sure. Well, Apolline, it's been really informative. Thanks so much for coming on. Thank you, Nick. It was a pleasure.