Bitcoin Sunset Mining: Has Its Time Come?
A discussion about putting a “sunset clause” to using stranded fossil gas to mine bitcoin.
Powering bitcoin mining with stranded fossil gas from abandoned oil and gas wells has a less negative impact on climate and the environment than simply flaring or venting that gas. But being less negative still doesn’t make this type of mining net positive from a climate and environmental point of view, as I have argued before.
Climate, the environment and human health still suffer the increasingly negative cumulative impact of CO2, methane and other harmful emissions. From a purely energetic perspective, the toolkit of solutions to replace fossil gas already provides a few efficient options. More options will come online in terms of flexible grids, packetization of electric energy, grid interconnections, high-voltage lines to transport power over long distances, batteries, etc.
The all-round net-positive solution right now would be using the revenues generated by bitcoin mining with stranded gas to pay for the closure of the same wells powering the mining and the environmental remediation of lands polluted by those wells, while still contributing to the bitcoin network.
As soon as bitcoin mining has produced enough revenues to pay for wells’ closures, land remediation and a margin for miners, it’s more net positive and effective overall to close stranded wells, remediate lands and move those mining rigs to other stranded wells for a new round of closures.
There is no counter indication to this net-positive strategy in terms of bitcoin’s resiliency. If we choose the lesser-negative-but-still-net-negative climate and environmental strategy, it’s purely for financial, Cantillon-based, seigniorage induced greedy motives unrelated to bitcoin.
As a shorthand, for lack of a better phrase, I will call this net-positive strategy “sunset mining” — mining with a source of energy that will fade and end when certain conditions are met, i.e., when a climate net-negative energy source has produced enough revenues to pay for its own site closure, land remediation and a margin for miners.
Ideally, sunset mining would ultimately wind down flared gas mining with stranded wells.
Several questions revolve around this possible course of action, especially figuring out costs of well closure and land remediation vs stranded gas mining revenues and profits. Do the numbers work? What price does bitcoin need to be at to make it work? How long would it take to generate enough revenues to close wells? Would it have to be a non-profit type of setup? Can it work as a commercial endeavour?
In the following sections, I will try and provisionally answer some of these questions, quantifying what sunset mining would mean financially and which limitations it has in the overall scheme of things.
The (Uncertain Yet Large) Addressable Market for Sunset Mining
The precise number of stranded wells in the United States is not certain, but there are estimates.
The number of abandoned oil and gas wells in the U.S. is about 3.7 million, with around 3 million abandoned oil wells and 0.7 million abandoned gas wells, according to the latest data by the U.S. Environmental Protection Agency (EPA), published in April 2022 and covering the 1990–2020 period.
Historically, after 150 years of oil and gas production, there are 10 million abandoned, orphaned, idled, marginal and legacy wells in the United States, not including State or federal waters, according to a 2016 report by the National Energy Technology Laboratory (NETL) of the U.S. the Department of Energy.
There are more than 900,000 active oil and gas wells in the United States, and more than 130,000 have been drilled since 2010 during the U.S. shale boom, according to a 2017 report by the Washington Post.
So, whatever the exact number, historical and more recent stranded oil and gas wells are likely in the millions in the U.S. and when they were abandoned, “many wells were not adequately sealed and now provide a potential conduit for the vertical movement of liquids and gases”, the NETL says.
For this discussion, I’ll stick with the latest EPA numbers: 3.7 million stranded wells.
Cost of Capping Stranded Wells
Just as it’s hard to identify the exact number of stranded wells, it’s also tricky to assign an exact price tag to the closure of such wells.
Based on Texas orphan well data, the authors of a 2016 paper published by the Society for Petroleum Engineers (SPE), projected an average cost of $33,000 per well to plug producing wells. Other studies, quoted by a Carbon Tracker 2020 report titled It’s Closing Time: The Huge Bill to Abandon Oilfields Comes Early, cite a Diversified Gas and Oil estimate of $28,400 per well and a Rystad Energy figure ranging from $20,000 to $40,000 per well.
