WBD126 Audio Transcription 

Michel Rauchs on The Bitcoin Electricity Consumption Index

Interview date: Monday 8th July 2019

Note: the following is a transcription of my interview with Michel Rauchs from the University of Cambridge Judge Business School. I have reviewed the transcription but if you find any mistakes, please feel free to email me. You can listen to the original recording here.

In this interview, I speak with Michel Rauchs from the University of Cambridge Judge Business School to talk about Bitcoin’s power consumption and its carbon footprint. We discuss mining efficiency, the energy mix, and whether Bitcoin needs to be inefficient to work.

Interview Transcription

Peter McCormack: Well, how are you Michel?

Michel Rauchs: Good thanks! How are you doing?

Peter McCormack: Good! We keep bumping into each other at conferences now, don't we?

Michel Rauchs: Yeah, we kind of seem to bump into each other all the time! It's been quite fun.

Peter McCormack: What did you make of Munich?

Michel Rauchs: So Munich I was presenting at the Value of Bitcoin Conference and so yeah, we essentially ended up being on the same panel, you as a moderator and with Alex De Vries and Dan Held. In the end, I think both of us were kind of like more casual bystanders and just watching the two debating. But yeah, it was a lot of fun.

Peter McCormack: What did you make of the whole event though, cause it was kind of that bringing together of bankers and Bitcoiners and the bankers still seemed quite resistant.

Michel Rauchs: I have to say, I was actually quite surprised how open they seemed to be, to be honest. So I've been to other events like this pretty much two, three years ago and the atmosphere was genuinely completely different. I'm not sure how to actually explain it, but it seems like Bitcoin's here to stay and they recognize it. So now it's really about kind of like, furthering their understanding and see, okay, so what can we do with this because this does not seem to be going away. So I found that quite interesting to actually observe that.

Peter McCormack: Yeah, well I found in the breakout sessions when we were grabbing lunch and grabbing a coffee, people were very open minded and like you say, accepting, but I felt during the debates maybe there was more resistance, but perhaps that may be was just the louder voices.

Michel Rauchs: Yeah, are things well on both sides to be fair? For example, safety net, just like jumping up and so on. But I actually thought it was pretty civilized and of course when they are on stage, they have to kind of support their official position, which is still pretty, let's say conservative, for most institutions. But even given that, I think it was actually fairly balanced and interesting.

Peter McCormack: We had a good debate about energy consumption as you said, but as you noted, we were the bystanders really. I think Alex is obviously very entrenched in his views and Dan is a staunch defender of Bitcoin. So I think those two went to war and you and I just... We got to watch it! But I did try and include you!

Michel Rauchs: Thanks for the try!

Peter McCormack: All right man, so before we get into the recent announcement, for people who don't know you, it would be good if you can just give a background, explain the work that you do at the Cambridge Center for Alternative Finance.

Michel Rauchs: Sure, so I'm some researcher at the Cambridge Center for Alternative Finance, which is a research center based at Judge Business School at the University of Cambridge. So I joined the Center three years ago to essentially launch their cryptocurrency and Blockchain research together with Dr. Gary Coleman back then. I've been leading that research program since two years now. So the key research that we do is really mostly empirical benchmarking studies.

So we reach out to companies and institutions in the space, so it could be exchanges, miners, wallet service providers and so on, but also institutions like central banks and public sector institutions. Then we get or collect data directly from them, which we then combine with the wealth of public information that's available, to essentially provide a bigger picture of where the industry is at and where it's heading.

Peter McCormack: But the Center doesn't just focus on cryptocurrencies, right?

Michel Rauchs: No. So we've got a larger team and we focus on four key areas. So besides cryptocurrency and Blockchains, it's all started with crowd funding and peer to peer lending. So that's really when the Center launched in 2015, that was really the key focus. Since then we've actually evolved to essentially include other types of alternative finance. Alternative finance to us, we really define it as any sort of system, channel or instrument that really emerges outside of traditional finance and that is a sector has really exploded in recent years. So yeah, there's quite a lot of research to do!

Peter McCormack: Especially because of cryptocurrencies? Is that a dominant part of the work now?

Michel Rauchs: No, I would just say it's one of several verticals. So just to give you like an orders of magnitude comparison, for example, when we do these benchmarking studies, we genuinely have around 300 to 400 entities that we studied in the cryptocurrency ecosystem. Whereas our colleagues from the Alternative Finance benchmarking, so most crowd funding and peer to peer lending, they've surveyed I think thousands of platforms in China alone. So it gives you just like a bit of an idea of the scale! But eventually we hope to get to similar numbers as well with the cryptocurrency research.

Peter McCormack: How much of a focus is there on Bitcoin versus alternative cryptocurrencies? Are you completely neutral as an organization or do you have any bias or focus?

Michel Rauchs: Well, as an academic institution we're supposed to be independent and unbiased, right? So we can't really have any preferences towards given assets. So we just look at the numbers and so yeah, we don't discriminate against any cryptocurrency or crypto assets, whatever you want to call them.

