NOTE: The methodology and underlying figures (specifically, emission rates for gold and banking) mentioned throughout are generally derived from my 2014 & 2018 works. The 2018 works is available as an easier to digest 10-part medium series, here)
Bitcoin’s energy consumption and environmental impact is something that is commented on very regularly, but ultimately, is something that is rarely understood. The majority of arguments stem from comparisons of Bitcoin to particular nation states, or some other apples to oranges comparison. Critics can’t even separate “energy use” from “electricity use”.
Over a billion people on Earth are not hooked up to the electric grid, but still have perfectly good access to energy via physical fuels. Just like they don’t need The Grid to survive, neither do Bitcoin miners, many who are now using flared methane in remote oil fields to mine Bitcoin. Of the 160,000 TWh of energy generated worldwide each year, 50,000 is lost to inefficiencies, and only 25,000 TWh of that is generated by “The Grid”. Bitcoin consumes around 120TWh, a mere 0.25% of the energy wasted each year, or, 0.5% of the World’s Grid, or, under 0.1% of Human energy production. Pick your poison with the framing.
While the energy use of Bitcoin is something trivial to define and then calculate, this is not the case for what Bitcoin is regularly compared to. Where does the environmental impact of the gold industry end? Is it with deep gashes in the Earth that open-cut gold mines leave that are so massive they can be seen from outer space? Or do we also factor in the electricity costs of all of the world’s jewellery shops and gold dealers? And the Banking System? Is it just all of the bank buildings, ATMs, and cotton, plastic and metal for notes, coins and cards? Or is it the money printing and relentless pursuit of profit at all costs which has spurred untethered consumerism and environmental destruction? Or the ability to print money out of thin air to spend trillions on state-sanctioned mass-murder? Scope definition can quickly, and emotionally, get out of hand…
The next issue is framing. There has been a dangerous conflation between “energy use” and “pollution”. To return to the example of gold, companies like Dell are now aiming to use 100% recycled gold in their computer chipsets, as, according to research commissioned by them, it is at least 99% “cleaner” than physical mining. It is “cleaner” because the use of trillions of litres of water, billions of tonnes of cyanide, and literal mountains of waste rock are avoided (among others). That said, recycling gold consumes more electricity than mining it. So is recycling “dirtier” than mining just because more CO2 is emitted? This is laughable. Further, you will never be mining gold with an electric Caterpillar D10 bulldozer — they will run on diesel for the rest of time. Gold Recycling is largely a “grid-dependent” process; as the energy grid gets cleaner and becomes emissionless, so will gold recycling (for the most part). This is also true for Bitcoin. You can expect Bitcoin’s energy use to tend to infinity in the long term, and the environmental impact to tend to zero.
So what do the actual numbers look like? Although I’ve just spent some time outlining why comparisons are disingenuous, I present the below, based on my works from 2014 & 2018, and indexed to today’s numbers.
2020 saw ~3500 tonnes of gold mined, and ~1300 tonnes of gold recycled. According to Dell, each kilogram of mined gold produces 20 tonnes of CO2 and 48.6 MWh of Energy, but they omit the energy needed for further refinement of gold due to data limitations. A new study from De Paul University suggests that this should be closer to 35 tonnes of CO2 for jewellery, which is roughly 50% of the entire Gold demand. It is assumed that refinement will use similar power to recycling due to the industrial process involved. Therefore, a total of 79.9 MWh is used for each gram of gold in jewellery (mining + refining). Each kilogram of recycled gold produces 37 tonnes of CO2 and 31.3 MWh of Energy. This brings the gold mining industry total for 2020 to 265TWh and 145 Mt of CO2 if you use the De Paul numbers and account for 1750 tonnes of Jewellery.
In 2014, I calculated a figure of 2380 million GJ (or 660 TWh) of energy used for the banking system’s branches and ATMs, alongside cotton, plastic and metal for notes and coins. Without too much wizardry, we can conservatively assume this has increased by 1–2% per year, and we can round up our figure to 700 TWh, and about 400 Mt of CO2 produced.
I’m very comfortable using the Cambridge Bitcoin Electricity Consumption Index (CBECI), as their methodology mirrors mine from 2014 and 2018. According to the CBECI, Bitcoin uses 113 TWh of energy per year. The proportion of renewables used is unknowable, but basic economics dictates that miners will continue to chase the cheapest power available, and this is increasingly becoming renewable. There is little doubt that Bitcoin will be running on a majority combination of stranded, waste and renewable energy by the end of the decade. That said, using the current global average grid figure of about 0.6 tonnes of CO2 per kWh electricity produced, this leaves Bitcoin with 70 Mt of CO2 emitted.
Although the comparisons are not like for like, and with only Bitcoin having a 100% defined scope for this exercise, we can say that, at a minimum, Bitcoin consumes/emits less than half of what the gold mining industry does, and less than one fifth of what Bank branches and ATMs do.