These average estimates are on the low to medium side of the range estimated by the Environmental Defense Fund, according to whom “closure costs average $25,000-$75,000 per well” in the U.S.
But this may still be a low estimate. Last year, the Colorado Oil and Gas Conservation Commission (COGCC) calculated the average plugging and clean-up cost at nearly $93,000 per well in its State. An oil & gas industry’s rule of thumb cited by the Carbon Tracker report to estimate plugging costs is to budget 5% to 10% of drilling costs, noting that plugging costs, like drilling costs, increase with the depth of the well. As the U.S. Energy Information Administration (EIA) estimates the average shale drilling costs range from $1.8 to $2.6 million per well, the industry forecast of average plugging and abandoning costs for shale wells in the Texas area considered should range from $90,000 (5% of $1.8 million) to $260,000 (10% of $2.6 million).
But “industry cost data from outside the U.S. (there is no available U.S. industry cost data) indicates that actual costs may be an order of magnitude higher on average. The costs to retire thousands of ultra-deep wells and those with wellbore damage or close to sensitive receptors can exceed $1 million” each, according to the Carbon Tracker study.
One possible reason for these discrepancies is that lower estimates tend to refer to older, shallower wells.
Shale wells, on the other hand, are deeper, making them more costly both to drill and to plug. Most orphan wells are straight-hole vertical wells that far predate the shale boom of the mid-2000s, but modern shale wells in the Western U.S. are deep, averaging 10,000 feet, and were excavated using modern unconventional drilling and completion techniques such as horizontal drilling and hydraulic fracturing (fracking), the Carbon Tracker report’s authors explain.
“Significant differences in vertical depth and well design mean that average plugging and abandoning (P&A) costs for orphan wells in Texas do not imply an average P&A cost for shale wells”, according to the Carbon Tracker study.
Looking at Australia’s experience, the cheapest well cost $25,515 to plug. The most expensive was $675,604. The average cost for wells more than 10,000 feet was $424,000, adjusted for inflation and currency conversion.
“These costs exclude charges for site preparation and rig mobilization, which are unnecessary in P&A completions where the drilling rig remains on site, but which are incurred when closing post-production wells. Notably, the one outlier in the Australia data (a 9,000-foot well costing $676,000) shows that outlier deep wells can significantly exceed expected average costs”, according to Carbon Tracker.
It’s safe to say that estimates for plugging and remediating oil & gas wells fluctuate a lot, depending on several variables. Without any pretense of exhaustiveness and detailed precision, merely with the intent to provide a rough picture, I will take a more or less arbitrary low, medium and high plugging cost, trying to profile what the economics of sunset mining might look like.
Cost Bases for Sunset Mining Economic Estimates
Rounding numbers, I will exclude the lowest and highest figures I found ($20,000 and $676,000). I will then take two ranges I mentioned for the U.S., from lower to higher, and use the average of each to establish a floor figure and a middle figure for plugging and remediating a well. The ceiling amount will come from the Australian example. Summarizing:
· $30,000 — minimum price tag (average of Rystad’s $20,000 to $40,000 range).
· $175,000 — middle price tag (average of industry’s rule of thumb’s $90,000 to $260,000 range).
· $424,000 — maximum price tag (average of Australia’s price study).
This is a somewhat arbitrary approach, but it should provide a more realistic picture, based on the newer reality of horizontal, deep drilling, whose plugging costs could reach in extreme cases as much as $1 million per well, according to Carbon Tracker.
To start establishing some preliminary economics of sunset mining I will use a handy online tool created by Great American Mining (GAM) and called “Gas-to-Ash Calculator”. This allows to estimate the value of bitcoin mining based on a certain availability and characteristics of stranded gas, and the type of miner used.
As GAM indicates, its estimates are calculated in real-time. That means that bitcoin network difficulty and bitcoin value in U.S. dollars are set at the time of calculations. Such values will obviously change over time. Unfortunately, GAM’s tool doesn’t allow to adjust bitcoin’s USD exchange rate and difficulty levels, so it can’t be used directly to play out different scenarios, but it still is a very useful base, from which one can try and extrapolate different scenarios.