Now, it just turns out that Bitcoin is actually still the most used, most dominant cryptocurrency, so it always comes back and you've got probably around five to six more, that the service providers that we survey are always mentioning and that's pretty much it. All the rest is really the long tail that for us, to be honest, doesn't really yield much interest or much... There's not much interest from our side to actually do more research on that, because they're just not really being used at all.

Peter McCormack: So you're not studying Dentacoin?

Michel Rauchs: No, not directly!

Peter McCormack: All right man. Well, so energy consumption is a massive area. You obviously had an announcement, was it last week now?

Michel Rauchs: Yeah, it was pretty much exactly one week ago, so Tuesday last week.

Peter McCormack: Okay and energy consumption is a hot topic within Bitcoin. It's a nice focal point for the haters to throw shade over Bitcoin. So you stated it quite nicely, I think it might have been in your tweet storm, but you brought it down to two points; A, Bitcoin mining boils the oceans and directly contributes to climate change and B, Bitcoin is the most efficient money ever devised and will drive the green revolution. Then you nicely put that the debate lacks nuance!

Michel Rauchs: Yeah, I kind of just want to summarize both extreme sides of that debate, just to show that there is a big gap in the middle and of course this has been a gross oversimplification. Of course there's more balanced voices, but I'd just like to kind of show how far those two positions actually or how diametrically opposed they are.

As a result, you get essentially all these different headlines, that either supports the argument that Bitcoin is directly contributing to climate change or on the other hand that's mostly Bitcoin supporters, who actually say no, it's going to drive to green revolution, which frankly I think at least right now is also farfetched.

Peter McCormack: What you tend to find though is when Bitcoin hits mainstream media, it tends to be something along the energy consumption, that always compares it to a country, "which country have we reached?" There's almost like two measures. 

There's price and then the country we can compare energy consumption to and what you're going to have, is a bunch of people who've maybe heard of Bitcoin, probably have never owned it, think of it as magic internet money and hear that it uses as much energy as Switzerland to make a few people rich and it's kind of misleading.

Michel Rauchs: It is. It's essentially about two main things, right? So we also are guilty of this because we also do a country comparison, but that's one of several comparisons that we do and essentially the main reason why we do it, is that so according to our model right now, Bitcoin consumes around 60 terawatt hours a year on an annualized basis. Now the thing is to most people, that doesn't really mean anything and frankly before I start with this research, like 60 terawatt hours have absolutely no clue, is that a lot, is it not a lot?

Then the first natural thing to do is to look for benchmark rates or something to compare it to. Now the problem with countries, obviously if it consumes as much as Switzerland or any other country, that just sounds enormous. So what we also do on our website, is that just after that country comparison, we also compare it against the major countries and there you basically can't even see Bitcoin, because it's just like a minor blip on the chart, as compared to China, the US and India.

So the idea really is to first put that figure, so that 60 terawatt per hour figure into perspective, so that people can make up their own mind, in terms of where does that actually fit? So to put it into perspective. Now this is different from actually suggesting that it is a good thing that Bitcoin consumes as much energy as a country, because that is more something of a cost benefit analysis that you need to do.

For that you also need to objectively evaluate what the benefits actually are. So what is the utility derived from users, by just using Bitcoin the system or Bitcoin the unit? Now that is a completely different type of exercise that we essentially don't want to go into right now, because it just requires a lot deeper analysis about how Bitcoin is being used and the exact utility that people all over the world get from using it, in a specific given situation and because that's mostly contextual, it's very difficult to generalize.

Peter McCormack: Of course. It's quite subjective because some people, well a lot of people state that holding Bitcoin is using Bitcoin.

Michel Rauchs: Yeah and I would actually agree with that. If you consider it to be a sort of synthetic commodity money as George Selgin puts it and that is being used as a store of value, so a long-term essentially financial investment, sure, then holding it is definitely a use case. Others like Bitcoin Cash supporters would probably disagree and say "no, it's primarily for payments", which is also fair point to say.

The thing is Bitcoin the network doesn't really care what you use it for and that's what I personally as a researcher find so interesting about it. As long as you just comply with the protocol rules and pay a fee so that miners will essentially include your transaction into a block, well you can use it for whatever you like and nobody can prevent you from doing that.

Peter McCormack: Yeah, this is where Dan Held would chip in and say "is the trustless settlement of $1.34 trillion between counter parties annually, with the added benefit of cheaper energy for all, worth the $4.5 billion in mining costs?" He makes a very solid argument that to have a trustless network, a secure trustless network, which is allowing value transfer across the world, it's worth the cost.

He would absolutely state it. One thing I would ask you though, have you done any work to look at, almost the cost per user in the network? So for example, you can look at how much energy that may be the airline industry uses, but then you could also then compare that to the number of people who fly.

Michel Rauchs: Sure, but for that you would first need to have an accurate measure of how many users they actually are.

