A primer for journalists and independent researchers
It is not uncommon for nascent technologies to be accompanied by claims that are not based in fact. Bitcoin mining is no exception to this trend. As a result, when researchers are attempting to understand Bitcoin, it’s important to distinguish the claims which are fact-based from ones that aren’t.
Below are some of the most common incorrect, and no-longer-correct, claims about Bitcoin mining, together with the correct version of each claim which is supported by data.
Section A: Claims that are no longer accurate
1. “Bitcoin mining reopened mothballed fossil fuel plants”
There are 2 examples mistakenly used by journalists, Greenidge: Lake Seneca Gas Plant, and a Hardin, Montana Coal Factory.
In fact as we’ll see, no Bitcoin mining currently occurs on resurrected fossil fuel plants
1a. The Greenidge Gas plant was not resurrected to mine bitcoin (this was misreported)but to supply power back to the grid, and did not commence bitcoin mining until two years later.
The SEC filings confirm this.
Context: We can see this evolution in the S-1 form with the SEC, here, which provides a detailed look on when and how the operation grew from a then inactive power facility. Greenidge brought the now natural gas facility to the wholesale energy market in 2017 as a merchant power provider, under newly issued air and water permits. There was no crypto mining operation, it was just a natural gas power plant when Greenidge restarted operations.
Additionally, you can see Greenidge’s initial press release announcing its exploration into bitcoin mining dated March 5, 2020 here. It started with 1MW of crypto-mining in 2019.
1b. The Montana coal factory is the other facility often cited as an example of Bitcoin mining perpetuating fossil fuel power. However it has not been mining bitcoin since late 2022. It was fully migrated to a wind farm in Texas.
Evidence: The company journalists are referencing is Marathon Digital Holdings. However in May 2022, Marathon announced their intention to fully migrate from their Montana coal site. At the AIM conference in Dubai in Dec 2022, Marathon CEO Fred Thiel confirmed that Marathon had fully migrated its coal power source onto a Wind Farm in Texas.
2. “Bitcoin’s carbon footprint is over 60 Mt CO2e, and increasing”
Alex de Vries is frequently cited in mainstream media for having published a paper which showed that Bitcoin uses 65 MT CO2e emissions, and that Bitcoin’s water footprint is enormous.
Less known, for it was not covered in mainstream news, is that Alex de Vries’/Digieconomist’s work was recently called to question in peer reviewed research (Sai and Vranken, 2024). His method of using a “per transaction” metric for measuring water and energy usage was dismissed as “not a meaningful metric” by Cambridge University in 2018.
The “papers” that journalists refer to are in fact commentaries (shorter pieces that academic journals are not obligated to go through the peer review process)
For example, De Vries’ academic commentary that showed 65 MT CO2 emissions from Bitcoin was published in Mar 2022, so contains data that is no longer contemporary.
3. “Most bitcoin mining is done using fossil fuels”
This was true prior to the exodus of mining from China, but is now 56.5% sustainable. There are no contemporary studies that still claim that bitcoin mining is mainly fossil fuel based.
Context: In September 2023, Bloomberg Intelligence stopped using the Cambridge model, in preference to an open source model which I developed over a 2 year period to build on exclusions in the Cambridge Model (see model methodology). They stated “Batten’s model appears to be more accurate”. At the time the sustainable energy percentage was 52.6%. This has since grown to over 55%, as can be viewed in this live chart.
The Chamanara et al paper that is often referenced, while written in 2023, used a 3-year old data set (see abstract below). Like political opinions, Bitcoin mining energy source is a highly dynamic system, as we can see from Cambridge’s miner map. As such, using a three year old data set is akin to attempting to predict the result of an election using a political poll taken three years ago.
4. “Mining got dirtier since migrating from China”
A careful examination of the data tells us that the opposite happened, as we can observe from this recent chart from Woonomic
Most mining in China were not hydro based, they were mostly coal based (migrating to hydro during the minority of the year when it was wet season), so when some miners migrated from China, it was net decarbonising.
Source: Forbes journalist Roger Huang observed: “It’s been two years now – and it should be said that Bitcoin and the hash power built on securing it is stronger than ever.
In an ironic twist of fate for the discussions about bitcoin’s environmental impact, it’s also highly likely that shifting from China’s energy mix has helped make bitcoin more “green” – and emit less carbon emissions.”
