Bitcoin Energy Consumption Index

Key Network Statistics



*The assumptions underlying this estimate can be found here.

Did you know?

Ever since its inception Bitcoin’s trust-minimizing consensus has been enabled by its proof-of-work algorithm. New sets of transactions (blocks) are added to Bitcoin’s blockchain roughly every 10 minutes by so-called miners. While working on the blockchain these miners aren’t required to trust each other. The only thing miners have to trust is the code that runs Bitcoin. The code includes several rules to validate new transactions. For example, a transaction can only be valid if the sender actually owns the sent amount. Every miner individually confirms whether transactions adhere to these rules, eliminating the need to trust other miners.

The trick is to get all miners to agree on the same history of transactions. Every miner in the network is constantly tasked with preparing the next batch of transactions for the blockchain. Only one of these blocks will be randomly selected to become the latest block on the chain. Random selection in a distributed network isn’t easy, so this is where proof-of-work comes in. In proof-of-work, the next block comes from the first miner that produces a valid one. This is easier said than done, as the Bitcoin protocol makes it very difficult for miners to do so. In fact, the difficulty is regularly adjusted by the protocol to ensure that all miners in the network will only produce one valid bock every 10 minutes on average. Once one of the miners finally manages to produce a valid block, it will inform the rest of the network. Other miners will accept this block once they confirm it adheres to all rules, and then discard whatever block they had been working on themselves. The lucky miner gets rewarded with a fixed amount of coins, along with the transaction fees belonging to the processed transactions in the new block. The cycle then starts again.

What kind of work are miners performing?

The process of producing a valid block is largely based on trial and error, where miners are making numerous attempts every second trying to find the right value for a block component called the “nonce“, and hoping the resulting completed block will match the requirements (as there is no way to predict the outcome). For this reason, mining is sometimes compared to a lottery where you can pick your own numbers. The number of attempts (hashes) per second is given by your mining equipment’s hashrate. This will typically be expressed in Gigahash per second (1 billion hashes per second).

Sustainability

The continuous block mining cycle incentivizes people all over the world to mine Bitcoin. As mining can provide a solid stream of revenue, people are very willing to run power-hungry machines to get a piece of it. Over the years this has caused the total energy consumption of the Bitcoin network to grow to epic proportions, as the price of the currency reached new highs. The entire Bitcoin network now consumes more energy than a number of countries, based on a report published by the International Energy Agency. If Bitcoin was a country, it would rank as shown below.

Countries by Energy Consumption



Comparing Bitcoin to other payment systems

To put the energy consumed by the Bitcoin network into perspective we can compare it to another payment system like VISA for example. Even though the available information on VISA’s energy consumption is limited, we can establish that the data centers that process VISA’s transactions consume electricity equal to that of 50,000 U.S. households. We also know VISA processed 82.3 billion transactions in 2016. With the help of these numbers, it is possible to compare both networks and show that Bitcoin is extremely more energy intensive per transaction than VISA.

VISA Network Statistics



Of course, these numbers are far from perfect (e.g. energy consumption of VISA offices isn’t included), but the differences are so extreme that they will remain shocking regardless. One could argue that this is simply the price of a transaction that doesn’t require a trusted third party, but this price doesn’t have to be so high as will be discussed hereafter.

Alternatives

Proof-of-work was the first consensus algorithm that managed to prove itself, but it isn’t the only consensus algorithm. More energy efficient algorithms, like proof-of-stake, have been in development over recent years. In proof-of-stake coin owners create blocks rather than miners, thus not requiring power hungry machines that produce as many hashes per second as possible. Because of this, the energy consumption of proof-of-stake is negligible compared to proof-of-work. Bitcoin could potentially switch to such an consensus algorithm, which would significantly improve sustainability. The only downside is that there are many different versions of proof-of-stake, and none of these have fully proven themselves yet. Nevertheless the work on these algorithms offers good hope for the future.

Model and key assumptions

Even though the total network hashrate can easily be calculated, it is impossible to tell what this means in terms of electricity consumption as there is no central register with all active machines (and their exact power consumption). In the past, electricity consumption estimates typically included an assumption on what machines were still active and how they were distributed, in order to arrive at a certain number of Watts consumed per Gigahash/sec (GH/s). This arbitrary approach has led to a wide set of electricity consumption estimates that strongly deviate from one another, sometimes with a disregard to the economic consequences of the chosen parameters. The Bitcoin Energy Consumption Index therefore proposes to turn the problem around, and approach electricity consumption from an economic perspective.