Peter McCormack: You'd have to do an estimate just to give some kind of ballpark. I think we know there's something like 50 million addresses say, and is that about right?

Michel Rauchs: I haven't looked at the addresses, but I don't think that's actually a really good measure, because you've got all these large concerted service providers, that hold money. So what we did is we reach out to service providers and so we found that there are around 35 million ID verified users at different exchanges and wallet service providers.

So if you were to add those to those that are probably not using... Essential self custodying Bitcoins, we would also estimate it's between 30 and 50 million Bitcoin users right now. But again, this is a very rough estimate. Now the thing is, in terms of the costs per user, it depends on what exactly you're doing. Then we're back again to, what can you use Bitcoin for? You can use it to do small value payments, like paying for coffee and actually people are increasingly doing that with Lightning now.

But you can also use it as a Hodler, you can use it essentially as a financial investment, then you can also use it to do cross border payments or do payments that actually others wouldn't let you do. So payments that you can't do on VISA or MasterCard, but then you can also use it as a public notary. So essentially time stamping documents to do proof of existence to verify the existence and also the integrity of a document at some point of time.

The thing is, these are already four completely different use cases, to have four different types of users and of course also who require different essentially transaction sizes depending on what you do and that is really where fees are being determined.

So how large is your transaction? Now the thing is, this just makes it actually very difficult to just generalize it and boil it down to, "oh, this is the cost per user" because each of these users, use it for completely different purposes and some might actually bring higher utility to the users, than let's say, just using Bitcoin for fun to pay for coffee, just to demonstrate that it works.

Peter McCormack: So essentially we have a free market for the use of Bitcoin and we have a free market for energy. So why do we even care?

Michel Rauchs: Well, I wouldn't say that we shouldn't care at all, but I think that the system's kind of self regulating in a sense. So even if you have some examples of let's say regional governments, that try to create disincentives for miners to actually stay there, so essentially willing to drive them out, by either culling subsidies, raising taxes and so on. Well what happens is that miners would just naturally leave to another place, where essentially you've got low cost electricity in high abundance and that's fine.

So it's like a free market. Now the thing is if every government at the same time would start enacting restrictive measures regarding mining, that would create quite a lot of problems. So this is also one of the main reasons why we actually wanted to launch this project; Cambridge Bitcoin Electricity Consumption Index, because we talk quite a lot to regulators and policymakers, both in terms of education, to be honest educating them about different things, but then also in terms of advising them on some of the policies that they can enact.

So one topic that always comes up is, "so how much energy does Bitcoin consume and what does that do to, first environment, but second also to our local economies?" So households and potential industries that might get displaced by Bitcoin miners. So there seems to be really a lot of misconceptions and also many questions and doubts and I would say legitimate doubts, that haven't really been resolved or clarified.

So unless there was really one reliable measure that provides an unbiased overview over, first, how much Bitcoin can actually consume in terms of electricity, but then more importantly, and that's the second part of our project, what is Bitcoin's environmental footprint? So what does it actually mean for the climate?

Because you can burn coal to produce one kilowatt hour of electricity, which has completely different environmental footprint than just using hydro-power to produce that same amount of electricity and the hydro-power would essentially just go wasted, because there is no ways to transport it over long distances or no way to store it over some period of time.

So I think doing that analysis will alleviate lots of the concerns that lots of regulators and policymakers have and that might prevent them then, from taking too restrictive measures that they might have taken otherwise based on sparse or incorrect information.

Peter McCormack: So you're thinking in terms of regulators who are under pressure to reduce energy consumption, I think the UK has now set a target of being carbon neutral by say 2050, they have to consider things such as Bitcoin mining. If they have this misconception that Bitcoin miners are using fossil fuels, as opposed to renewable fuels, they might actually put the wrong kind of tariffs and taxes on it.

Michel Rauchs: Yeah, exactly and in some cases that might be warranted, given the local situations. So if Bitcoin miners actually displaced other industries that provide also utility to the region, just because they're competing for the electricity, then you need to consider again and do a cost benefit analysis. So what do we actually get from Bitcoin mining and is it worth that it displaces these other industries.

So for that we would probably need some more studies, but as far as I can see right now, this really hasn't happened in a lot of places, so it's not like an immediate concern, but it's still something that policy makers and regulators are essentially required to look at. Then you've got, of course politicians, who are always driven by short term goals, which is also totally understandable. So exactly as you mentioned, you have these headlines saying that "oh Bitcoin money for criminals, consumes as much electricity as a country like Switzerland."

Well that's quite an easy target then for at least short term measures, whether in the future that really brings you a lot, I would doubt it. So our goal is really with this website, is to provide an unbiased estimate of how much it consumes and then in the second phase, we really go deep into the environmental footprint discussion, to actually really quantify what Bitcoin's impact on the climate actually is.

Peter McCormack: So you can really stack up the FUD then? You can say that we are boiling the oceans and killing polar bears to help finance terrorists?