Journalists accounts should not be relied on as a sole source of evidence, so I offer a second piece of supporting evidence.
This conclusion is backed up by Cambridge Centre for Alternative Finance, who acknowledge that the age of their data set (last updated Jan 2022) impacts emissions estimates, and state in their methodology for calculation Bitcoin emissions: “ today’s electricity mix associated with Bitcoin mining is likely markedly different from that of January 2022. Based on initial data, this discrepancy likely leads to an overestimation in our emission estimate by approximately 25%. “
Source: Cambridge Centre for Alternative Finance
5. “Mining moved to cheap gas in US and Kazakhstan” after the exodus from China
This is inaccurate, as mining only briefly moved to Kazakhstan before finding a permanent home in the US. Most mining moved from China to the Texas grid which has a high and increasing percentage of renewable energy. The migration to Kazakhstan was temporary and Kazakhstan currently has under 3% of global hashrate. Miners re-migrated after that to countries such as Paraguay, Ethiopia, Uruguay which are predominantly hydro-powered.
Sources: There are two contemporary mining maps in existence, giving us an estimated range of (0.55%-3.5% for Kazakhstan. Well down on the 13.2% when Cambridge last updated their mining map.
- Bitriver estimated in 2023 that Kazakhstan had 100MW of mining remaining (at the time, Global hashrate was ~18GW, which puts Kazakhstan at 0.55% of global hashrate.
- Hashlabs estimate that 3.5% of global hashrate still exists in Kazakhstan.
6. “Cambridge has the most accurate assessment of Bitcoin emissions”
The emissions estimations of 65 Mt CO2e are now acknowledged on Cambridge’s own website to be overstated due to the age of their dataset (30 months old, as acknowledged in their emission estimate notes)
Again, more up to date mining statistics can be found here
- Hashlabs Miner Map (2024)
- Bitriver Miner Distribution (2023)
de Vries estimated a similar amount in 2021 in a commentary in Joule Magazine. However. as we can see in the mining map graphic, this figure was based on the assessment that 25.2MT CO2e came from Kazakhstan, which is no longer contemporary.
A more accurate methodology which uses the Cambridge Model as its base, but uses more contemporary data and factors in the impact of offgrid bitcoin mining shows that emissions are significantly lower, and have remained static over a four year period despite rising hashrate.
7. “Bitcoin must become/ is struggling to become environmentally friendly”
Framing bitcoin as not environmentally friendly is an inaccurate assessment. All nascent technologies, including the solar industry had carbon debts for a long time before they paid it off. In solar’s case it did not pay off its carbon debt until 2011 (Louwen et al), more than 50 years after the invention of the silicon based solar cell in 1954. It’s critical that we assess not only current carbon footprint but trend line in making an assessment about whether a nascent technology is net positive.
The idea that Bitcoin is struggling to go green was based on a 2022 Reuters article, which used now debunked Cambridge University data to show only marginal gains in sustainable energy use.
8. “A lot of cheap energy that can be used for bitcoin mining has high carbon emissions”
This is no longer an accurate statement. Today, the cheapest forms of electricity are generally from sustainable sources (Ray and Douglas). Bitcoin’s drive to use cheap energy therefore coincidentally also drives it to use green sources of energy. This trend will likely continue into the future, given that renewable energy is trending cheaper with time.
Section B: Claims that were never accurate
9. “Bitcoin uses as much energy as whole countries”/ “Bitcoin uses too much energy”
Cambridge University has discredited this type of comparison, saying that is not only not meaningful to compare countries to industries, but that it is an example of “presenter bias”.
The “Bitcoin uses too much energy” claim is usually used by critics interchangeably with the notion “it is bad for the environment”. This is a highly dubious claim, based on the invalid assumption that more energy use equates to environmental damage. The immediate follow-on logic is generally to suggest “moving to proof of stake”. But putting aside the question of the feasibility of the suggestion, there is a more fundamental issue with this logic: “energy use” is not synonymous with “environmental damage”.
There are seven questions that we should ask in assessing whether any technology’s energy use is net positive or net negative to the environment
- Was that energy from sustainable sources, and is its use of sustainable energy growing?
