hi and Welcome to our webinar on blockchain energy consumption where we will explore one of the most pressing topics in the world of blockchain technology I'm pleased to introduce our speaker today our Chiefs of Staff Juan ignasio Juan is renowned for his expertise in ESG and will share a science-based uh methodology for calculating the energy consumption of blockchain networks during this webinar we will delve into several key areas including the if Energy Efficiency of proof of State systems the intricacy of Bitcoin carbon accounting and the role of renewable energy in Bitcoin mining this is an excellent opportunity to gain valuable insights into the environmental impacts of blockchain and explore Innovative approaches to sustainable technology we encourage you to engage with us throughout the webinar by submitting your questions and thoughts in the chat which we'll answer at the end of the presentation thank you for joining us today and without further Ado let's begin our Deep dive into into blockchain energy consumption thank you very much we're recording all right hello everyone it's a pleasure to be here as you heard my name is Juan I'm the chief of staff uh D Science Foundation I've been working on blockchain for a number of years and before that I used to work in sustainability um and we have been doing a lot of research search here um on on blockchain energy consumption blockchain cover footprint um blockchain sustainability in fact I encourage you to see our initiative DT Earth uh at dt. Earth um and we want to share some of these uh findings some of our methodologies with you uh see what their reaction is hopefully in the future maybe we can even do some follow-up seminars so without further Ado uh let me walk you through what is going to be the structure of the seminar so um first of all U we're going to provide some context about why this is important explain a bit of our work and set out the key questions that we are going to be trying to answer um now the key aspect of this seminar is that I'm going to be positing that um it is good to organize your thinking when you're thinking of the sustainability of crypto uh into different typ types of systems types of objects of study that you want to to look into so we are going to be sharing some of the framework that we ourselves are using to study and the scientific methodologies that we use to understand um blockchain carbon footprint and blockchain energy consumption so for that reason we're going to provide first a basic model we're going to modify this model to understand how things are with proof of work we'll we'll delve into a special case on storage BAS blockchains um the the case of layer twos and then we'll live we we'll move on to some other special cases and we will extract in the key lessons so let's take a step back and look into the context why we are trying to answer these questions why we care about all of this you have probably heard um lots of things about the carbon footprint of uh blockchain technology you have probably heard that uh Bitcoin consumes as much energy as Ireland or as Argentina I have here a screenshot uh some websit saying that one Bitcoin transaction is equivalent to watching 120,000 hours of YouTube so this raises some questions first of all is this true is this a meaningful thing to say um so we'll be diving into that you'll also see here um this um comparison uh analogies saying well if bitcoin's energy consumption were the bush Khalifa then the ethereum prior to this event called the merge uh would be equivalent to a tower Piza and uh after the merge equivalent to a screw uh continuing along um the path of metaphors um our chairman Dr paasa said not so long ago at the world economic Forum that the global network of hia could function on less energy that that used used by a regular household so the question is how we calculate these things and can we U replicate this further can we expand this can we cover more crypto assets and how and what we can do to understand the results in and to just to generate good knowledge about this um this is all more important in the context of renowned criticism renewed criticism against crypto assets Green Piece has set up this um initiative called change the code trying to um to force the change in the Bitcoin ecosystem and everything is becoming much more relevant now with um the emergence of one uh new comprehensive set of regulations in the EU so-call Mika or marar which stands for markets in crypto assets regulations so Mika is a regulatory framework that covers crypto Assets in uh so many regards it's very comprehensive but two uh um dimensions of this that really interest us are the fact that by the 1 of January 2025 uh crypto asset service providers are going to have to publish sustainability indicators on all the crypto assets that uh they are dealing with in addition the issuers or other parties of the crypto assets will need to publish white papers and these white papers will have to have a number of sustainability indicators so the science um it becomes more important than ever they it is important to make figures comparable to make them scientific um to make them fair so that's why we're going to be exploring some of these things so we have been doing work uh on crypto energy consumption cryptocarbon footprint um mining with renewable energies low carbon crypto activity as I mentioned our initiative D the Earth before and also we have been working on policy submitting responses to the Europe the European Securities and markets Authority esma um and also the US White House Office in of Science and Technology policy which published a report not so long ago heavily citing some of our work we're going to be trying then to share uh in as little time as we have a bit of the framework that we use to think about this so that um this knowledge