Michel Rauchs: You could say that!

Peter McCormack: There were a number of other tools out there. So why did you decide to create your own one? I mean Alex de Vries has his own one, The Bitcoin Energy Consumption Index, so why another one?

Michel Rauchs: So essentially, yeah, you mentioned pretty much the only alternative right now, that is also based on a real time model and that is essentially updated every... I don't know how often he updates it, but at least on a daily basis I would say. But apart from that, you've just had a few studies, but genuinely they use different methodologies and they essentially just provide a one time snapshot and given that price plus hash rate functions are quite volatile, that is not very helpful.

So we figured that there needs to be a real time model for that. Now looking at Alex de Vries model, I think it's an interesting approach. So it's a top down model that actually looks at estimating miner revenues and then using an assumption that says, "they use x% of the mining revenues are spent on essentially just electricity", so to cover electricity costs and I think right now he uses 60% for that and then he goes on to find an average kilowatt hour price globally.

From there he essentially calculates the estimated energy usage. So that's an economic top down model. So our model is based on an approach developed by Marc Bevand in a blog post in I think early 2017, if I'm not mistaken. So that is a bottom up approach.

So it's also based on economic factors, but it starts really with the actual hardware, so the mining equipment that exists and then tries to build in some assumptions in terms of who is using what equipment, at what time and then using the mining equipment efficiencies, so essentially their specs, we do a bottom up analysis. So I can go into details of how that exactly works if you like?

Peter McCormack: Yeah, definitely and I'd like to know a bit more about the theoretical upper bound and the theoretical lower bound as well, because there's quite a quite a range in there.

Michel Rauchs: Yeah, so as you mentioned, we added theoretical just two days after, because we noticed that the big gap between these two bounds was actually causing quite a lot of confusion and of course people don't read the methodology. So we just wanted to make it absolutely clear that this is just an absolutely purely theoretical floor and ceiling.

Peter McCormack: Well the methodology is quite detailed as well. So I've read through it this morning and it isn't the case that you can just read it and go, "okay, I instantly get this." You almost need to sit down with a piece of paper and a pen and go through it just to... I could probably spend a day, just trying to get my head around the model and like you say, most people aren't going to do that. They're just going to look at the number and go, "here's my conclusion."

Michel Rauchs: Which is absolutely fine. But I think adding the word "theoretical", hopefully clarifies some of this. So speaking about the theoretical, the lower bound is probably the easiest to calculate. So that is really based on a very simplistic assumption, which is also very unrealistic, hence the theoretical nature of it and that just assumes that all miners always use the most efficient hardware that's available on the market. So just by doing that, you just essentially need to have the hash rates and that's it.

So you just calculate the hash rate using the most efficient equipment out there and that's how you get to this absolute floor, in terms of how much electricity Bitcoin would consume, if everybody used the most efficient hardware. Now that is very unrealistic because of course for lots of different reasons, not all miners use the same equipment and they may also have or actually in many cases, still have some old equipment that depending on the Bitcoin price and on the current hash rate, might still be profitable to actually mine.

So for that, then you can use the same approach for an upper bound. So we need to provide a ceiling, which is essentially a boundary and for that, you then assume that all miners always used the least efficient hardware. Now the thing is of course you can take that to the extreme and say, "well, if everybody used a CPU, where would we be?" Now the thing is that you quickly get really to a figure that consumes more electricity than the Earth actually produces, so that is not very reliable or useful ceiling either.

So the thing is what we did, and there's essentially based on Marc Bevand's initial idea, is that then we take all the equipment that exists. So essentially have a list of more than 60 different ASICs and then we model the economic lifetime of each of these, just in terms of mining revenues, minus the electricity costs. So leaving out all the operation costs like maintenance and so on, as well as any capital expenditures. So we model the economic lifetime of each of that equipment, just to see at any given point in time, and we do the models since November 2014, so any of these days, when is the given equipment still profitable, given Bitcoin price and the hash rate.

Then we assume from that, "well we just used the least efficient hardware that is available on the market, but still profitable at least in terms of the electricity purely." So of course that changes pretty much every other day, because Bitcoin price and has rate is quite volatile, but so that's how you get to this theoretical upper bound. Now because hash rate often lags a bit behind price, so like recently the last two weeks, Bitcoin price is spiking, then that doesn't necessarily immediately translate into higher hash rate.

So difficulty stays the same, which then means that mining revenues are a lot higher than they used to be, given a constant hash rate and that also means that a lot of previously unprofitable hardware, now becomes profitable again.

Peter McCormack: It comes back online.

Michel Rauchs: Exactly and this is why this range... As soon as Bitcoin prices really spike, you see the upper bound figure just skyrockets completely. But then again, that's also a quite unrealistic assumption, because lots of miners used the latest generation hardware and the most efficient ones. So this is also just a purely theoretical ceiling essentially.

Peter McCormack: Right, so in some ways the trend is also probably more important than the actual numbers, because you can't get to the actual numbers. So it's more like you start to follow the trend?

