Demystifying the Environmental Impact of Cryptocurrencies

by Ryan Shea

Key Take-aways

• Common knowledge has it that cryptocurrencies are environmentally unfriendly.
• Based on market cap, 40% of cryptocurrencies already use much less electricity intensive consensus protocols. This is set to increase to around 60% when Ethereum makes its switch to Proof-of-Stake.
• Usual estimates of Bitcoin’s negative environmental impacts are overstated.
• Cryptocurrencies are a lot closer to ESG criteria than generally considered and, could well become sustainable in the not too-distant future.

“Cryptocurrency is a good idea on many levels and we believe it has a promising future, but this cannot come at great cost to the environment.”  Elon Musk, Tesla CEO

Late last month Ripple co-founder Chris Larsen, together with Greenpeace, launched a “Change the Code, Not the Climate” campaign to put pressure on Bitcoin and its community to follow in Ethereum’s footsteps and transition away from computationally (energy) expensive proof-of-work (PoW) in favour of less computationally expensive proof-of-stake (PoS).

Let me be polite and say that this campaign announcement has not gone down well with the crypto community. This is important because, as I pointed out in my last article discussing the evolution of Bitcoin narratives, it relies upon the consensus of its community to enact changes and judged by this reaction the consensus is firmly against moving away from PoW to PoS. In short: Chris Larsen is just wasting $5 million funding a campaign that will never succeed. None the less it does keep the public gaze on the environmental impact of cryptocurrencies, a topic investors need to understand more deeply given that ESG is becoming a increasing factor in the investment decision-making process.

But how environmentally unfriendly are cryptocurrencies?

Consensus Protocols

Environmental concerns are predominantly focused on the amount of energy expended in mining cryptocurrencies. I am sure you are all familiar with the media headline stating that Bitcoin mining uses the same amount of electricity as [insert country of choice]. However, as I have noted before cryptocurrencies come in a wide variety of flavours. Bitcoin uses PoW as its decentralized consensus mechanism to validate transactions and mine new tokens. Ethereum, the second largest cryptocurrency by market cap, uses the same method, at least for now. But, this is by no means the only consensus protocol. There are a plethora of them including, but not limited to, proof-of-stake (PoS), proof-of-authority (PoA), proof-of-burn (PoB) etc etc and these alternative methods consume considerably less power. For example, a transaction on Tezos, which deploys a PoS protocol, is lower than either Bitcoin or Ethereum by over a factor of two million.

Using market cap as a measure of a cryptocurrency’s popularity out of a total of $2.1tr, roughly 60% use PoW at present. As alluded to above, Ethereum is continuing to transition its protocol to the much less energy intensive PoS. When complete this will mean only 40% of cryptocurrencies measured by market cap will be using the energy intensive PoW protocol. The vast majority of this relates to Bitcoin as the market cap of cryptocurrencies other than Bitcoin or Ethereum using PoW  is just 2.5% of the total. Hence, as and when Ethereum makes the switch to PoS, the question as to how environmentally unfriendly cryptocurrencies becomes a narrower question of how environmentally unfriendly is Bitcoin?

Environmental Evils

The typical source cited for Bitcoin environmental impact statistics is the platform Digiconomist created by Alex de Vries, a researcher at the School of Business and Economics at the Vrije Universiteit Amsterdam, as a hobby project.

Despite being widely quoted, the reliability of is particular source of information on Bitcoin’s power usage is open to question. To understand why consider the following chart, which plots  Digiconomist’s estimate of Bitcoin’s network annual electricity demand compared with the University of Cambridge Judge Business School’s alternate daily electricity demand estimate and Bitcoin’s global hash rate published by blockchain.com.