can flow and others can build on it as well well so what are the key questions that we are going to be trying to answer uh they are going to be well what is the energy use of a crypto assets what is the carbon footprint of the crypto asset typically we think of them uh in line but actually they may not be um so much in line right and keep in mind that in this case we're talking about what is called scope two uh emissions right um crypto activity does not involve a lot of direct emissions so if you go to a mining Farm you're not going to see a lot of greenhouse gases coming from there they come mostly from the electricity purchased by these players and um and at the point of origin is where the the common food print occurs so we're going to be asking what are the methodologies used to investigate this and what metrics we can use and how can we make make them Fair how can we make them comparable and to what extent we can um may they may not be perfectly comparable and also finally what is the right way of interpreting the results that we get so sustainability metrics uh they are metrics right metrics have limitations uh they have usefulness they have value but they also have limitations um it's never the fault of the metric in itself um but usage of a metric without proper context or interpretation uh then becomes not just a limitation but a flaw so we are going to try to provide some caveats and some guidelines on how to do exactly this so as I was saying and to organize what is going to be the rest of this webinar I'd like to uh POS it that when you're thinking of the climate impact of crypto you should be organ organizing your thought process into four main types of crypto asset systems now there are others and probably you could break this down into subgroups but I think this is a good framework to start and to organize your thought this is emerging from what we do ourselves so first of all we would have proof of stake systems and some other proof of and some other systems that are not proof of stake but they are quite similar from the energy point of view they may be different in from from another perspective but from the energy point of view they are similar so why well the reason for this is that for these systems the approach the that is at the basis of calculating carbon footprint and the energy consumption of the system is rather simple you need to count how many nodes there are and you need to investigate how much energy is consumed by the average node if you have the average consumption of a node and you have the number of nodes the number of participants in the network you get a total right now you can create bounds time series Etc but this is the core of the idea it's rather simple now we'll get into the details shortly for other types of systems um in what we call layer ones it's not exactly like this one participant in the network um in a proof of work system for instance or also what we'll call Storage uh blockchains may have a lot of unit of hardware and these units of Hardware may be may be used much or much more intensively or not so intensively so really that's not a good framework to think of their energy consumption what we should do instead is focus on the core activity in proof of work systems that leads to high energy consumption and that is mining mining is the process of uh in essence and I'm butchering the concept A bit trying to guess some random number that um through this process called hashing right so every time you try to guess one of these numbers every time you do a you you you make a hash you are spending a bit of energy so then the question is what is the energy consumption per hash and how many hashes are being produced at every second once you have one and the other you can multiply and you get a total energy consumption and from that you can go to the carbon footprint a similar but different thing happens with some blockchains that I'm setting aside CA Chia filecoin um they have very different from each other but I choose to lump them into one group because they storage is rather at the center of all of these systems so in these cases U to calculate their energy consumption what you need to do um is really look at how much data the systems are storing and what is the energy consumption per bite of data stored so we'll get into that shortly now there's one more type big and important group of systems that I want to draw your attention to and that is layer twos and here really I am lumping together two rather different things like tokens such as stable coins and also Layer Two platforms now they're different right for instance tokens have uh don't have their own nodes right but a platform like polygon has its own nodes a platform like polygon has uh its own throughput a token does not um however they are they have something in common and that is a layer 2 system is a platform that is built on top of a layer one that uses a borrows a part of the infrastructure so you know typically with systems like proof of work systems there is a vision to build some sort of digital cash or digital gold system and I'm always simplifying here but I think this is a good heuristic that's not typically exactly the vision of the other the first group of blockchains that we were looking into in these cases it's not really a digital cash or digital gold the analogy use often is digital oil um these platforms have the sort of vision to become an abore for the decentralized internet that's where the utility comes from the Native coin is used to um as a fuel for all of these decentralized apps built on top of them platforms and tokens so you could say that really the usefulness but also the activity the energy consumption of the layer one is caused by the layer two so then you need to ask yourself the question of what percentage of the activity of the layer