The index is built on the premise that miner income and costs are related. Since electricity costs are a major component of the ongoing costs, it follows that the total electricity consumption of the Bitcoin network must be related to miner income as well. In order to calculate the electricity consumption of all miners in the Bitcoin network, one should start by determining the total miner income and the average percentage of miner income spent on ongoing costs. The former is quite easy to determine, as miner income depends on the block reward (including fees) and the price of Bitcoin. Both variables are known, and the number of blocks mined is too.

This second part is a complex problem by itself, as new and increasingly energy efficient mining farms are constantly being added to the network while older ones are slowly forced out. With mining rewards always being equal to the proportional contribution to the total network hashrate (on average), the profitability of mining hardware ultimately depends on the variable (electricity) costs per GH/s. This provides manufacturers with a strong incentive to keep creating more efficient hardware. Starting a new and more efficient farm with these machines can then be very rewarding, although it cannibalizes the profits of other farms.

The total ongoing costs for a new farm may range from somewhere between 20 to 30 percent of the total income. At the same time, the ongoing costs for a mining farm at the end of its lifetime may approach or even exceed 100 percent (as they receive an increasingly smaller slice of the same pie). These farms are first of all trying to make the most out of their investment, and may keep on running even after they start losing money due to contracts not having expired yet, expectations of price increases (what is a loss one day can suddenly be a big profit due to Bitcoin volatility), or simply due to privacy reasons (mining is a way to acquire coins anonymously) for example.

Economic theory suggests that the marginal product of mining should theoretically equal its marginal cost in a competitive market. This would mean we can calculate the network’s energy efficiency by solving for the break-even electricity costs. Adam Hayes details this appraoch in his paper titled “A Cost of Production Model for Bitcoin” released in 2015. Tomaso Aste (2016), on the other hand, implies that the average costs of mining are closer to 55 percent of the available miner income. Aste, however, doesn’t provide many details along with his estimate. To be on the conservative side, the average cost percentage used to calculate the Index is set at 65 percent. The effective percentage cost will be lower than this number when the price is rising, due to the price lag introduced in the next section.

After getting an estimate of the total mining costs, one can turn this into an estimate of the total network electricity consumption once it is known how much is being spent per kilowatt-hour (KWh). For industrial scale miners this is a limited amount, although it still differs significantly per country. It can be 4 cents per KWh in some Chinese regions, as confirmed by professional miners such as, for example, ViaBTC. The latter offers cloud mining contracts that include electricity costs of 0.35 CNY per KWh (per November 2016). This translates to roughly 5 cents per KWh in USD. On the other hand it can even be less than 2 cents in Washington State. The former number of 5 cents per KWh was chosen to calculate the index.

Limitations

By design the Bitcoin Energy Consumption Index is strongly linked to the Bitcoin price (after correcting for fees and average time between blocks), but under certain circumstances this may cause a discrepancy with the actual mining hardware in the network. Intuitively it is to be expected that an increase in available miner income is followed by more miners on the network trying to take advantage of this. It, however, might take some time before these machines can be up and running. This happens typically if price increases exceed expectations miners may already have had on the future price of Bitcoin.

For this reason, the Bitcoin price used for determining the Index is based on a moving average over the last few weeks. Should the current price fall below this moving average, then the current price becomes the one used for setting the Index value. The latter behavior captures the fact that it is easier to shut down existing hardware, than it is to add new machines to the network. Even with these adjustments, the Index may still reflect a value that is a bit too far ahead of time. This is something to consider when interpreting the Index, although it should also be realized that reality is expected to catch up soon. In the end, responding slowly to an increasing price does represent an opportunity cost for miners, so the gap is expected to exist only for a minimal amount of time.

Lastly, one may reach different conclusions on applying different assumptions. The chosen assumptions have been chosen in such a way that they can be considered to be both intuitive and conservative, based on information of actual mining operations. Within reasonable economic boundaries one might expect to find a number that is 25 percent higher or lower (given limited price volatility). In the end, the goal of the Index is not to produce a perfect estimate, but to produce an economically credible day-to-day estimate that is more accurate and robust than an estimate based on the efficiency of a selection of mining machines.

Recommended Reading

The Bitcoin Energy Consumption Index is the first real-time estimate of the energy consumed by the Bitcoin network, but certainly not the first. A list of articles that have focussed on this subject in the past are featured below. These articles have served as an inspiration for the Energy Index, and may also serve as a validation of the estimated numbers.



If you find an article missing from this list please report it here, and it will be added as soon as possible.