Bitcoin Power Usage vs. Global Hash Rate (Log scale)

One would naturally expect the three series to show similar trends and while this is true generally, a notable divergence occurred last year when China banned domestic mining, which due to the dominance of Chinese miners led to a sudden and dramatic drop in overall Bitcoin mining activity. This decline in computational power behind the Bitcoin network, as evidenced by the drop in the global hash rate (green line), was reflected in the Cambridge business school’s power estimate (blue line) but doesn’t register at all on the Digiconomist estimate (red line). Strange. The reason given for this lack of downward movement was explained in a July 2021 blog post. To wit,

“such a large drop in active equipment also affects the energy consumption of the network. Unfortunately, it’s never possible to say by how much exactly. A 50% drop in the computational power does not necessarily equal a 50% drop in energy consumption.

Whatever the precise amount may be, neither the Bitcoin Energy Consumption Index [Ed note: The Digiconomist estimate] , nor any other model or tracker, is equipped to properly deal with a network disruption of this magnitude. Because of the previous challenges in determining the most likely energy consumption impact, any adjustment would be arbitrary. For this reason, no adjustments were made to reflect immediate impact of the ban. [my emphasis]”

What this episode underlines is the extent to which the Digiconomist estimate of Bitcoin’s electricity consumption is “manufactured” in the sense that it does not make use of the multitude of publicly available data generated by virtue of Bitcoin being on a decentralized blockchain, which is in marked contrast to the approach followed by Cambridge Business School. Whether this is due to Digiconomist’s attempt to be intellectually rigorous (when in doubt, leave it out) or it reflects his personal negative bias towards cryptocurrencies (a bias all-too-evident when one reads the numerous blog posts on the Digiconomist website) is open to debate. What is certainly should do is raise questions about the credibility of his electricity consumption estimates, something that we - and we hope others - will take into consideration when seeking to assess Bitcoin’s environmental impact.

Carbon Footprint

That said, let’s take Digiconomist’s electricity consumption numbers at face value. The annual energy consumption of Bitcoin is on their estimates 204.5 Twh. Converting this to C02 production relies upon knowing how electricity used by the miners was generated. Given the decentralized nature of the Bitcoin network there is no straightforward way to determine this information. It must be estimated. The method used is to attempt to identify the location of the Bitcoin miners from their IP addresses and then combine that information with the carbon intensity of electricity generation at the given location. This is not a perfect method given VPN services can be used to disguise the country of location, and is almost certainly being used by Chinese miners seeking to circumvent its government’s mining ban, but it is only practical way to attempt this exercise. Based on these inputs, Digiconimist estimates fossil fuels (coal, oil and natural gas) accounts for 60% of Bitcoin’s energy consumption – see chart. The remaining 40% comes from hydroelectric, nuclear and other renewables.

Bitcoin Mining Electricity Source

The chart below shows how this compares with the energy source breakdown for both the world in terms of primary usage and electricity generation. The split between fossil fuels and renewables for the Bitcoin network is very much in keeping with the global trend. Indeed, the only substantive difference in energy make-up is the dominance of oil in terms of world primary usage, which for the most part reflects its importance in global transportation. Bitcoin is, in other words, nothing special.

Energy Sources: World vs. Bitcoin

Based on these electricity generation shares and the consumption figures for the Bitcoin network as a whole it is possible to generate an estimate as to how much CO2 Bitcoin mining produces. Digiconomist estimates Bitcoin’s carbon footprint to be 114.06 million tons per annum which, deploying the old country comparison method, approximates to the carbon footprint of the Czech Republic. That sounds like an awful lot of CO2. But how much is it in comparison with other sectors of the economy?

Thankfully, Digiconomist provides us with a useful comparison metric. In a blog post published in 2017 examining the carbon footprint of a coal-powered Bitcoin mine in Inner Mongolia for comparison he noted that

“on an international flight a Boeing 747-400 typically emits 92 kg CO2 per passenger per hour, given that all seats are taken. On 416 seats that’s 38 tons of CO2 per hour in total.”  