one is due to the layer two or comes from the layer two and once you answer that question you know what percentage of the energy consumption uh of the layer uh one is equivalent to the energy consumption of theer that's a basic framework now let's take a deep dive so as I was saying there's a basic model which we could apply for proof of stake chains and here I'm lumping together nominated proof of stake delegated proof of stake liquid proof of stake Etc um but also you can apply this to others like the xrp or Stellar the the system as I was saying is you first need to estimate the power usage of one node one average representative node in a network since you don't know or often you don't know exactly what um Hardware is being used by each one of the participants of the network what you're going to want to do is probably set up a lower bound an upper bound and a best guess you can do this as estimation through an empirical assessment uh or you could do um literature review uh but that's the basic the first step that you need to do so once you have the power usage of one node at least a range of what the power usage could be you you need to investigate how many noes there are multiplying one one by the other you get a total power usage which you're probably going to want to convert to energy I won't get into the distinction between power and energy in very much detail now have to to connect later on this but in essence you know obviously um for some things measures of energy like wat hours instead of powers like Watts they're more more useful because there's a Time Dimension very implicit there you can work with cumulative metrics and it's easier for carbon footprint calculations Etc so yeah actually speaking of carbon footprint we have now calculated the total energy usage of the network how can we go from energy usage to carbon footprint well we need to know where the nodes are if we know where the nodes are we can then investigate the carbon intensity of the grid where these nodes are getting their en electricity for from and once you are able to construct a carbon intensity for the entire grid an average which will be derived from a weighted average of the locations of the gnes you can calculate the total carbon footprint so what the key questions to keep in mind here are how much power does one node use where to get the data from and a very important question is will this data be comparable across chains and how comparable can this be to what extent and a question here that comes a lot right and a complaint often heard when metrics on electricity consumption for pro stake systems for instance are presented is that a dead network doesn't use a lot of energy that doesn't make it C friendly in any and a climate friendly in any meaningful way so let's dive a bit into this problem of comparability what I was saying before is metrics are metrics right they have their value they have their usefulness um they have their limitations uh and we need to be wary of these limitations that's not the fault of the metric in fact if we can construct multiple metrics and that's what we do when we do our research that allows us to paint a more comprehensive picture of uh of the state-ofthe-art and the actual uh carbon footprint or carbon emission or energy consumption of the network so in the previous slide we were talking about the aggregate energy usage or the aggregate power usage of a network and that's an important measure in itself right but often um this is not enough we want to measure of Energy Efficiency because a Deb Network it doesn't consume a lot of energy but we we want our networks to be live right so we tend to consider um input and output not just inputs right the input is the electricity what is the output well the basic output in many of these things is the transactions so the more transactions you do with the same amount of energy you are cons considered more efficient and therefore um there is this metric energy consumption per transaction so that gives you another Insight on to The Climate friendliness of the blockchains now you may say this is not enough because an a network may have very very few nodes so because it has very very few nodes you can say well the energy consumption per transaction is low but that's only because of the number of nodes there's no particular Energy Efficiency going on here so you can take this into account and set up another netk another metric energy consumption per transaction per node this begins to paint a more complete picture but all of these metrics have problems as well and even if you do this you still get comparability problems because you can always say well maybe let's put it this way every time there's a new blockchain coming out they have metrics that they are extremely energy efficient and that's usually because there's not a lot of activity not a lot of um um transactions or they could be some transactions but not a lot of members not a lot of users so even if you do these metrics of energy consumption per transaction or per transaction per node may still not be um it may not still not be valid to say well if this network manages to become the new ethereum it will maintain these levels of efficiency when it becomes that big you you have no real Reon to believe that this ratio of energy consumption per transaction per node is going to actually hold constant over time so uh what we do in in our research is we also try to calculate the energy consumption per transaction at comparable levels of throughput for that um it's a bit more complicated but what we do is we try to model the relationship between throughput and validator account right typically you can think of it in this way the more um acity happening in a network measur in transactions the more people are joining the network to become validators so uh and if no activity