For the calculation I’m reporting here, the BTC-USD exchange rate was $29,771.80. The price of bitcoin could fall even further, for instance around $22,000-$24,000, toward its 200-week simple moving average, which would make it even harder for sunset mining to reach the breakeven point needed to fund stranded wells’ closure, But a ∼$30K average price — a more than 50% discount compared to its all-time high — should provide a somewhat prudent and sound base to assess the feasibility of sunset mining in the medium-long-term for these gas-to-ash estimates.
GAM’s tool also assumes a “2% mining pool fee, and 100% generator load (with 2.5% of total available capacity allocated to other operational requirements)”. In a real-life situation, there would also be other factors to consider, like taxes, depreciation of the miners, possible downtime due to repairs and maintenance, etc. I will not consider these factors and focus instead on gaining a rough understanding of basic conditions.
Based on 3.7 million stranded wells to plug and remediate and the floor and ceiling prices I took as benchmarks, total costs could range anywhere from $111 billion to $1,56 trillion, with the actual figure being probably somewhere in between.
Gas Volume Bases for Sunset Mining Economic Estimates
The U.S. Department of Energy (DOE) has analyzed the volume of flare activities across major producing regions, such as the Bakken play in North Dakota and the Eagle Ford and Permian Basin plays in Texas. The DOE’s identified more than 78,000 individual flare units as of 2018. These flare units range in size from less than 100 thousand cubic feet per day (Mcf/d) to up to 0.1 billion cubic feet per day (Bcf/d). Flare units less than 100 Mcf/d represented 89% of the total number of flare units and more than 26% of the total flare gas volume in these selected producing plays.
“Achieving a 55% reduction in annual flare gas volumes would require operational modifications to greater than 75,000 individual flare units (96% of total number of flare units). If larger flare units were targeted to decrease flare volumes, for example, a 45% flare gas reduction could be achieved through operational modifications to approximately 3,000 flare units (4% of total number of flare units) with flare sizes greater than 300 Mcf/d”, according to the DOE’s Flaring and Venting Reduction Research and Development Activities report.
The following two illustrations from the DOE’s report show numbers, volumes and percentages of different flaring activities in the sample, which can be considered representative of the wider flaring context in the U.S.
As one can see, smaller flares are responsible for larger overall volumes of gas combustion, while larger flares make up a minority of the gas being burned. And as the DOE has noted, whatever use one can imagine for that gas would require tens of thousands of single actions — a very piecemeal type of work.
Possible Economics for Sunset Mining
Payback times for closing and remediating stranded wells could range from less than 2 years to almost 4 years, with total costs fluctuating between less than $0.5 million to $9.5 million per well and revenues ranging between $260,000 and $5 million a year per well, depending on the closure and remediation cost estimates and the gas volumes one considers. These figures are based on some number crunching I did using GAM’s tool, with the following assumptions:
· Bitcoin-USD Exchange = $29,771.80
· Miner: Whatsminer M30S (88TH/s @ 3,344w)
· Bitcoin Network Difficulty: 29,897,409,688,833
· BTUs contained in a standard cubic foot (SCF) of Natural Gas: 1,100
· Gas volumes: between 50 and 950 Mcf/d.
In these estimates, I factored in mining and cooling investment costs (CAPEX), a rough approximation for site preparation costs and a 20% profit margin for miners. I don’t know if this margin is realistic in the long term, but it seemed prudent in the short to medium term. Profit margins would also vary depending on the for-profit vs non-profit model one could adopt. Operating expenses would consist mainly of energy costs, which in this case are zero or close to zero. Mcf/d levels have been chosen to reflect actual gas volumes coming from stranded wells, even if volumes could be higher if gas came from pipeline bulk sourcing. Here are the exact estimates for the three price tags I chose as benchmarks to close and remediate each well.
