The purpose of this article is to help bring real world comparisons to light to help the general public understand Bitcoin’s energy consumption while comparing it to everyday activities they never question.
Before we delve into the numbers, we should clarify a few things.
We are not here to defend Bitcoin’s energy usage, as it uses this energy to secure the network and will continue to do so as long as Bitcoin is useful.
To note, we also treat Bitcoin in this scenario as a payment only infrastructure at a detriment to Bitcoin, when in reality Bitcoin is much more. The last few years have seen the introduction of Layer 2 technologies such as Lightning and more recently we’ve seen Layer 3 protocols introduced such as Impervious all with negligible additional energy usage. Lightning also has uses outside of just “payments”, an example is Sphinx chat, a Lightning based chat app, just like WhatsApp, but every message is securely sent over Lightning as a transaction.
We should also make it clear that due to Bitcoin’s incredible transparency, we can easily estimate the energy usage from mining in a simple manner from one source. This is something that is not possible in other industries, making the energy consumption of Bitcoin an easy attack vector for people opposing Bitcoin.
Another myth we should dispel before we begin is that Bitcoin does not require more energy for more transactions as these 2 metrics are independent of each other. It’s often incorrectly reported that for Bitcoin to scale it would have to scale at a 1:1 rate with energy consumption, this is not true. To help understand this, we could create “BlueBitcoin”, an exact clone of Bitcoin and be the only miner on the whole network with only our laptop. This would still result in the same amount of transactions per second as Bitcoin. Energy usage is a security metric, not a speed or transaction metric.
For our numbers we are benchmarking 13,479 transactions per second via the Lightning Network, which we will explain later.
Bitcoin using energy isn’t a problem, the real issue is the source of the energy and its impact on the environment.
Background and Sources
There has been a significant amount of noise around the energy usage of the Bitcoin network ever since the price soared along with the hash rate earlier this year. This has been accompanied with comparisons of Bitcoin’s sluggish transaction rate and excessive GHG (Greenhouse gas) emissions. Many people have come to realise that standard BTC transactions are not the future of payments. Enter the Lightning Network. The Lightning Network is a layer 2 technology built on top of Bitcoin’s secure base layer. At a high level it uses smart contracts to create a network of connected peer-to-peer nodes able to make almost instantaneous and near zero cost transactions, all while the payment is able to be finalised and settled instantly, something not possible within other payment networks. Lightning payments are more efficient in both energy and fees. But how much more efficient are they?
Throughout this article we will use grams of CO2e (the e adjusts other GHGs and equalises to C02) as a means of comparisons. This allows for adjustment to specific energy mixes.
For the ease of calculation, we will take the average of “non-Renewables” as Coal, Gas and Oil and “Green sources” as the average of, Bioenergy, Solar, Geothermal, Wind, Wave, Nuclear*, River Marine and Hydro. While the specific mix will vary between countries this captures the fact that generally green sources produce vastly less GHG than non-renewables.
*(Nuclear is included as a green technology as while it does create toxic waste in the form of radioactive material this does not manifest itself as a GHG)
In order to estimate the total cost of a Lightning transaction we must run Bitcoin underneath this, therefore we will now estimate the carbon cost of the network. According to the Bitcoin Mining Council (BMC) the network is currently estimated to be using 189TWh, this is on the upper end of estimates from other studies by the University of Cambridge. Previous Studies have estimated that the proportion of Bitcoin mining that is mined through green energy sources is between 39–74%, however more recently the BMC has estimated the total network is running at 56% renewable. There is clearly wide variation in this but erring on the side of caution and siding with the mining experts we shall use the 56% figure. Using this energy mix we would put the total carbon generation at approximately 61.4 trillion grams of CO2 per year.
Now we have the total carbon cost of the Bitcoin network we can talk about the Lightning network built on top of Bitcoin. The Lightning Network is still in its early days but the key element of Lightning is its ability to scale non-proportionally to energy usage. A Lightning network transaction is designed for efficiency and works on a peer-to-peer network leveraging other people’s existing connections. A transaction takes a fraction of a second and can take just 2 nodes to confirm a transaction. For transactions to be connected across the world we will assume 6 nodes, each running for half a second. This is in line with the standard 6 degrees of separation where no person can be more than 6 connections away from anyone in the world. If and when the Lightning network scales and becomes widely used the majority of transactions may not pass through more than just 6 nodes. In order to estimate a per transaction carbon cost we need to make some assumptions about the size of the Lightning network. The VISA payments network alone facilitated 185.5 billion transactions in 2019. Adding in UnionPay and Mastercard with another 131.2 billion and 108.4 billion respectively this works out as 425.1 billion transactions within just 3 companies. This excludes cash and other means of transferring wealth such as PayPal or bank transfers. Taking 425.1 billion transactions a year, we are looking at 13,479.8 transactions per second.