Bitcoin is estimated to use between 140-204 Twh per annum. According to the IEA, US electricity production in 2019 generated 475g of CO2 / kWh, so assuming the electricity used by Bitcoin miners is approximately equivalent, this means that Bitcoin produces 7,591-9,711 tons of CO2 per hour – or in 747 terms Bitcoin’s network produces the equivalent C02 to 175-300 aircraft flying continuously (the lower figure is based on the lower Cambridge estimate of 140 TWh). Compare that number with the number of planes actually in our skies. According to FlightAware there are somewhere between 7,782 and 8,755 commercial aircraft flying at any one time – see chart. In other words, from the perspective of CO2, Bitcoin is equivalent to about 2.5% of global air travel.

Map of Global Aircraft In Flight

Regards global C02 emissions, the latest data published by World Bank puts the global estimate at around 45bn metric tons of C02 equivalent per annum. That’s a big number compared with  Bitcoin’s 60-100mn metric tons. In fact, the Bitcoin network accounts for about 0.12-0.18% of global CO2 emissions.

When looked at in these terms, rather than the more provocative country comparison, it is clear that in the grand scheme of things the Bitcoin network is not a major factor in CO2 generation. For countries seeking to achieve their 2050 net zero goals, curtailing air travel will be much more effective than banning Bitcoin mining as the later is barely more than a statistically-insignificant rounding error.

Of course, this only provides a snapshot of Bitcoin mining’s current C02 emissions. Looking ahead, this is likely to change quite markedly for the simple economic reasons. Electricity constitutes a great deal of a Bitcoin miner’s operating costs so they are strongly incentivised to find the cheapest source of electricity. In light of technological advances over previous decades, the cost of electricity from non-carbon sources has fallen sharply and is now comparable, if not cheaper, than fossil fuel equivalents (even more so this year given the huge impact Russia’s invasion of Ukraine and the resulting sanctions have had on oil and gas prices). It therefore makes financial sense for Bitcoin miners to transition away from electricity generated from higher cost fossil fuels to lower cost renewables.

LCOE ($ per MWh)

Obviously, the same cost incentive applies to all industries, however, transitioning to greener energy is not easy. For many industries it is nigh-impossible to substitute their primary energy source, for example, commercial vehicles and airliners require fossil fuels as there is no large scale green alternative at present. Ditto for many energy intensive heavy industries.

Conversely, for Bitcoin miners it is relatively straight forward. Electricity is electricity regardless of how it is produced and Bitcoin mining rigs are geographically mobile. This allows them to shift location in order to source the lowest electricity prices. Indeed, the cyclical rise in hydroelectric in the earlier chart showing the Bitcoin mining energy source was attributable to Chinese Bitcoin miners relocating their rigs during the wet season to Sichuan province to benefit from the low electricity prices due to the proliferation of hydroelectric capacity in the region.

This mobility also allows them to extract electricity from other forms of “stranded” energy, such as flared gas - a source of energy that is often hard to extract commercially but which adds to CO2 emission stats regardless. Indeed ExxonMobil recently announced that it is running a pilot project with Crusoe Energy Systems to turn otherwise wasted flare gas into electricity for mining Bitcoin. Because few other companies are able to utilize such “stranded” energy sources, the electricity is provided at low cost and as an added bonus they tend to have lower carbon footprints than electricity generated from more standard sources. This confluence of factors means Bitcoin mining is not only likely to become greener over time, but market forces as opposed to government “green” incentives - such as tax breaks - will be the driver.

E Waste – Another Biggie...

Obviously electricity consumption is not the only area where Bitcoin mining can have a detrimental impact on the environment. As Digiconomist correct notes, there is the related issue of e waste produced when mining rigs come to the end of the economically useful life cycle. According to his estimates, again widely quoted in the media, Bitcoin miners generate 30,700 tonnes of e waste per year. To put this number into context, the typical country comparison is used and in this case it equates to the small IT equipment waste of the Netherlands.