happens in the network um well people will sooner or later abandon the network and there will be less validators than before also nodes have a higher marginal energy consumption the more transactions they proc processing on an individual level so all of these blockchains and here I'm I'm sharing some papers in which uh we we investigated this all of these metrics have their usefulness and I I I would argue we need to take them together to paint a complete picture and um this is doesn't even cover all uh there's some other things to take into account there's some transaction types right um consider for instance uh Solana a metrics of throughput typically consider vote transactions now vote transactions transaction that are a part of the consens the mechanism of achieving consensus in itself you cannot consider that a unit of output right that's a unit of input so you need to subtract that what about transaction complexity you may say well there's this blockchain that processes many more transactions and my blockchain is um looks bad because we process less transactions but actually our transactions are more complex right and that's delivering more value that's certainly an argument you can make but you will also need some um to to provide some additional information to argue this um is is really the case that the transactions are systematically more complex and are they more more complex due to a good reason that is that they are delivering more value or are they more complex because the underlying protocol is less efficient and you need to do more uh acrobatics to manage to achieve the same result also I've been talking about nodes validators they're not actually the same thing um uh and there's obviously validators are nodes but there's other types of nodes there's other forms of of participants in the network so there's the question there of uh who are you considering a participant now at the core of the network there's always the validators but you need to be consistent when you're comparing across networks because if you don't use the same type of node uh you will get incomparable figures so that's it we have the basic model proof of stake um um Stellar consensus Ripple consensus uh and a number of other blockchains can be analyzed with this basic model but as I was saying proof of work is different right proof of work think of Bitcoin and the various Forks think of Litecoin when Arrow see cash uh in this case as I was saying we need to understand what is the energy consumption of one hash and then how many hashes there are this will get us total energy usage now there's tricks here already now to come from energy to the carbon footprint we need to understand the carbon intensity of the network and that's the most contentious issue of all so we'll stop there for a second and then once we have sorted that we'll get to the total common footprint so the key questions are going to be what Hardware is being used where we get the data from how can this be comparable and here we also need to double click on the metric of energy consumption per transaction because in the proof of work world this is a highly disputed metric and how we know Carbon intensities so let's get to it so to calculate the energy consumption of a hash that is of the the energy consumption that each unit of mining activity uh um requires we need to understand what Hardware is being used what A6 uh generally uh depending on the on the blockchain type of course what as6 are being used and that's a problem because there's a lot of as6 there's so many different models of machines so how can we know which machines are being used there's miners all over the world we don't know how to contact them even if we were able to contact them we wouldn't be able to trust the results even if we were able to trust the results they would become outdated very soon right um we could ask the manufacturers what they are selling but even if we know what they're selling we don't know what has what is being used at the moment and what has has been already disposed as garbage or it's being stored it's not being used anymore until the Bitcoin price goes back up um so it is hard to collect this data from the very bottom but we need to do some sort of bottom up ex estimation anyway so what we do in this context is um there's two two approaches that you can follow one of them is well first of all you look at the Bitcoin price to take an example with Bitcoin you look at Bitcoin cash rate you look at the energy prices and you filter out which machines are profitable to use at a given moment and which machines are not profitable under any circumstance and then you discard those so now you have the subset of all the machines that are still profitable to use now you don't know which share uh belongs to each machine to each model so how many Machines of each type are being used so that's where you uh set up lower upper bounds and guess assuming that there's a higher concentration of the more efficient machines or of the less efficient machines or what Cambridge does or has done in the past U an equally weighted basket of machines now that's not the only way you can do it there's another approach pioneered by con Matrix um this approach is quite different I was saying that hashing is a process of guessing a ROM a a number right randomly so the more attempts you do the more chances you have of guessing correctly turns out that machines are very bad at doing things randomly um this includes bitco miners so um while it may look like they are guessing the numbers randomly you actually can identify some patterns in how um how they are guessing so if you manage to identify patterns in the numbers in the in the hashing activity you may be able to identify What machines these hashes are coming from what is the share of each machine type and therefore