Michel Rauchs: I wouldn't necessarily say it's really a trend. The main reason we put those two numbers there is just to delineate the boundaries really of the asset. So you could say, well it's very, very unrealistic that Bitcoin consumes at a total more than this particular upper bound figure. So just going to the index right now, how much is it? It's a very, very high number. So in terms of power, it's 25 gigawatts power capacities, which would translate into 186 terawatt hours per year, which is just enormous but also completely unrealistic.

Peter McCormack: I still feel like, when following the charts, the trend kind of follows the price.

Michel Rauchs: Oh yeah.

Peter McCormack: And it was like when we were in Munich, like Alex's point, which I think was slightly misleading, but when he started to predict future value of the Bitcoin network, if it was to be worth trillions, then what would the impact then be on hash rate and therefore power consumption.

Michel Rauchs: Yeah, so the thing is we purposefully don't include any predictions, because I just don't think you can predict either the Bitcoin price or Bitcoin's hash rates in the future and that is essentially the two main dynamic parameters that go into the model. So what he essentially did, is more like a thought experiment, which essentially says that, "okay, according to my economic model, there is always an equilibrium. 

So if price goes up, mining revenues goes up." My model assumes that also the hash rate will automatically go up, so that in equilibrium, marginal revenues equal margin of costs. So that essentially there is no profit that can be made anymore, because the market is so efficient, so will always get to this equilibrium. Then of course using that particular logic then yeah, sure. If the Bitcoin price goes up to a million, then there will be an equilibrium, which essentially then means that Bitcoin will just consume pretty much the entire energy produced on the planet.

How likely that is, I'm not exactly sure, but I think still trend wise, it's not totally unreasonable to expect something like this, just because for miners, as long as it's profitable to mine, they will just add more hash rate. Now there's practical limitations to that, with how much electricity do you have access to? What are the electricity costs?

And especially also hardware equipment, like ASICs don't just grow on trees, so they need to be manufactured, produced, shipped, which of course causes significant delays and lags in between. So that's also things to consider. But overall I think that if the Bitcoin price rises, then the electricity consumption will automatically wise as well necessarily.

Peter McCormack: Yeah and I guess over a long enough timeframe, with innovation in ASICs and as the machines become more efficient, you could see an increase in hash rate, but with a lower proportional increase in energy consumption?

Michel Rauchs: I'm not so sure about this and the reason for that is really that over time...

Peter McCormack: Really?

Michel Rauchs: So the thing is Bitcoin needs to be inefficient right? So in order to work it needs to be inefficient and it needs to have high costs in order to add new blocks, which also means that it needs to be very expensive to change past blocks. As ASICs really get commoditized, so essentially in terms of acquiring them, really that cost over time will really drop significant and it has already dropped significantly.

Then that means that over times on long time horizon, the only essentially cost that will remain, like substantial cost, will be electricity costs. So even if ASICs get more energy efficient, that will just mean that people will buy more ASICs, because there needs to be a cost that needs to be there and it trends towards being electricity costs being really by far the main cost driver. If that doesn't work anymore, well then there's no incentive to essentially secure Bitcoin.

Peter McCormack: Okay, that's fair enough! Did you do any work looking at the number of transactions? I know again going back to Alex de Vries' Bitcoin Energy Consumption Index, the comparison was about, from my calculation, that one Bitcoin transaction equals around 350,000 VISA transactions. Now I don't think that is a fair comparison, but they still did the work.

Michel Rauchs: Yeah, so that's again the main problem about comparing Bitcoin to something else, right? So we also compare to countries, for example, which frankly doesn't really make much sense. It's more about putting things into perspective. Now, of course, just to put things into perspective, you can compare it to a payment system like VISA, but then you need to be clear that those serve completely different value propositions.

So it's really just about putting the figure into perspective, but not suggesting that Bitcoin does exactly the same thing as VISA and therefore it just is a hundred thousand times less efficient, which means that VISA is the best alternative. This is kind of like my reading at least when I go to that website, and this is something that essentially we wanted to change on our website, but some of the feedback we actually got is that right now we do similar comparisons and it doesn't become immediately clear that we essentially just want to put figures into perspective.

So we are really planning to essentially split that comparison section into two. So first, having something to put figures into perspective, but then a second section about alternative uses, for things tend to use or produce a similar value proposition as Bitcoin, using those four different essentially use cases that exist and then directly comparing those to essentially Bitcoin's electricity consumption.

But also using things like, untapped hydro sources that are in remote regions that can't be easily accessed and essentially go wasted right now. So comparing how much of that particular electricity could essentially power Bitcoin, without causing any interruption or disruption to existing industries or other uses. So I think it's worth spitting those things up and generally in comparisons they tend to get conflated, which creates a lot of confusion and I would say misperceptions about what Bitcoin actually does.