Events that could have an impact on these estimates are mainly variations in the USD price of bitcoin, which would indirectly affect network difficulty adjustments, and miners’ profit margins. A 50% profit margin, for instance, would lengthen total payback times to a maximum of more than 5 years. A doubling of bitcoin’s price closer to its all-time high would double revenues and cut in half payback times. A further decline in bitcoin’s price would slow down sunset mining’s payback times.
Based on my present assumptions, payback times seem reasonable, both in economic terms, especially for a non-profit model run by either private or public organizations, and in climate terms, being fairly quick. One element that seems daunting is the extreme fragmentation of stranded wells, which would require thousands of separate mining setups. It remains to be seen how much hashing power can be deployed at any given time for sunset mining, which means it could still take some years to plug stranded wells with this type of action.
Let’s try and see how many years.
Taking the latest estimates by the Cambridge Bitcoin Electricity Consumption Index (CBECI), for instance, we see that the U.S. hosts roughly 38% of the world’s hashpower, whose global electric power is estimated at 13.55 GW. That means that in the U.S. the electric power dedicated to bitcoin mining should be roughly 5.15 GW or 5,150 MW. As an example, let’s say that no more than 5% of mining in the U.S. will consist of sunset mining. That would be 257.5 MW.
The volume of U.S. natural gas that was reported as either vented or flared in 2019 was 1.48 billion cubic feet per day (Bcf/d) or 540 Bcf per year, according to the EIA. The same report estimates flaring in Texas averaged between 0.80 and 0.90 Bcf/d or 328.5 Bcf per year through the last quarter of 2018 and the last quarter of 2019. The World Bank estimates onshore flaring in the U.S. had decreased to about 292 Bcf in 2021.
Assuming these estimates are ballpark accurate (they are often underestimated), taking a round number closer to the more recent World Bank flaring figure, how long would it take for 257.5 MW of cumulative bitcoin sunset mining electric power in the U.S. to fund the closure of about 300 billion cubic feet of annual gas flaring? Is this a viable scenario? Will there be enough miners earmarked for this type of action?
Assuming a theoretical worst-case scenario, as if all sunset mining consisted of small and fragmented 50 Mcf/d gas sources (0.01825 Bcf per year) requiring small 0.22 MW systems, there would have to be about 1,170 of these setups to reach 257.5 MW of total sunset mining capacity powering 76,050 mining machines. To be economically sustainable, these smaller systems need to run from 2 to 4 years. So, that’s how long it would take for each sunset mining setup to retire one stranded well in the scenario I just described.
A 257.5 MW sunset mining capacity would be able to absorb 21.35 Bcf of stranded gas per year. If one year were enough for sunset mining to generate enough revenues to fund wells’ closure and remediation, the whole 300 Bcf of stranded gas volume could be retired in about 14 years, assuming no new stranded gas was added to the annual total. But because in this worst-case scenario economic viability is reached only after 2 to 4 years, to absorb the whole 300 Bcf of stranded gas it could take 14 x 2 to 4 years = 28 to 56 years.
If, instead of 5% of total mining its sunset variant was able to mobilize only 2.5% of the total power, well-closure funding times would double to 4 to 8 years per well. If sunset mining was able to mobilize 10% of total U.S. hash and electric power, the time would halve to 1 to 2 years per well — with the total time to retire 300 Bcf of overall gas volume at roughly 14 to 28 years (it then takes from a few days to a week to plug a well, with several more weeks or months necessary to bring back the site to its original state).
These estimates seem to show that sunset mining could be well worth the effort, both in terms of 2050 climate targets and environmental remediation, while contributing to bitcoin’s resiliency and security — especially if closure costs remain contained around the minimum or middle price tags.
Another positive factor is that the extreme fragmentation of stranded wells could help employ thousands of former oil & gas workers who have lost their jobs in the latest booms and busts of the fossil fuel sector.
Carbon Pricing Contribution
It’s worth noting that the estimates I just made rely solely on sunset mining’s “internal” revenue and cost dynamics, without taking into account external factors that could change its rates of return and speed of implementation.
One such key factor is the extra revenue that some kind of carbon pricing or carbon certificate trading scheme could provide.