U what is the the actual distribution of of as types at a given moment now this is tricky um methodology is a bit nent uh but it's an interesting approach as well so we have identified the energy consumption per hash now we need to know how many hash rate how many hashes there are that's called the hash rate is measured in Tera hashes or EXA hash per seconds millions or millions of millions of of hashes and it's not typically a very controversial thing right there's plenty of sources online on how to do it in fact is a bit more tricky than it looks because the hash rate is not uh information that is really reported directly right you you don't really know directly how many hashes are being attempted at every moment you only know um about successful hashes so all the all this hashing activity which is mostly unsuccessful hashes you don't find out about so how do we know what is a hash rate at a given moment we look at the difficulty level right the difficulty level indirectly uh reflects what the hash rate is now difficulty in Bitcoin is adjusted every two weeks you can do some things to understand where has rate is at shorter intervals most people doing research in energy consumption of crypto to be um completely Fair don't usually go that deep there's plenty of sources on hash rate and they're not very disputed so now you know the total energy use because you have the energy consumption per hash and the total number of hashes so now we need to go to the difficult part C footprint I was saying before well what we did for the first group of blockchains is we just calculate we we just um find out where the nodes are we get the grid intensity and that place that gives us a average carbon intensity for the whole network and with that we get the carbon footprint well but it's actually a bit more difficult and you cannot really use this methodology you can use it but you'll get an overestimation of the carbon footprint and that's because in proof of work mining notably in Bitcoin mining there's a lot of mining happening off grid and not just off grid but it turns out that the mining happening outside of the grid is typically happening in um near renewable sources so there's a lot of Bitcoin miners going near constant renewable energy science like hydroelectric um uh power plants or um variable renewable energy like solar and wind FS so um if you only consider the grid carbon intensities you will end up over representing um and you will be introducing a bias you will be overestimating really the the Caron footprint of the network this offgrid mining even includes carbon negative mining right um in in using methane to mine Bitcoin methane that otherwise would be flared or vented uh and that's the key condition there um because of the higher radiative forcing of methane than carbon dioxide um by capturing the methane putting into a more efficient combustion engine than the efficiency of flaring and um turning most of that methane into CO2 the actual Caron footprint of the activity can be considered negative right now there's also a problem on of just finding out where the miners are um it's not easy you kind you don't even know who they are and um you can ask the the mining pools the mining pools they may give you some data they may not give you it they may have more or less reliable data and also you may have a more or less representative set of mining uh pools so for instance the the the gold standard in this area is the work by the Cambridge Center for alternative Finance but in in their initial work the the mining pools they had selected were under representing Texas which is a lower carbon intensity location uh within the us and that end up and ended up rep over representing the the C footprint of Bitcoin right so there's a lot of things to take into account uh what we do uh is we collect uh it's a huge bottom up effort of collecting actual minor location information right um and um we try to build from that uh using the Cambridge data which unfortunately hasn't yet been updated since 2022 uh as a subsidiary data right um this also relevant um to understand the possible effectiveness of uh Bitcoin mining bands so you see this this map I was showing before this is where the miners are and this other map I'm showing now shows the carbon intensity of each country you see they are not a perfect match uh and in fact many of the countries where there's most the mining activity like the EU um in general Canada um um and the us actually have a lower car footprint so it is interesting to look at this because it could mean that a mining band in these jurisdictions could mean an increased carbon footprint once these miners relocate to other countries so a lot to take into account uh in this um I'd like to share a few more thoughts about carbon accounting so carbon accounting is a very difficult um area um in fact uh it's not it's it's often invisible U people use different approaches to carbon accounting uh and you end up just hearing about the numbers uh these many grams or tons of CO2 but you're not usually told what carbon accounting approach has been used and this introduces a problem of comparability so in fact there's actually multiple ways of doing carbon accounting I would say some of the main ones are what we call attributional approaches and marginal approaches right within attributional approaches we'll find something called egalitarian approaches or Market based approaches so what I'm talking about here because this is um a bit Niche right so um the intuitive way of approaching uh carbon footprint is well you you it would go about something like this right say there's Bitcoin miners in a grid in a network and they consume 1% of the energy then what are the emissions of these Miners and the the answer is well 1% of the emissions right you take the total emissions and you