This is perfectly feasible under the Lightning network. We are not there yet, but it would be unfair to measure Lightnings capability whilst still in its infancy. We will therefore use 13,479.8 transactions a second for comparison. It should be noted that even this massive number of transactions is a low end estimate as the Lightning network has the potential to replace even cash transactions of which there are billions of a year and not to forget this is Lightning usage for payments only, which as we mentioned earlier Lightning’s use case can go beyond payments, as can Bitcoin’s.
A raspberry pi including an SSD that can run a Lightning network node draws just 5.1 volts and 3 amps, we can assume a worst-case scenario of 1 power factor. This works out as 0.015 kWh. Adjusting this for the 6 nodes and a half second use per transaction we can assume 1 transaction consumes:
Converting this to gCO2 with 100% non-renewable energy (as this is up to the user, and we can assume a worst-case scenario) we can say with some reasonable level of confidence that the variable carbon emissions of running enough raspberry pi’s to send a transaction worldwide is 0.009922325 gCO2 per transaction.
Estimating the carbon cost of manufacturing a raspberry pi and determining how many are needed to support the network is fiendishly difficult. For simplicity we will model a world where every node is an iPhone 11 as the carbon cost of this is known at approximately 105kg. In reality raspberry pi’s are much less carbon intensive to construct. We will estimate the number of transactions as 13,479. This is equivalent to one node completing one transaction per second. We shall not go too deeply into this but we have used this due to the fact nodes can be centralised through exchanges such as Strike or CoinCorner. Adding it this way would also require that every node is replaced every year (nodes can last multiple years) so is erring towards a worst case scenario for Bitcoin. This would be 1.4 million tonnes of CO2 for the Lightning node infrastructure, or to put it another way, 0.0033 grams of CO2 per year per transaction.
We can now calculate the total cost of sending a Bitcoin transaction by dividing the fixed costs of the Bitcoin and Lightning infrastructure by the transactions per year and then adding the marginal cost in electricity of running the pi’s.
We now have a reference point to measure the efficiency of the Lightning Network as a means of payments. It must also be said that this is putting the entire carbon cost of the Bitcoin network into a means of peer-to-peer payments. Anyone who has read anything about Bitcoin knows this is only one element of Bitcoin, the network is the most secure savings platform in the world as well as providing a base layer for future technologies. It must be stressed that this is a worst-case scenario of energy usage and as time passes the use case for Bitcoin grows stronger and stronger.
Email works through a net of servers and centralised exchanges that can operate at extremely high speeds. In 2019 nearly 300 billion emails were sent per day and this number is expected to rise with forecasts of nearly 400 billion per day by 2025. While the technology does allow for an extraordinary number of emails the amount of energy to facilitate such transactions is equally enormous. A large email is said to cost approximately 17 grams of CO2. You could be looking at just 9 emails to be on par with the carbon emissions of a transaction that can theoretically transfer millions of pounds worth of Bitcoin, and nobody has issues with sending hundreds of emails a day. To make matters worse, this is multiplied out by mailing lists. Both domestic and commercial emails sending out hundreds of emails at once can cost thousands of grams of CO2.
Leveling Bitcoin and emails energy usage is a business that needs to take into account the relative use of both. On average people receive 75 emails a day, with some estimates as high as 120 for business users whereas the average number of transactions in the euro zone is 1.5 per day. Using the 75 figure at approximately 90% green energy email and Bitcoin become on par with one another. And we must not forget this is pinning the entirety of Bitcoin onto payments and disregarding the many other use cases of Bitcoin.
We can take this further, comparing Lightning to other “green” technologies. An Electric Vehicle (EV) is an enormously complex object to manufacture requiring rare minerals such as lithium to be mined for the battery and steel to be produced for a chassis. The majority of carbon emissions for a car are made before the wheels ever touch the road. The international council of clean transportation estimated that the EU average for an EV to be produced made 18,750,000 gCO2 per car. This is quite simply enormous. The emissions from producing 1 EV is equivalent to some 129,704 Lightning transactions! An average person does 1.5 transactions per day, so with an 81 year life expectancy you would expect to need 43,000 transactions if you started spending from the day you are born.
To put this into perspective, if you were to buy an EV with a Lightning transaction the transaction’s CO2 footprint would be just 0.00081% of the car. In terms of weight a windscreen wiper represents 0.0112% weight of the car (assuming 180-gram wiper and 1.6 tonne vehicle (electric vehicles on average are heavier than internal combustion engine vehicles)). Comparing these two percentages:
We can say that 1 Lightning transaction is equivalent to 3.45% of a windscreen wiper in respect of the whole car. It’s hard to overstate what a small proportion this is. 3.45% of a windscreen wiper equals the carbon emissions from buying the entire vehicle. While this can be seen as far-fetched and unrealistic comparing windscreen wipers to electronic payments, the fact of the matter is the carbon cost of a Lightning transaction is positive, but it’s of a magnitude that people do not even register when applied to other areas of the economy.
