Skip to main content

How much energy does cryptocurrency use?

It is estimated that in 2019 cryptocurrencies consumed 70-90 TWh of electricity globally, with 60-70 TWh of that from Bitcoin mining. In comparison, all data centers worldwide used between 196-400 TWh of energy in 2020 (although the latter figure includes some crypto).

However, these numbers are quite uncertain. Some calculate a range of 60 TWh – 204 TWh for Bitcoin and 17 TWh – 105 TWh for Ethereum (as of Apr 2022) and there are many hundreds of crypto assets, all with different energy profiles.

Analysis of these other assets suggest that Bitcoin is responsible for ~66% of all crypto-related energy consumption, with the variance in energy consumption related to whether the hashing algorithm can be run on specialist equipment or not.

Table of the top cryptocurrencies and their energy
The top 20 mineable currencies with their respective algorithms, efficiencies of suitable mining devices, and rated power of the networks (as of 2020-03-27). Source: Gallersdörfer, U., Klaaßen, L., and Stoll, C. (2020). Energy Consumption of Cryptocurrencies Beyond Bitcoin. Joule.

Electricity is the main energy input into crypto mining so the changing grid mix throughout the day has an impact on the carbon footprint of the cryptocurrency being mined. This means that just like with manufacturing physical products, the country where the cryptocurrency is mined makes a difference.

In 2019, 44% of Bitcoin mining was happening in China and around 39% of all proof-of-work cryptocurrencies were mined with renewable energy, primarily hydro. However, in May 2021 China cracked down on Bitcoin which resulted in the halt of almost all mining activities in the country. Moving elsewhere, they inherited the grid mix of the countries they moved to. This resulted in the overall carbon footprint increasing:

Based on average emission factors (557.76 gCO2/kWh) and the Bitcoin network’s estimated electric load demand (13.39 GW as of August 2021), we estimate that Bitcoin mining may be responsible for 65.4 megatonnes of CO2 (MtCO2) per year.

de Vries, A., Gallersdörfer, U., Klaaßen, L., and Stoll, C. (2022). Revisiting Bitcoin’s carbon footprint. Joule.

This is equivalent to around 0.19% of global emissions, the same as Greece.

Illustration of Bitcoin energy fuels
Estimated electricity mix that fuelled the Bitcoin network from September 2019 to August 2021. Source: de Vries, A., Gallersdörfer, U., Klaaßen, L., and Stoll, C. (2022). Revisiting Bitcoin’s carbon footprint. Joule.

Crypto mining equipment waste #

Crypto mining has become a specialist activity that is performed on dedicated equipment designed for the job. Unlike more general IT equipment, there is a direct and urgent economic incentive for miners to always use the most efficient equipment. As efficiency doubles every 1.5 years (Koomey’s Law), Bitcoin mining equipment becomes obsolete at the same rate.

When buying a new consumer laptop, most of the carbon footprint is in the manufacturing. The opposite is true for servers – most of the carbon footprint comes from operating it. You get the worst of both worlds with crypto mining equipment – you combine the high manufacturing emissions with high energy consumption during the use stage, then it becomes useless within just 1.5 years.

Calculations about useful life of Bitcoin equipment across the whole network in 2018 suggested that each transaction generated 134.5 g of electronic waste, compared to 0.0045 g per Visa transaction. There are lots of assumptions in this calculation, and it is several years old, but for Bitcoin to be as efficient as Visa as a method of transaction would require improvements of several orders of magnitude.

Challenges with calculating cryptocurrency mining energy consumption #

The amount of energy required varies significantly depending on the type of underlying technology. In proof-of-work assets like Bitcoin, calculations are performed to create new blocks, which mint new Bitcoins. Those calculations consume computing power, which is where the energy is used. The reward for performing the calculations is the value of the minted Bitcoin.

The hash rate can be used to track the number of calculations happening at any given time – currently averaging near an all time high of around ~200 million terahashes/sec. The amount of energy required per hash is not constant – it is based on the computational difficulty, which varies per block – but each computation requires energy, so the greater the computational rate, the greater the energy consumption. This is the “proof-of-work”.

Not all cryptocurrencies involve mining and there are efforts to transition some, like Ethereum, to less energy-intensive methods (proof-of-stake). The official position is that moving Ethereum to proof-of-stake will reduce its energy consumption by 99.95%.

Graph of Bitcoin hash rate
Estimated number of terahashes per second the bitcoin network. From

Asset prices are another important factor because this determines how much can be spent on mining energy:

Bitcoin’s price directly effects the value of the mined coins and therefore the amount of resources miners can afford to spend on mining. Given that Bitcoin mining is a competitive market, economic theory suggests that it should cost a bitcoin to mine a bitcoin. If it costs any less than one bitcoin to mine a bitcoin, miners can profit by adding more units of computational power to the network. The opposite is also true because miners would be operating at a loss and start removing units of computational power from the network if mining costs exceed one bit-coin.

de Vries, A. (2021). Bitcoin boom: What rising prices mean for the network’s energy consumption. Joule.

Therefore, annualized energy consumption will increase as the price increases. Placing a bet on the appreciation of the price of a particular proof-of-work crypto asset is implicitly betting that the energy consumption will increase. Although it is as difficult to predict the energy consumption as it is to predict the asset price, as hash rates have generally trended up we can assume that energy consumption will also continue to increase.

Is cryptocurrency / Bitcoin bad for the environment? #

Answering this question requires you to make a judgment about the utility of a specific crypto asset, or crypto as a whole. I am generally very skeptical about most crypto assets, particularly NFTs in the form they exist today. It feels like most “investors” are actually just speculating on a future price increase and the communities around them act like cults. There are concerning examples of fraud and major security breaches, and the major innovation in digital technology is the idea of abundance and zero marginal cost whereas the value in many cryptocurrencies is from scarcity.

However, I appreciate the value of decentralization, individual ownership, and the idea of self-custody of assets (assuming the UX is well designed and individuals can manage their security correctly). It is difficult to cut through the noise and figure out if this is indeed like the early days of the internet or whether the whole thing is just a massive collection of Ponzi schemes.

We are used to the idea of consuming energy for something useful. It’s not correct to ask “is the internet bad for the environment” or “is video streaming bad for the environment”. Rather, we should ask “what is the environmental impact of internet usage” or “what impact does video streaming have on the environment”. We have decided that both the internet and video streaming are useful – the question is how do we minimize their environmental impact? This is why I spend time working on sustainable computing – computer technology is a net good, so how do we make it sustainable so that more people can benefit?

So when someone asks: is crypto / Bitcoin / blockchain / Etherum bad for the environment? We should first question whether there is any usefulness in the technology itself. If there is utility, we can consider how to minimize its impact. If not, then we should be thinking about how to minimize the use of the technology itself.

Given how long crypto has been around, and how much money has gone into it, I find it unlikely that it will all suddenly disappear. Also remember the Jevons paradox – if something is made more efficient then it will use more energy, not less. Bitcoin may not be the end game – I expect (and hope) more interesting use cases will be developed – but crypto (and web3) is here to stay. That means work needs to be done to make it sustainable.