Peter McCormack: Yeah, so another issue I have with the comparison to the VISA transactions, is that I believe that the energy consumption secures the Bitcoin network. To only compare it to base chain transactions is quite misleading, because it doesn't take into consideration off chain transactions and layer two transactions, which will obviously be a much higher number if you calculated them all. They're not taken into consideration, therefore it just feels a little bit, I don't know, unfair?

Michel Rauchs: Yeah, exactly. Also the thing is, the throughput of the Bitcoin network, so the number of transactions, is completely unrelated to the electricity that's being used by the network. Those are two completely separate things, but doing that comparison actually suggests that they are very closely linked, which means that people thing that okay, if if Bitcoin now were process not 4/5 transactions a second, but 1000, that means that the electricity consumption will just jump up by a factor of 1000s as well, which is not correct because those two are completely unrelated.

But then also as you mentioned, one transaction can actually hide different semantics. So you can, for example, timestamp billions of documents using standards like OpenTimestamp for example, to do proof of existence that wouldn't really be noticeable by anyone just looking at the operator and field. But yeah, as you mentioned as well, there's off chain transactions or payments... Well even one single transaction can have multiple payments.

So you have one company that essentially pays out, let's say a mining pool, that pays out to different hashers using one transaction, but there are hundreds of payments in there. Another argument essentially for saying why the energy cost per transaction doesn't really make much sense, the metric, is I think it comes from LawrenceMT, that's his Twitter name, that every transaction, every proof of work essentially secures the entire chain.

So with every additional of block, you get another layer of security on top, which secures all the UTXOs, that are in previous blocks. So this is also something that in traditional payments systems, we don't really see, because it's something that doesn't really exist in there. So this is also a factor to take into account.

Peter McCormack: Okay, so listen, what is the carbon footprint of Bitcoin? Is it something we should be concerned about?

Michel Rauchs: So that's something that we purposefully left out at this stage, just because we do not have enough data on that. The reason it goes back to what I was saying before is that, you can burn coal to produce one kilowatt hour, which is fundamentally different from using hydro or any other type of renewables to produce the same amount of electricity. Now in order to estimate or quantify Bitcoin's environmental footprint, so the carbon emission of carbon dioxide emissions, you would need to actually know exactly the energy sources, so the energy mix of each hashing facility to actually get to, let's say a reasonable estimation.

Now obviously this is pretty much impossible, because we actually don't know how many hashing facilities there are, where they are based, what energy sources they use. But that would be really the only, I would say, accurate way to do that. Now there's been, of course, lots of different estimates from some other academics, but Alex de Vries does it as well and there was actually a pretty good paper recently by the Technical University of Munich, who also attempted to do this by actually just looking at pool locations and IP addresses of mining equipment that were connecting to those pools. Now there's quite a few issues with that approach, but it was still an interesting methodology I haven't really seen before.

So Alex de Vries, I think he estimates by just looking at, for example, one of the mining maps that we produced and then essentially extrapolating from that where miners are based and then looking at the statistics of in that given region, say in the Chinese region like in Mongolia, what is the average electricity mix or energy mix and there happens to be mostly coal and then that gets extrapolated to essentially get to a global carbon emissions figure. The problem with that approach is really that, what we can see in practice is that most miners are essentially off the grid.

So they are mostly really based in very remote areas that essentially are completely off the grid. So you can't really store electricity for long time, you can't transport it over long distances and they have special electricity deals with the local power stations. It often happens that those power stations are in remote areas where there is an excess, but of course that doesn't appear on the grid.

So if you just use the grids figures to extrapolate, then you won't get to those figures either. The problem is, in my opinion, the most recent estimates, they don't really take this factor into account and as a result they overestimate carbon emissions. But again, to really empirically prove that, you would essentially need to do a bottom up analysis, by really going through the hashing facilities, getting the exact energy mix and then extrapolating from that.

Peter McCormack: That sounds too bigger a task. It's almost impractical to do?

Michel Rauchs: Well of course you can always approximate it. So yeah, you never get to the exact figure... So this is part of the second stage of our project where we will essentially build on our mining map that we had in our last report, so the second global benchmarking study, where we already identified around 120 different hashing facilities all over the world.

For some of them, they are actually very open about their energy mix and so we essentially plan to make that map available on the website, so make it an interactive version and actually collaborate together with miners and power stations to essentially maintain a list on an ongoing basis. So not really the individual locations, because we don't want to be responsible for any miners getting their equipment seized or doing things that will run counter to op-sec.

So we want to do it on an aggregate level, so mostly region based, which would still give us enough information to actually to, not only then have an approximate carbon estimation, but also would enable us to essentially get more information about how centralized mining actually is, both in geographical terms but also in terms of how many operators there are, how many facilities do they operate.

So it would actually give us a lot more information about the health of the Bitcoin network and the security of the Bitcoin network as well, as a kind of like nice byproduct.

Peter McCormack: Have you taken a look at the Coinshare's report, where they estimated around 77% was from renewables?