assign them the same share this is actually one approach is attributional it's pretty much egalitarian however um that's not the only way you can do it right another way that uh you can go about this is say well let's say we have a grid here the United Kingdom grid and a new Bitcoin Farm turns on and that's more demand than before and therefore more energy electricity generation instantly needs to be generated to match this new demand this uh instant matching right requires firing up an electricity generation PL somewhere this firing up the question is what's the the carbon footprint of that firing up so if you if to meet this additional electricity demand you're firing up a coal plant or you're using a wind Park that changes everything and it it's not the same of saying oh I'm just going to I'm just going to consider 1% of all the emissions so this is something called marginality the marginal uh carbon footprint is the marginal effect of the marginal uh consumer of electricity now there's this uh impossibility trim identified by TR cross electricity rits uh and and carbon accounting and there and it's that there's three properties you would typically want your carbon accounting system to uphold but you can never have all three you can have two of these three at best sometimes only one but you can never have the three so you can have marginality that I um the effect of the marginal consumer is repres ented um an egalitarian approach does not mean that uh but you would also like to have a fungibility right that is that one kilowatt hour of my consumption is equal in carbon footprint to one kilowatt hour of each other's consumption uh that is that our energy consumption is fungible in terms of their carbon footprint now egalitarian approaches would meet that but marginal and even Market based approaches do not meet that and finally you would want something called compositionality that is that the sum of the C footprint of all the participants in the network uh amounts to the total C footprint of the network right that the whole is equal to the sum of its parts but only attributional systems uh can meet this requirement marginal systems don't so all of these systems are valid none of them is wrong or flawed it's just that we need to use them in the correct context and that we need to be very careful uh because it may make the figures incomparable some more thing a few more things about carbon accounting we have we have been talking about carbon footprint in general but you may also want to consider the um the carbon footprint of an individual transaction or the energy consumption of an individual transaction and as I was saying before the proof of work field is typically very opposed to energy consumption per transaction or carbon footprint per transaction action metrics because they find them to be very flawed for many reasons right um Bitcoin in in core in particular is a network that is a wholesale settlement Network right now in practice so it it doesn't uh conduct that many transactions uh but each one of these transactions they are extremely valuable right so you canot compare retail transactions to hold wholesale transactions and there's also Layer Two transactions that are often not considered right so transactions in lighting Network we consider them in our research but I understand that most um most of the parties in the market they don't right and it also creates all sorts of problems um of extrapolation so people may say well Bitcoin has a thousand times less transactions than Visa but it wants to replace Visa therefore if Bitcoin succeeds in it need to multiply its energ electricity consumption by a thousand and that doesn't follow that doesn't follow at all because actually this transaction is not what is causing the electricity consumption it's really the mining so actually this has a name right it's called transaction accounting and other people they follow this other uh perspective called origin accounting and that is you don't divide the electricity usage or the carbon footprint by the number of transactions conducted but by the number of coins might or the value of the coins mind depending on the approach that's called origin accounting now they are still tricky um it could lead to some strange results if you mine Bitcoin say in 2010 uh you sell them and in 2024 you buy them again and you can demonstrate it's exactly the same Bitcoin right and because you m them in a carbon a neutral way in 2010 can you say that you are carbon neutral now when you bought them again and likely you you cannot really say that and that's because um when you are demanding this Bitcoin well by increasing the demand on on the coin you also increasing the price if you increase the price you increase the reward for miners you incentivize more mining activity if you incentivize more mining activity they may use all sorts of electricity sources so it doesn't matter what you did in 2010 what electricity Source you used back then because what is causing the mining is the demand for Bitcoin the demand to hold Bitcoin this is called holding accounting or maintenance accounting there's another approach uh hybrid accounting which basically mixes transaction and maintenance accounting right saying well there is demand for uh to hold Bitcoin but there also also there is demand to transact with Bitcoin so you should consider both so as you can see lots of different schools all of which you could use in a given context but we need to be very careful when um extrapolating some um some charts here to um take into account I'm just going to briefly go over this you see um in the left um a chart illustrating how the Energy Efficiency of um Bitcoin miners is really increasing year over year and the top two charts at the at at at the top are comparisons of Bitcoin um against other Industries right this