Sounds bad. But yet again Digiconomist has overstated his case. This estimate is based on mining devices having an average lifespan of only 1.29 years. However, Compass mining – a Bitcoin mining service provider who should naturally know a thing or two about Bitcoin mining rigs - on its website makes clear that ASIC mining machine...

“operating in well-managed facilities can last for many years. Three to five years is typically a machine’s average lifespan, although even longer periods aren’t unheard of. Newer models are expected to have at least five-year lifespans.”

These machines are not cheap, with one capable of generating hash power around the 100 TH/s costing around the $10,000 mark. Aside from electricity (which Cambridge business school estimates as accounting for roughly 80% of operational expenses for miners) this is the other major cost associated with mining so it makes sense that miners do their utmost to maintain the longevity of their rigs in order to maximise their profitability.

The website Asic Miner Value provides real time estimates of profitability per mining rig. It confirms that for those rigs designed to mine on the SHA-256 encryption algo used by Bitcoin, the majority of rigs released since late 2018 remain profitable based on the average cost of electricity in the US of 13 cents per Kwh (note this is 8 cents higher than the estimate Digiconomist uses in his electricity consumption calculations meaning that even older mining rigs would still be profitable). While newer mining rigs are more profitable, it seems unlikely that older rigs will be scrapped if they are still profitable and certainly not after just 15 months as Digiconomist assumes.

In addition, the Digiconomist estimate does not take into consideration recycling of no longer economically useful mining rigs: 2.9 million mining rigs, weighing 13kg each, junked every 15 months equates to 30.7 kilotons of e waste annually. However, the bulk of the 13kg weight of a typical Bitcoin mining rig is the metal casings and aluminium heat-sinks. In a world where metal prices have been rising sharply, these are increasingly valuable commodities and, unsurprisingly, a number of companies offer Bitcoin mining recycling services.

… Or Not

If instead we assume a more realistic number for the economic life span of a Bitcoin mining rig, say three years instead of 15 months, the estimated e waste from Bitcoin mining drops for 30.7 kilotons annually, to under 13 kilotons. Incorporating a 50% recycling ratio by weight, which as mentioned is very do-able, then Bitcoin mining would generate only 6.4 kiltons of e waste annually – a fifth of the Digiconomist estimate based on perfectly reasonable alternative assumptions.

Now, 6 kiltons of e waste is not great, it’s still a lot. But, it is important to get some perspective. According to the latest estimates from the World Economic Forum, globally we generate 50 million metric tonnes of e waste per year, or expressed in the same units I used for Bitcoin mining e waste 50,000 kilotons. Even without assuming any recycling, just a more accurate lifespan estimate, Bitcoin mining contributes less than 0.02% of global e waste.

So, in conclusion, despite all the hype about the environmentally unfriendly nature of cryptocurrencies, the reality is nowhere near as bad as generally presumed. Once Ethereum makes the switch of PoS, electricity demand will mainly be driven by Bitcoin, whose share of global CO2 emissions and e waste is miniscule. Moreover, technological innovation, the geographic flexibility of miners and their high price sensitivity to electricity costs creates an incentive structure for Bitcoin miners to embrace greener renewable energy sources without the need for government subsidies. If only the same could be said of many other industries.

Cryptocurrencies, and Bitcoin in particular, may not satisfy most people’s criteria for considering it an ESG compliant asset class, but the reality is that it is a lot closer than generally considered and given the likely direction of travel it will be in the not too-distant future.

Until next time.

Ryan Shea, crypto economist at Trakx

Footnotes

[1]    See: https://trakx.io/bitcoin-the-inside-out-narrative/

[2]    PoW is easier to implement, especially in a decentralized set-up. It is also considered to be more secure, because Bitcoin validates transactions after mining. Miners are not trusted to provide valid transaction. This task is performed by the multitude of nodes in the Bitcoin network. In PoS one must trust the staker as the validator, which given the way blocks are created has a centralizing effect because the probability of being selected to validate a block is based on the number of coins staked by the validator. Ethereum sets the minimum stake to be a validator at ETH 32 – or  $103,000. Such a high cost threshold has raised some concerns about how equitable or inclusive the PoS protocol is. Moreover, PoW has a 13 year track record, unlike PoS which is still in its infancy.