Michel Rauchs: Yeah, I'll have to say personally that I find that number to be very high. I was very surprised to actually hear that figure, but it's interesting so we're going to actually have a call with the one of the core authors tomorrow to discuss this and see how we could potentially collaborate and also so we understand better what their methodology is.

So just using our own data that we got by talking directly to hashing facilities, it's more about in the 20-30% range. So there's quite a big gap between those two figures and I, personally, have no idea which figure is closer to the actual reality. I would say it's probably something in the middle?

Peter McCormack: So maybe about 50% then?

Michel Rauchs: Exactly! So right now there really are two estimates that have a very huge gap in between. So it's a bit like the lower bound and upper bound for the index itself.

Peter McCormack: Well whilst we have a lot of FUD, we do have a history of Bitcoiners tending to either post rationalize criticisms or go with the most positive news possible. So I've heard many Bitcoiners respond to the energy consumption potential FUD, depends how you look at it, because I think it's subjective and bring out the Coinshare statistic, "well 77% is renewable". So obviously that's potentially misleading itself?

Michel Rauchs: Yeah, exactly. That's exactly that lack of nuance in the debate from those two extremist sides and let's say some mainstream economists and commentators in the news do exactly the same thing, but just using the other side, so saying, "oh Bitcoin mining is powered by 60% or 70% of coal in China, which is just devastating for the environment." Frankly, I think that figure is just as far away from the reality than a 70%+ share of renewables.

Also with all due respect to the Coinshare's report, so I don't want to say that the numbers are completely wrong, it just personally to me seems very high given the conversations that we have with some miners and just the fact that renewables are intimate in nature most of them. So even like during the wet season, miners actually move in China to Sichuan and other provinces just because the dams are just full, so there's lots of abundant hydro-power.

But in the dry season that essentially dries completely up, so they need to go back to coal or to other fossil fuels. So it's also seasonal and so if you have like a general renewables figure, I think that needs to take into account both seasons.

Peter McCormack: Do you think that Bitcoin gets a disproportionate amount of attention for energy consumption, because one of the really interesting things on your website, because I didn't know this. If somebody asked me, I'd have my suspicions, but idle home devices in the US could power Bitcoin for three years.

I mean that's just idle home devices and then the other thing I noticed, particularly recently, it's like if I'm coming into a city late at night, the amount of office blocks which are just lit up, where you know those rooms are empty. So you've got idle home devices, but I'm assuming globally there's a huge amount of energy waste?

Michel Rauchs: Oh exactly and that was the main reason essentially for the comparison page. We don't want to do any value judgment ourselves. We just want to provide a neutral platform for visitors essentially to make up their own mind, in terms of is that now a lot of electricity or not? For that you need to present both sides of the debate as well, using examples to illustrate that.

So going back to the country example, it's something that's been used quite a lot by the media to essentially demonize Bitcoin. But then on the other hand we also wanted to show, well there's actually a lot of ways, as you just mentioned, in conventional ways of using electricity, that could power the Bitcoin network for quite a long amount of time or at least multiple times given the early consumption.

So providing both sides of that debate I think is very important to bring some balance to the debate as well, because both supporters as well as critics, genuinely tend to use the comparisons that sued their narratives or fit their narratives and that doesn't really contribute to healthy debate either.

Peter McCormack: Okay, so that is really interesting and I'm probably going to get a kick from the Bitcoiners for saying this, but I do have a slight concern over energy usage of Bitcoin. How much security is enough? I know you can't put a limit on it, because as the network grows, it is a market, the miners will buy more equipment and they will put more online as long as they can make money from it. But I do think we get to a point where, surely there is enough security? I just don't see anywhere limiting it and I'm not very keen on Alex de Vries' suggestion that we should move Bitcoin to proof of stake either.

Michel Rauchs: No. So that's also something that actually lots of, well I wouldn't say necessarily academics, but let's say researchers in general, who suggest that proof of stake is just orders of magnitude more environmentally friendly, so it has proven itself, so let's switch to that and it would solve all our problems. Now the thing is, it just really assumes that proof-of-stake has the same security assumptions and the same security levels as proof of work and right now there is absolutely no evidence to support that claim.

Now personally, if essentially it turned out to have the same security guarantees, I think we should always use the most efficient way of doing things if all else is equal. But right now, frankly there is no evidence that supports their claim, which means that proof of work still remains the most trust-minimized essentially way, of essentially reaching consensus on the global level, where you don't need to have parties participating in consensus processes that need to be identified.

Peter McCormack: Okay, that's pretty compelling. I think that's one of the other problems. I'm probably going to sound very mean towards Alex, but I think one of the problems I have is that I guess your tool and your approach is very neutral. You've said it's very neutral.

You've recognize both extremes of the debate and you think there's more nuance to this, but I feel like maybe with the work Alex is doing, he's coming from the little bit more anti Bitcoin energy consumption, less anti Bitcoin, and it feels there's maybe more of a bias in the work he's doing and I felt that actually at the debate too.