Unlike the European Union or China, the United States doesn’t have a federal system to price carbon or trade carbon certificates for the most polluting industries. Carbon pricing in the U.S. at present is either voluntary or takes place at State level, and only a handful of States have some form of carbon scheme — most notably California, with newer or more limited implementations in Massachusetts, Oregon, Pennsylvania and Washington.
Carbon prices in California are now about $30 per tonne of CO2 equivalent ($30/tCO2e), while in Massachusetts, for instance, they are only $1/tCO2e. That compares with $88/tCO2e in the European Union and $9/tCO2e in China.
The Report of the High-Level Commission on Carbon Prices and Competitiveness, a voluntary, multi-stakeholder partnership hosted by the World Bank, identified a $50–100/tCO2e range as the price needed by 2030 to keep global heating to below 2°C — the upper end of the limit agreed in the Paris Agreement — as part of a comprehensive climate policy package.
Research and rating company S&P Global expects the EU’s carbon allowance prices to exceed $100/tCO2e from 2025 onward, topping $120/tCO2e by 2030 as the EU steps up its transition to net zero emissions. The overwhelming consensus among analysts and researchers is that carbon prices will have to increase significantly in the next decade or so to reach global decarbonization goals.
Taking again a somewhat arbitrary yet prudent reference point, it might be plausible to assume that within the next few years different U.S. jurisdictions could price carbon certificates as California does today: $30/tCO2e.
Natural gas (methane) generators have an average emission rate of 898 pounds of CO2 per MWh, according to the EPA, equal to 407.3 Kg/MWh or 0.407 tCO2/MWh. Assuming for simplicity’s sake that a 0.22 MW sunset mining system works 24/7/365 (it won’t due to maintenance, repairs, etc. but close enough to it), it will emit roughly 784 tCO2 per year.
As the 100-year global warming potential of methane is 25 times higher than CO2, 1 kg of methane equals 25 kg of CO2e. Therefore, avoiding the release of methane into the atmosphere (as methane is turned into CO2 in the electric generation process) means decreasing emissions by about 18,816 tons of CO2 equivalent per year. This carbon offset, valued at $30/tCO2e, would be worth $564,480 a year, for each 0.22 MW sunset mining system. This calculation would produce even bigger financial results if instead of the 100-year average global warming potential we shortened the timeframe to a more human scale, as over a 20-year period methane is 86 times more potent than CO2 in terms of its warming effects. That means that avoiding emissions of 1 kg of methane equals preventing 86 kg of CO2e over 20 years.
Miners will obviously want to poket this revenue to increase their profits, but they could decide to earmark some of it to help pay for well closure and land remediation. It’s impossible to say what percentage of carbon credits would be used to fund these expenses. A non-for-profit operator might decide to use all the carbon credit revenue for well closure and remediation. A commercial operator might decide to use less. But if a commercial operator is into sunset bitcoin mining to begin with, it’s likely it will be sensitive to environmental and climate issues. So, let’s take another arbitrary, yet plausible, benchmark and say that on average sunset bitcoin mining operators will dedicate 50% of carbon credits’ revenue to fund well closure and remediations.
Sticking to the more diluted 100-year horizon for global warming effects, even in a theoretical worst-case scenario (all sunset mining made of small and fragmented 50 Mcf/d gas sources requiring small 0.22 MW systems), with the additional revenues potentially created by a $30/tCO2e carbon scheme just in the 1st year, payback times for closure and remediation costs of $30,000, $175,000 and $424,000 would shorten, respectively, to 13 months, 16 months and 22 months — basically from just over one year to less than 2 years.
As carbon credits would continue being generated even after the 1st year, up until mining and carbon credit revenues break even with closure and remediation costs, the “net” payback time would be even shorter, i.e., presumably around 1 year for all cost scenarios. If this is true, 257.5 MW of sunset mining capacity would be able to retire the whole 300 Bcf of stranded gas volume in about 14 years, assuming no new stranded gas was added to the annual total.