- Was the energy from an otherwise wasted/stranded renewable energy source (such as curtailed solar energy), or emission negative source (such as vented landfill gas)
- Did that consumer of energy obviate a more carbon intensive technology (for example, Electric Vehicles obviating fossil-fuel based transport)
- Did the energy consumer recycle the heat produced, in a matter than measurably decreased their prior reliance on fossil fuel based heating?
- Did the energy consumer render a renewable economically project viable that otherwise would not have occurred.
- Was the energy consumer flexible in their use patterns (flexible users have been shown to help grid operators decarbonize grids by being able to place more intermittent renewable energy onto them)
- Did that technology have measurable environmental benefits from its use (for example, solar has an emission intensive manufacture process due to the coal furnaces required to melt silicon, but over its lifetime will pay off its carbon debt 20x)
In Bitcoin mining’s case, there is strong evidence both in real life applications and in the peer reviewed literature that all seven of these questions have positive answers, suggesting that Bitcoin’s energy use is tracking towards, and may already be, net-decarbonizing to the environment.
10. “Bitcoin could improve its energy efficiency by moving to Proof of Stake”
The argument that Bitcoin should move to Proof of Stake to mitigate environmental impact is like suggesting that a child be given a prefrontal lobotomy in order to curb her energy levels: a non-starter.
It generally follows on from the incorrect assumption that “using energy” is tantamount to causing environmental damage, which as we’ve seen in 1. above, it is not.
Of those who have taken the time to research what “sound money” is and why “proof of work” is integral to the creation of sound money, no one has advocated Proof of Stake. Similarly, for those who have taken the time to research Bitcoin’s positive environmental externalities, no one to my knowledge has ever advocated a move to Proof of Stake. It is in short, a solution only suggested by those who have not researched Bitcoin or Bitcoin mining deeply.
Changing the consensus algorithm would break money as a decentralised form of money. People have already created bitcoin with a Proof of stake algorithm, but almost no one uses it because the value of the technology was broken in the eyes of the Bitcoin community in the process.
This graph shows the existence of Bitcoin Proof of Stake. It currently trades at $0.008219 per coin, compared to Bitcoin Proof of Work, which currently trades at $69,602.10 per coin. The price difference (market cap of 8.4 Million times less) is a rough reflection of the relative number of users of each coin.
Because Bitcoin uses a peer-to-peer consensus mechanism, all users are free to issue a BIP (Bitcoin Improvement Proposal) at any time. The fact that BitcoinPOS has been proposed but not adopted suggests little appetite for the change in the Bitcoin community. The primary reason for this is a consensus view that Proof of Stake is insufficiently decentralized to allow Bitcoin to remain “sound money”. A balanced take on when POS and when POW is the better system can be found here.
Secondly, even if it were academically possible the move is ill advised for environmental reasons. Moving Bitcoin to move to Proof of Stake would mean that Bitcoin no longer had the ability to mitigate methane, provide stability to the energy grid, increase renewable energy capacity, monetize wasted renewable energy, support renewable development and climate action, or facilitate clean microgrid development, as demonstrated in the linked peer reviewed research and independent reports.
So in short, Bitcoin would lose its ability to be sound money, and lose its ability to ever contribute to the environment by a move to Proof of Stake.
11. “Bitcoin uses a lot of energy/water per translation” (and other research quoting Alex de Vries/ Digiconomist)
This claim, propagated widely by Alex de Vries (also known as Digiconomist) and widely reported, was never accurate. The claim was first rebutted by Cambridge Judge Business School, then debunked in a peer review paper by Sai and Vranken titled “A critique of Bitcoin energy consumption estimates”. Their paper criticizes the data gathering techniques used by Alex de Vries in his estimations of Bitcoin’s energy consumption, sustainable energy mix and emissions. Here’s a breakdown of their critique:
- Limited Transparency: Sai and Vranken argue that de Vries’ methods lack transparency regarding the specific data sources he uses for his calculations. This makes it difficult to assess the accuracy and reliability of his estimates.
- Focus on Hashrate: De Vries’ method relies heavily on the Bitcoin network’s hashrate (computational power) as a key data point. Sai and Vranken argue that this oversimplifies the relationship between hashrate and energy consumption. Other factors, like the efficiency of mining machines, also play a role.
- Inaccurate Efficiency Modeling: Sai and Vranken argue that de Vries’ modeling of Bitcoin mining machine efficiency may not be fully accurate.