Using the website carbon cost calculator we can estimate that 1 visit to a well known social media site generates 0.93 grams of carbon. If we were then to take some account that has, say, 60 million followers and reasonably assume a worst case scenario of every one of their followers simply logging on to view a post, this would have a potential carbon cost of 53.7 million grams of carbon. 53.7 million grams from one person’s admittedly large following! This would facilitate a potential 385,999 Lightning transactions. This puts the carbon cost of the world around us into sharp relief.
Food and drink
An orange in itself is relatively carbon friendly at just 150gCO2 (just over a Lightning transaction). But your morning “not from concentrate” orange juice is incredibly carbon intensive. It is estimated that 1L of orange juice produces the same carbon footprint as 6kg of oranges! This would mean a 250ml glass of OJ would come in at around 10 transactions.
A box of eggs, which are seen as a healthy and sustainable source of protein come in at 2kg of CO2. This is equal to 13 transactions, effectively enough for an average persons transactions for 10 days.
1 pint of milk produced and consumed in the UK takes approximately 1.1kg of CO2 to produce. This is equated to 7 transactions equal to enough milk for 2 large coffees! While this doesn't sound a huge amount, we must remember that a single cow produces 16,700 pints of a milk a year and is kept for 3–4 years. This means one female cow will produce upwards of 66,000 pints in her lifetime. Forgetting about meat production and focusing solely on milk production you could have over 500,00 transactions for the same cost as 1 cow.
It should be noted that all of these numbers are from the current energy mix of the Bitcoin network. As the hash rate moves to more renewable sources the carbon cost of transactions go down, meanwhile the emissions from your morning OJ are relatively fixed for now.
Current Banking System
Galaxy digital have previously estimated the total energy consumption of the existing banking system and estimated a total of 263.72 TWh usage. Other studeis run by Hass McCook have estimated the total carbon cost of the financial sector at over 380 million tonnes of CO2, this is some 6x that of Bitcoin. These numbers have a lot of underlying assumptions as very few banks report, or potentially even know what their total annual energy usage is. However, taking this as given, Bitcoin represents a potential saving in total energy consumption. Comparing the Bitcoin network with the entire global financial system is a task that will always be fraught with difficult assumptions and is not something we will attempt to do here. But the vast complexity that the financial system brings with insurance, stock and bond markets all have the potential to find their way onto the Bitcoin network, and the brilliant thing about Bitcoin is it can do this in such a way that the energy consumption can grow disproportionally to this through efficient scaling as we have seen so far with the efficient Lightning network.
In a potentially more controversial comparison humans emit approximately 500L of CO2 per day, simply by breathing. It is important to note that this does not contribute to global warming as over our lifetime we absorb carbon from plants and other food and then emit these when we breath and eventually decompose, therefore over the long run we will have a net zero emission. This is countered by activities such as burning fossil fuels which release trapped carbon from millions of years ago. However, we can still use this for comparison purposes and use it to put a Lightning transaction into scale. 500L of CO2 works out at approximately 1kg of CO2. This is to say just by breathing an average human will emit the same amount of carbon as approximately 6 Lightning transactions every single day.
The above article shows that while Bitcoin’s energy usage may seem large at face value when broken down even on a purely transactional level the carbon emissions are comparable or even lower than some everyday activities that few would describe as excessive energy use. While providing a carbon effective way of transacting around the world, BTCs energy use also provides the most secure computer network in the world, relying on and discriminating against no one.
Throughout this article we have used 13,479 transactions a second, the same number as the three biggest payment providers in the world combined. It is impossible to actually know how many Lightning transactions happen every second but the capacity of the network is certainly capable of this and well beyond with no theoretical limit on transactions.
As for a goal for the future we can see miners flocking to the cheapest energy sources in the world. This is largely possible due to the incredible competitiveness of the industry and the speed at which it can adapt to changing geopolitical conditions, proof of which we have seen over the past months with the move out of China. 90% green energy sources may be possible within the next 5 years and possibly 0 emissions by 2031, making it one of, if not, the cleanest industry in the world. We have included the comparisons here for these numbers and even taken into account a fairly extreme 5x increase in power usage as the hash rate will have continued to grow as mining becomes more profitable as energy becomes cheaper through efficient green energy production and the Bitcoin price continues to rise.
With adoption of Lightning even more layer 3 technologies can build on this taking advantage of the decentralised nature and incredible security of the Bitcoin core layer. This sets the scene for many, many more Lightning transactions in the near future which will continue to drive down the carbon cost per transaction. With the rise of more sustainable energy generation, the widespread adoption of Lightning and the potential for even further scaling, Bitcoin looks to become one of the most efficient payment providers, store of value and influential technologies in the world.
Help us build the inevitable.