A young wizard named Forrest took me aside at a new years eve afterparty and told me his concerns about the energy use of cryptocurrency. He was very passionate about this problem and so I dedicate this post to him. We really ought to address this as carefully as we can, and with an open mind to this problem. After all there could be real danger of a "Economics 2.0" situation a la Charles Stross's book "Accelerondo", in which economic competition drives some rather nasty developments which eventually leads to the building of a Dyson sphere around the sun, taking all the energy available in our star and using it towards some financial processing. Scary stuff indeed!
First off, lets try to recognize the factors that are involved.
When we see a line of thousands of idling cars waiting on a packed highway, as we do all too often, each with a single occupant, spewing carbon dioxide and worse into the atmosphere, we don't immediately say "cars are evil". Nor do we do so when we see a crushed armadillo or frog or human lying mangled by automotive impact. We know that there is more to this problem than might at first meet the eye. The price of gas and the history of the oil refinement industry are part of this story, as are the social factors that went into the construction of the highways, and the adoption of drive-to-work suburban mentality, and the fiat empire that watches over and feeds this hierarchy that makes it possible. It's easy to have sympathy with the "cars suck" viewpoint, and cars are often compared to cryptos for pedagogic reasons.
When we see that people are being murdered over marijuana deals, we don't immediately say "the plant known as cannabis is evil". We know that prohibition is the evil thing here, and drug abuse, and lack of education, and various other factors.
So to see a ton of electrical power being used by bitcoin miners, lets just say this is power that might have gone to something more useful, or perhaps carbon that might have remained in the ground. If the heat isn't being used, it's wasted. Like the nuke plant that dumps the tepid water in the river, or the flare-off at the local oil field.
Why are these folks using so much electricity and isn't it wasteful? How could it be avoided?
That's the problem in a nutshell.
Before we continue with looking at answering this question, lets look at some metrics to define the problem.
Many commentators state that bitcoin miners use as much power as a small country, or give some numbers to describe the amount of power being used. However this belies the difficulty of analyzing the issue: in fact we don't know how much power is being used. It might be enough power for 100 countries, or it might be a lot more or less. The only way we can estimate is by looking at the efficiency of commercially available miners and making some assumptions about what people are using. Technically, there is the possibility that a big fraction of mining is being done by one guy in a closet with a set of AA batteries. However we know from all the mining porn we've watched that there are indeed large farms spinning lots of fans and at least making it look as if they are big power-hungry mining ops. And logic says there would be, as there is spare capacity and people who want to turn it into money.
If we want to know how much energy is being used by the miners, we need to know something about the efficiency of their machines. This is a little difficult because we know for sure that somebody is out there now mining with some old x86 CPUs powered with a diesel generator, while others are mining with stealth super-efficient technology which we don't know the name of yet.
So what can we say? Well a good order-of-magnitude estimate is the upper bound on rational miners in the steady state : they aren't going to be using more energy than the mining reward provides. For BTC today that's about 13 BTC per block (12.5 in coinbase and half a coin or so in fees) or 1872 BTC per day. Let's call it about 2 million dollars a day. To find some comparison for this figure we can look at the earnings of Exxon-Mobil [XOM] which are declared to be something like 53 million dollars per day. So that's kind of the hard cap on energy consumption of Exxon-Mobil (not the energy you use when you buy it, but what they use).
The argument here is that marginal cost = marginal revenue is the upper limit of the cost, as presented by Paul Sztorc. This is an economic equilibrium so I argue it is unlikely to be actually reached, however it is as good an estimate as we can come by. The actual power use will be somewhere between 0.01 and 0.99 of what we calculate as all the miners block reward spent on electricity at some nominal price. Since we are speaking in order of magnitude one-significant-digit terms here lets call it ~ 1 million USD per day spent on electricity.
Wasteful? Perhaps. However it's worth pointing out that this is peanuts compared to some other energy sinks. For example, US military energy costs are estimated to come to some 20 billion USD annually, or near ~60 million USD per day in energy costs (apparently down quite a bit recently as measured in real units). 3000% higher than bitcoin miners, and probably much more. How about street lights? Cars? Airplanes? The waste goes on and on. But before we get carried away, lets try the analysis of Paul Sztorc on the energy requirements of fiat issuance.
How much fiat is issued daily? Well of course we don't even know that much, that's the nature of the beast. But we can look at some estimates. My quick search of FED stats suggests that the figure is close to 3/4 of a billion USD per day. What are the costs associated with this production? At first one is tempted to dismiss out of hand Paul Sztorc's claim that marginal revenue tends to equal marginal costs here. What are the costs of issuing fiat currency? Basically nothing, as no energy or cost is a associated with fiat issuance. However, this leaves off the associated indirect waste of the process. Sports cars rotting away in garages, empty mansions being heated, massive development projects that are unneeded or worse, and of course warfare, fascism, monopoly business and destruction of free markets towards even greater inefficiencies. All these things produce waste far beyond what the measly 2m usd per day of BTC rewards can do.
While this comparative analysis is useful in the present day, we should be clear it doesn't necessarily work in future economic conditions. The energy costs and potential wastes of proof-of-work in other political conditions also needs careful consideration. Back to that later.
OK so by now we've addressed how energy use is certainly a concern, we've put an upper bound on the problem and compared it to some other problems. Now lets take a look at some other folks discussing this badly:
What nonsense! The bitcoin network has NO minimum energy requirement per transaction. People are choosing to use whatever spare capacity they can find to mine as they see fit, as this is how the system is designed to run - on spare power capacity. Tying money to energy results in LESS energy waste, not more. Maybe we should turn the lights off after all, now that money doesn't grow on trees and every joule could be satoshis in the wallet. Which scenario are you more likely to watch carefully your fossil fuel usage: 1- fuel price tied to fiat which you can issue to yourself in arbitrary quantities at no cost 2- fuel price tied to bitcoin which nobody can issue privately and is tied to energy use ? Go on answer the question as to which scenario you are more likely to be efficient about your fuel usage.
Another example, this one from an in-color whitepaper that pushes a 100% premined proof-of-stake coin:
All four points are dubious here.
So there are many examples of people incorrectly stating the nature of this problem, but does that help us?
In fact many people push the agenda that proof-of-work is an out of control energy disaster requiring tons of energy, when in fact the networks function at any level of energy and are designed to run on spare capacity. The issue is not that proof-of-work is somehow an energy hog, but that people are choosing to put their energy into mining rather than into other projects. This is the issue we need to address.
Another opinion is that a tie from currency to energy would be a huge boon to the energy budget of humanity, leading to far less wasted energy. How would this work? Well once there are tools that let us turn energy directly into money, and once we cannot get money from nothing by knowing the right people, we will be much more careful with our energy usage.
Such a future is that which many public coin users hope for. Ask yourself who is more likely to waste energy, somebody who can issue currency or get it very cheaply and trade it for energy? Or somebody who had to use energy to get currency.
1 - The excessive use of energy by miners of POW coins is both a problem itself and a symptom of other problems inherent in our financial system.
2 - This problem is often misunderstood and misportrayed by people pushing their own agenda
3 - The rise of public proof-of-work coins has a potential to vastly decrease the waste of energy currently occuring during the anthropocene extinction
4 - The energy and other wastes produced by private currency issuance are vastly worse than those produced by public currency issuance
5 - The jury is out on whether other techniques of public ledger maintenance could be improvements
That's all for now. Tune in later for an updated quantitative analysis of satoshis-per-joule.