- Reliance on anecdotal examples to back up model assumptions
- Use of small samples, assumed to be representative of a whole system. For example, De Vries’ uses a small (34–40%) geographic dataset from Cambridge and presents the results for the whole of the network
- Making assumptions not backed up by empirical evidence. For example,
- de Vries’ study which assumed old devices (specifically Antminer S9) were still the dominant mining hardware, without empirical validation
- assumption without evidence that the life cycle of Bitcoin hardware is “around 1-2 years”
- assumption without validation that “electricity consumption only accounts for 60% of all revenue costs”
de Vries’ work was also called to question by Cambridge Judge Business School, who called his method of estimating Bitcoin’s resource use per transaction “not a meaningful metric”. Sai and Vraken’s paper confirms that this is indeed a faux metric, as Bitcoin energy use does not come from its transaction. Further, it is semantically misleading, as “laypeople” outside the Bitcoin community are unaware that on the Blockchain one Blockchain “transaction” can contain billions of payments.
The paper suggests that de Vries’ data gathering techniques might not be comprehensive enough to capture the full picture of Bitcoin’s energy consumption. This lack of transparency makes it difficult to verify the accuracy of his conclusions.
12. “Bitcoin using more renewable energy does not make bitcoin good for the environment”
This is untrue because there are ways, unique to bitcoin, in which the technology uses renewable energy that do make it directly good for the environment. For example, its flexible load consumption of electricity has been shown to have a direct decarbonising impact on the grid (see below). Similarly, Bitcoin’s unique ability to profitably methane mitigation has a net emission reducing impact. Thirdly, consequence of Bitcoin mining flexibility is that it can obviate gas peaker plants (see below). Peer review research has also shown that Bitcoin can allow entire grids to be built out using higher proportions of renewable energy than would have otherwise been possible
Evidence:
Bruno et al found that Bitcoin mining utilizing demand response was almost carbon neutral compared to the alternative, using a gas peaker plant without Bitcoin mining.
However, Bitcoin mining did not impose the proposed $8 Billion price-tag to the grid of Texas. (Texas’ grid ended up using Bitcoin mining in 2021 to give demand side flexibility, and did not purchase these additional gas peaker plants).
Meanwhile, an independent report from Rhodes et al found that Flexible datacenters such as Bitcoin mining allowed the grid to be built out using a higher concentration of renewable energy.
The Whitehouse’s OSTP report (2022) reported on Bitcoin mining’s ability to mitigate methane, stating “crypto-asset mining operations that capture vented methane to produce electricity can yield positive results for the climate, by converting the potent methane to CO2 during combustion.” This finding was echoed by the Institute of Risk Management.
13. “Bitcoin may throttle down its power-use, but it strained the grid with high energy demand the first place”
This view has been proffered by Rhodes. However, whose thesis is that Bitcoin mining should not be credited for providing stability to the grid by powering down in times of high demand, because if it wasn’t there, the grid wouldn’t have reached the same level of strain.
However, this seemingly convincing argument is overly simplistic because Bitcoin mining does not merely create more demand on the grid in isolation. Bitcoin mining has also been shown both in peer reviewed research and by ERCOT grid operator Brad Jones to incentivize the development of more renewable-energy supply in tandem.
The best analogy is that Bitcoin mining permanently opens up an extra lane of highway, uses that lane itself at offpeak times, but frees it up at onpeak times by taking the nearest offramp.
The end result is less congestion for everyone on the grid.
At the same time, because it has brought more renewable energy onto the grid, it has helped grid owners achieve decarbonization targets and stability targets in parallel, something very difficult to do without Bitcoin mining.
14. “Bitcoin may be trending to use more clean energy, but all this means is that it is wasting energy much more efficiently.”
This statement has never been accurate because, as peer reviewed literature and case studies show, bitcoin mining has prevented the waste of significant quantities of renewable energy.
For example, Moghimi et al found that traditional Microgrids produced a large amount of excessive energy, which is wasted without usage due to mismatch between renewable generation and demand. However when Bitcoin mining was integrated it prevented almost all energy waste, while decreasing microgrid operating costs by 46.5%.