Sunset Mining Limits
Even counting the positive contribution that sunset mining could give to bitcoin, climate and the environment, its scope is limited by the fact that flared and voluntarily vented methane from sources that are easy to tap make only a small fraction of the total volume of gas leakage.
Methane emissions originating from abandoned wells in the U.S. amounted to only 3% of the total oil & gas industry emissions in 2020, according to the country’s Environmental Protection Agency. Based on a different industry breakdown used by the Intergovernmental Panel on Climate Change (IPCC), abandoned oil & gas wells were responsible for 6.9 million metric tons (MMT) of CO2 equivalent out of total emissions of 4,854.7 MMT CO2 Eq from the U.S. energy sector and total emissions from all human-related U.S. sources of 5,981.4 MMT CO2 Eq in 2020, according to another EPA report.
That means methane emissions from abandoned oil & gas wells make up as little as 0.14% of the energy sector’s greenhouse gasses and 0.11% of total human-made emissions, as this EPA graph of emissions by category shows, with methane emissions by abandoned oil & gas wells indicated by the red arrow.
Stranded gas flaring in the U.S. was responsible for 0.2% of total gross U.S. greenhouse gas emissions in 2020 (13 out of 5,981.4 MMT CO2 Eq), according to the EPA.
Even if these figures were grossly underestimated, the great majority of emissions come either from final uses and downstream sources or upstream and midstream sources that can’t be readily tapped — things like leaks in oil & gas drilling and extraction activities and infrastructure, in pneumatic controllers, offshore oil platforms, equipment, gas engines, produced water, chemical injection pumps, etc. It’s all dribs and drabs that are not easily usable to power bitcoin mining.
Sunset mining can contribute to plugging and remediating whatever gas volumes we are able to harness effectively and economically, but it can’t fix the bulk of the fugitive emission problem. It might be able to punch above its weight if mining revenues were pulled and used to finance the closure of other types of more widespread leaks. Oil & gas companies might be mandated to or voluntarily use bitcoin mining fueled by larger gas flows to fund well plugging and abandonment costs that are now insufficiently secured by bonds or other collateral.
“If all methane leaks from fossil fuel operations in 2021 had been captured and sold, then natural gas markets would have been supplied with an additional 180 billion cubic meters of natural gas. That is equivalent to all the gas used in Europe’s power sector and more than enough to ease today’s market tightness”, according to the International Energy Agency (IEA).
Expressed in a different unit measure, the IEA estimates that “the global energy sector was responsible for around 135 million tons of methane emitted into the atmosphere in 2021”.
There’s clearly a lot of methane leakage to be plugged all over the world. But fixing all the leaks and emissions of methane, CO2 and other greenhouse gases upstream, midstream and downstream along their supply and consumption chains would be very hard and costly. There are countless leaking points of emissions. They go hand in hand with the technology infrastructure. Emissions are usually underestimated and the idea that the oil & gas sector will suddenly start doing much more than in the past to fix the problem seems hopelessly naïve. Whatever effort is made risks being not much more than putting a Band-Aid on the climate problem.
The bulk of the solution is to break our fossil gas and oil addiction, build as much renewable energy capacity as possible, electrify as many final uses as possible as quickly as possible, strengthen smart flexible grids as rapidly as possible, boost storage capacity, develop long-distance high-voltage lines transporting electrons through interconnected and intercontinental grids, and discourage as much as possible fossil fuel demand with an effective, technology-agnostic carbon tax.
That doesn’t mean sunset mining is useless or irrelevant. On the contrary, I think any decimal point of emissions sunset mining can fix through sound economic and ecological action within a 2050 timeframe should be pursued. It’s the net better option both pragmatically and philosophically — in bitcoin terms, climate terms and environmental terms.
It will help rebalance the relationship between societal development and the planet’s ecosystem, between human beings and the oikos, the “common home” referred to in the etymology of the term “ecology”, of which bitcoin can be a foundation.
This article has been slightly modified a 1st time in relation to sunset bitcoin power and timing estimates. And it has been modified a 2nd time with the addition of the “Carbon Pricing Contribution” section.