[3]    This is not a typo, it is left blank deliberately. Various countries have been mentioned in the media at different times.

[4]    See: https://tezos.com/carbon/

[5]    For latest numbers – see: https://coinmarketcap.com/

[6]    Recognizing the long-standing joke that this switch has been six months out for years.

[7]    See: https://digiconomist.net/how-chinas-bitcoin-mining-ban-affects-energy-consumption-estimates/

[8]  Digiconomist uses mining profitability, an estimate of the percentage of revenues spent on electricity and an assumption on the cost of electricity (US 5 cents per Kwh) to generate their energy consumption estimate.

[9]  Especially journalists in the mainstream media.

[10]  For example, Germany and Ireland have higher hash rates than seems plausible as neither country is known to have large scale mining operations. This suggests miners (probably Chinese) are using VPNs to redirect IP addresses to these two countries – see: https://ccaf.io/cbeci/mining_map

[11]  The Bitcoin Mining Council, a voluntary forum for Bitcoin miners, Q4 survey which covers 46% of the global Bitcoin network puts the renewables figure at 66% -  see: https://bitcoinminingcouncil.com/q4-bitcoin-mining-council-survey-confirms-sustainable-power-mix-and-technological-efficiency/

[12]  See: https://www.iea.org/reports/global-energy-co2-status-report-2019/emissions

[13]  Lowcarbonpower.org provides estimates for carbon emission per energy source per kWh. Using these inputs weighted up by the actual estimated energy sources for Bitcoin miners provided by Digiconimist the best guess for Bitcoin carbon emissions per kWh is just over 400gCO2/kWh, so if anything the estimate I used is high – see: https://lowcarbonpower.org/

[14]  See: https://flightaware.com/

[15] It shows 10,539 aircraft in the air. This was late afternoon (UK time) on a Wednesday in early March. Not exactly a travel hotspot.

[16] Unless the drilling site is closely located to a gas pipeline, it is uneconomic to ship the gas so it gets either flared or simply released into the atmosphere. The latter is particularly problematic as methane is 25 times more potent as a greenhouse gas than CO2.

[17]  See: https://www.bloomberg.com/news/articles/2022-03-24/exxon-considers-taking-gas-to-bitcoin-pilot-to-four-countries

[18]  I have come across arguments stating that Bitcoin mining makes it profitable to overbuild renewable sources of energy production, since it allows that surplus supply to be monetized, and that helps mitigate the sporadic nature of many types of renewable energy. While I agree with the monetization aspect I am not convinced that overbuilding intermittent renewables is the way forward, absent improving storage methods, ie batteries. Yes it can help balance intermittent supply and bottlenecks on the electrical grid, but it doesn’t do anything to provide power when the wind is not blowing or the sun is not shining. This is something I need to ponder more on. If anyone thinks they can convince me of this argument please get in touch.

[19]  Google bitcoin e waste and his 2021 article is the source material for every single hyperlink on the first page of results searched. He clearly knows how to market himself!

[20]  Tera hashes per second – see: https://www.buybitcoinworldwide.com/mining/hardware/

[21]  See: https://www.asicminervalue.com/

[22]  See: https://www.bitnand.com/miner-recycling-program

[23]  See: https://www.weforum.org/agenda/2019/01/how-a-circular-approach-can-turn-e-waste-into-a-golden-opportunity/

[24]  After I finished my Bitcoin narrative piece, I came across a new narrative: Bitcoin as a national security tool. The narrative has been espoused by Jason Lowry, who is a commissioned officer in the US Space Force. It’s an intriguing idea – see: https://bitcoinmagazine.com/culture/can-bitcoin-miners-replace-global-consensus