is useful context uh considering that sometimes uh Bitcoin is compared to countries uh you can do that but obviously is a very problematic comparison so Bitcoin supporters and this comes from nydig has a fantastic report uh on on on the state of of of Bitcoin mining I really recommend they were uh compiling this um other points of reference and there's also at the bottom two um charts with some projections uh that I find like useful for context you'll see that there is an expectation of uh Bitcoin emissions to continue increasing um probably in the most scenarios uh but not for that long and they should Peck soon that's because of the mcoin Bitcoin emission Schedule H and uh possibly some uh peaks in price are also expected not so far from now at the same time we're seeing that uh although the the growth of Bitcoin money has been high in previous years it's being dwarfed more and more by data centers and in particular AI service so um interesting context to have for projections one more thing I'd like to say about proof of work is um and this is um important and going beyond the main Fork of Bitcoin which is Bitcoin core and that's that proof of work does not necessarily mean a very high carbon footprint but not only that so it proof of work is energy intensive it can mean a high C footprint but it may also not it depends on the energy sources being used but I'm going to go a step further and say even proof of work does even though proof of work is energy intensive that does not mean it is high on the energy consumption intrinsically because there's other aspects of proof of work architecture that may affect the energy consumption even if the system is energy intensive and these are mostly block size and block time so uh obviously this is not relevant to bitcoin core but it is relevant to other Forks of Bitcoin like Bitcoin cash Bitcoin SV with a higher block size what do I mean what block size right is well how big is a Bitcoin block how many transactions can you put there is it one megabyte is it two is it 10 is it 100 megabytes the more transactions you can put there the lower the transaction fees which will reduce minor Revenue could potentially also reduce uh bitcoin price and that will reduce mining activity and at the same time because it's funing more transactions you can argue that it makes it more efficient at least under one metric of Energy Efficiency which is energy consumption per transaction um because not only do you have a lower energy consumption you also have more transactions uh for that energy you can also instead of increasing block size increase increase block time it is the same really to have one 10 megabyte block uh every 10 minutes or one one megabyte block every one minute is the same now I'm not advocating for higher block sizes or greater block time what I'm saying is what this means is that you can just it would not be probably a very good policy intervention to Simply ban proof of work in itself because even in in those terms proof of work can be low energy so interesting thing to take into account obviously mining Revenue the price of the coin but also the emission schedule will also affect this energy consumption so briefly I'm going to touch upon the final two types of um systems and one would be storage based blockchains I'm lumping here three very different uh systems Fon Gia and CIA um but I I think it's it's it's important to group them because these three systems they have storage at their core so when you look into what's going on in one of these networks you will see well there's this many petabytes of information being stor St right now that is important thing not how many NOS there are right now in terms of energy consumption I mean now if you do a dip dive and by the way the the gold standard in the literature for this is paper by the team of Alan ransil from protocol Labs there's a follow-up paper that came out rather recently uh it's really great work um and it establishes a b basic framework that I think is correct right to understand the energy consumption of a storage based blockchain you need to understand how storage works and that is it takes two steps there's a first step when you are doing the ceiling or the plotting that is the recording the bringing the D data onto the hard drive and that's a very energy intensive step but it's also a oneoff energy uh a one-off step once you have recorded the data then you need to sustain that storage over time so you have the energy consumption of a ceiling that is high but one off and the energy Assumption of storage that is constant but smaller and in addition if there's a consensus mechanism on top of that you need to consider that as well that is the basic framework you should follow for these uh systems um you should also here raise the issue of energy intensity and this is also relevant uh in the context of for instance MAA this law that establishes right now at least in the in the uh final report of the Asma um the metric of energy intensity is considered to be an consumption per transaction however is this really a fair metric for systems whose um use case is not um how many transactions they can process in a decentralized way but how much information they can store in a decentralized way so maybe you should look into energy consumption per bite of storage now just disclaimer they are very different systems from each other so in case you're not very similar I will walk you through some of the main differences uh just to avoid butchering any one of of these systems so um in the case of finecoin but also seia they they intend to be decentralized storage networks right um that's not the case of chia right GI is more of the sort of app store for the decentralized internet that I was mentioning before GI instead uses storage um as a way to achieve cons consensus so they have a sort