Similarly, Lai and You found that Bitcoin mining prevented the significant waste of power often associated with wind and solar due to being produced at the wrong time of day, or in the wrong location without adequate transmission capacity. They found that “Leveraging crypto as virtual carriers can enhance renewable capacity and potentially achieve up to 98% and 92% utilization of available solar and wind power, respectively.”
Further, “wasting energy” is a value judgment is it not? One can only claim that energy is wasted if no good to humanity is produced in the process. There are 18 recorded social benefits rendered possible by bitcoin mining’s Proof of Work Algorithm.
Full list of 18 with references to sources
15. “Nation states around the world such as China are banning Bitcoin Mining”
China did not ban bitcoin mining. 15% of global hashrate still comes from China and is mostly renewable based. It did however put a stop to coal-based bitcoin mining and some larger operations that were being used to get money out of China.
Source: (We can see from Cambridge’s mining map that after a one month “suspension”, mining in China resumed and had recovered to 19% by Dec 2021 (the last data update to their model).
It is currently estimated by Hashlabs to be 15% of global hashrate.
16. “Bitcoin mining uses energy that could be used for something better”
This is factually incorrect.
Bitcoin mining because of its economic incentives to use cheap power is, as ERCOT Grid operator Brad Jones attested, a non-rival energy user. In other words, it “powers down when the price of power becomes too expensive; the very time other people want/need that power.
The argument that bitcoin mining takes energy that could be used for something better is analogous to saying “if we took all the crumbs eaten by ants each year, we could feed a small African nation.”
Also, evaluating what is a worthy user of energy is by definition a value judgment. Assessments of Bitcoin’s energy use often fail to not explore the value that bitcoin provides. Without this comparison we cannot legitimately claim that using energy for bitcoin mining could be better served for AI, Netflix or any number of things.
17. “Using flared gas as a power source is not decarbonizing because flaring destroys the methane anyway
This claim is factually incorrect. Flaring does not completely destroy methane. Only 91% is destroyed according to the latest study (source). Because methane is 84x more warming over a 20 year period, this 9% still has a significant environmental impact. Using this flared gas for power generation is a recognised net decarbonising technique, that bitcoin is now providing.
18. “You could use landfill gas for things other than Bitcoin mining such as eV charging stations”
This comment misses the point. The question is not “could landfill gas be used for other things” but “can it be used for other things profitably”.
In most cases, running power generation projects is the only economically viable solution. Also, it is not true that bitcoin mining is only a solution for small landfills. It is a solution for any landfill where it is economically nonviable to sell power back to the grid (which is roughly half of the world’s landfills, because many grids would need major substation upgrades to take MWs of power back.
This is backed up by one of the world’s LFGTE and LFG Carbon Credits experts, Nuno Barbosa, CEO of Unicarbo who stated
““50% of the world’s landfills have no option to sell their power back to the grid, but if they had an onsite customer, that would change everything.”
Barbosa also stated.
“Landfill gas infrastructure is capital intensive so eV charging stations, while a nice idea, seldom pencil out economically, certainly not as an initial user of landfill gas.”
Nuno Barbosa – CEO, Unicarbo
19. “Bitcoin miners earned $31Million by not mining Bitcoin, and this cost is passed on to consumers”
Reality: The evidence from Grid operators, and peer reviewed literature both suggests that Bitcoin may lower costs for everyday consumers of electricity.
Numbers such as the ‘$31M paid to a bitcoin mining company” are often presented without context so as to shock readers. For context: these payouts are done as part of a demand response program. Demand response is a critical example of “ancillary services” which grid operators use to stabilise grids and has existed well before bitcoin mining did (Demand Response has been around for roughly 40 years).
The entry into the demand response market of bitcoin mining means more suppliers compete for grid owners’ demand response revenue, making the demand response market more competitive and lowering the total cost of demand response payments made by grid owners. Brad Jones, former interim CEO of Texas’ grid confirmed “the capability [of bitcoin] to meet our ancillary services means lower cost for all consumers in the state of Texas.” Jones also confirmed that crypto-mining helps “balance our grid more efficiently” because of their unique ability to shut off quickly, when compared to other industrial loads offering demand response services.
One last important nuance, the $31M figure was mis-reported. RIOT earned $7M for selling demand response revenue. The remaining $24 was earned by selling back energy that RIOT had already purchased, not by “not mining Bitcoin”, as some news articles suggested.