Michel Rauchs: I don't know to be honest, I don't want to get into that. I have a huge deal of respect for Alex and for what he's done, for essentially having built the first real time index. I think his approach is really interesting because it's just a fundamentally different methodology and as you recently you mentioned, so he benchmark our best guess estimate compared to his one and essentially the figures are not really that much different. So there's like a 20%-30% gap, which I think is very reasonable for essentially two fundamentally different methodologies.

Now in terms of the way he maybe presents some of the findings, I had a long discussion in Munich with him in person about using these energy cost per transaction metrics or comparing it to households, because those are two completely different things. So the energy that's being used by Bitcoin could not power, let's say a few million US households just very simply because that electricity or that energy is in very remote regions in different parts of the world, so it can never power US households.

So in that regard it's not a very good comparison either. But on the other hand, I accept the criticism that he essentially made on our index and that essentially those comparisons are not a lot better than his, because they don't really compare apples to apples. I think that's also fair point to actually say, and so this is why we really want to split up those comparison pages into things just putting it into perspective by presenting both sides to make it look large or small, depending on what view or what side you are you on the debate.

But also then having a separate site, where we've tried to compare likes for like, so essentially Bitcoin to gold production or something like that, because we consider it to be more a store of value as well or also to other waste of electricity, that could essentially be used for different things. So that is going to be in a second version of our project.

Peter McCormack: What's the general feedback been from the tool then?

Michel Rauchs: I have to say we've really been overwhelmed, so it's been very good, especially the first two days. I think the Tweets and the thread that I did on Twitter is probably my most popular ever. Also just lots of media interest and also from mainstream journals and mainstream news sites, but especially of course the crypto community.

So yeah, we have received tons of messages on Twitter. Also a lot of people actually bothered to really have a look at the methodology and get back to us via the contact form on our website, providing suggestions, comments, feedback and so on. So it's been really great so far.

Peter McCormack: Yeah and I'll share that out in the show notes because I did find reading through the methodology quite interesting. How much work went into this? You must have been working on this for quite some time.

Michel Rauchs: Yeah, so we're actually a very small team, so it we've been three people working on this and not even full time because we also have other research projects going on. It's been probably in the making for like 2 to 3 months. Then again, we're standing on the shoulders of giants to quote Isaac Newton, in the sense that we didn't really come up with that methodology ourselves.

So we were looking at all the previous estimates that existed, combining all these different methodologies together and then really picking the best practices. So essentially the main methodology was really developed by Marc Bevand and so yeah, credit where credit is due.

Peter McCormack: And you've alluded to some things coming soon in the future. So what can you tell us? What's coming when? What should we keep an eye out for?

Michel Rauchs: So I don't want to give any precise timelines because again, there are lots of different research projects. I don't want to create too high of expectations here, but the first thing we're going to add is really splitting that comparison page to provide a fair comparison, because I think some of the criticism we've received is actually is quite valid in that regards. Then a major update will really be the first step into actually quantifying carbon emissions and so for that, we're going to start with that live interactive map of hashing facilities by regions.

So for that, we'd love to actually hear from miners that operate hashing facilities and would like to contribute to that project or local power stations, other researchers, especially in areas where we have language barriers or difficulties. So yeah, in case someone's listening to this and would like to get involved in a project in some way or another, just feel free to reach out to me.

Peter McCormack: Cool! So how do people find out more and how do they get in touch with you?

Michel Rauchs: They can go first to do websites; cbeci.org and we have a contact form there. Otherwise can just message me via Twitter, that's actually easiest probably. So that's @mrauchs. I'm actually also fairly active on Twitter and generally respond to most comments I get. So yeah, that's the two best ways I would say to reach out.

Peter McCormack: What event am I going to see you at next? What have you got coming up?

Michel Rauchs: Well, fortunately for now I have got an event pause, because I need to finish the report! I've been traveling so much over the last 1.5 months. So yeah, it's good to actually get some focus to get some work done. But yeah, otherwise looking forward. I'm going to be at Scaling Bitcoin. I'm not sure if you're going to be there?

Peter McCormack: Is that in Israel?

Michel Rauchs: Yeah, Tel Aviv.

Peter McCormack: Yeah, I am thinking about going to that. So the next couple I'm thinking of going too is that, and I'll be at the HoneyBadger Baltic.

Michel Rauchs: Yeah, so that's just immediately after Tel Aviv.

Peter McCormack: Yeah, so I'm thinking I'm just going to head over to Tel Aviv and then go to Latvia for that afterwards. I think that's going to be quite good.

Michel Rauchs: Yeah, probably same for me. So I guess we're going to bump into each other again!

Peter McCormack: Definitely! All right mate, well listen, thanks for coming on Michel. I know it was short notice, but I appreciate you coming on. I'm going to get this out as a bonus release this week, so it's going to be a three show week, but I appreciate you coming on and doing this.

Michel Rauchs: Perfect, well thank you very much for inviting me and yeah, it's been good discussion!