of proof of work system but when instead of doing the work of hashing they're doing the work of storing that's not how fcon Works has Fon has a rather similar to proof of stake consens of system and CI us is proof of work also information sealing it is actually different across uh the systems because for instance uh while filecoin green has has done a great job of publishing this data this data is hard to find for gania um and you cannot simply take variations in storage to be representative of of the amount of ceiling going on not only because you need to consider negative you need to exclude negative variation so it doesn't mean there is something has been unsealed but also because um with the case of chia for instance variations in storage will represent really variations in compression levels not in storage now we're getting a bit too technical but I just wanted to make the disclaimer that they have very different systems I'm just trying to provide a basic framework right so to finalize the final group layer twos and I think I already introduced at how you should think of layer twos um so recapping a bit what you should think of is about two things one is there an intrinsic energy usage of the layer Tu and two what is the extrinsic energy use of the layer Tu the word Layer Two is a bit not super technical for all these sort of applications some people are going to say well Layer Two is not exactly this but overall this is a good schema to think about things so obviously say usct does not have its own noes right so it doesn't have an intrinsic energy use only the energy use that it borrows from the layer one but other networks think of polygon does H will will have this but when you're thinking about the exic energy use so the question I mentioned before is what percentage of the activity of the layer one can be attributed to the layer two so what is this percentage of activity right and there's some great paper by scer Lena Classen and Christian stall about this uh really good work that um explains how you can um do this and actually that takes us back to the carbon accounting uh theories so these carbon accounting theories which look like a theoretical thing it actually becomes very practical when you need to attribute cality like it happens in these cases so basically what you're going to be doing is looking into how many transactions in the layer one are caused by the layer two and how many coins in the layer one are being demanded by the layer two that's a short uh summary of uh what this is now there are other systems that you need to think slightly differently about like layer zeros right chains that have a base layer and then relay chains thinking of PKA do Avalanche Cosmos uh exchanges where they hold many different coins or oracles which who have the energy consumption of the Oracle and wrapped coins right wrapped coins like wrapped Bitcoin or they are a token on ethereum for instance so you can say well this is token on ethereum it's an percentage of the electricity consumption of ethereum but maybe is you you could ask the question is rap Bitcoin causing any activity on the Bitcoin Network as well so a bit more complicated Food For Thought for future things which your future exploration uh I hope you have been finding this useful also also keep in mind that some of the building blocks of the Reon we have used today can actually help us with other metrics that are useful also for for instance the maao right uh based on energy usage you can have the building blocks to calculate water footprint based on water intensity or land footprint based on land intensity of electricity and based on the number of validators that you have identified you may have the building blocks to calculate electronic waste so that's not the end of that story there's much more to those things but these are some key building blocks that when you start following this framework that we have outlined you will be able to explore so some key lessons I would like to extract from all of this um not all of them I have introduced but I will say them now uh most crypto assets they actually have a very low Comm footprint something we we should keep in mind now we are trying to be very scientific about exactly how low we should keep in mind it is actually quite low uh in the case of proof of work now it tends to be higher uh but it doesn't necessarily need to be higher and even if it is higher now it doesn't necessarily need to be that high in the future uh be aware of comparisons across blockchain types right energy consumption per bite versus energy consumption per transaction uh comparison comparisons with between Industries uh Bitcoin and a country they're complicated uh be aware even of comparisons within the same blockchain type because blockchain systems are very um different let's be careful with simplistic extrapolations what I was saying with the case of Visa also with double counting because if you're going to say that um USD usdc and polygon they use a percentage of the electricity of say ethereum then you cannot then add all of these figures up because you will need to net otherwise you are going to be double counting and some final thoughts on policy a high carbon footprint for a consensus mechanism does not mean that a ban is the best policy recommendation um and also the best ways to offset the car carbon footprint uh will actually depend on your accounting Theory and that's something you should pay attention to now I would like to get your thoughts um we would love to do more seminars uh on this topic and here what you have to say here what you would like to learn about if you would like U to get some support on how to calculate your sustainability indicators coming into compliance with Ma uh please let us know uh I'm open now for for questions and um thank you for for your attention