How bitcoin makes burning fossil fuels more profitable than ever

The road to the village of Dresden, New York, on Seneca Lake’s western shore, is lined with wineries. Neat rows of spindly grape vines cover the rolling hills as far as the eye can see. Deepest of the state’s Finger Lakes, this long and narrow water body warms the region in winter and cools it in summer—making for an ideal microclimate for viticulture, and a popular tourist destination.

Just as the road drops down to the village, an old power plant looms. Behind the chain link fence and “Posted” signs thrum thousands of computer servers, or miners; by September they will number close to 50,000 (Greenidge 2022). This is where a new industry has taken root, alarming climate activists around the country: bitcoin mining.

Greenidge Generation first came online as a coal-fired plant in 1937, but by 2011, like so many other coal plants, it was no longer profitable. The owners shut it down, presumably for good, and declared bankruptcy. But a few years later, a private investment firm acquired the plant and converted it to gas, only to find selling energy to the grid still wasn’t very profitable (Christensen 2019) Since coming back online in 2017, Greenidge has sent very little energy to the grid. In fact, since 2020, instead of powering people’s homes and businesses, most of the energy the plant produces has been used to mine the digital cryptocurrency bitcoin.

In March 2020, the plant was using 14 megawatts of power to mine bitcoin (Kharif 2020); by October 2021, it was up to 45 megawatts (Greenidge 2021a). The company plans to use 85 of the plant’s 106-megawatt capacity to mine bitcoins by the end of the year.

It is now well understood that bitcoin mining uses vast amounts of energy. Alex de Vries, one of the leading experts on bitcoin and energy, estimates that the network uses 204.50 terawatt hours annually, about as much as Thailand. A more conservative estimate by the Cambridge Center for Alternative Finance puts it closer to 115.4 terawatt hours, or more than the Netherlands.

Bitcoin’s electricity guzzling has increased dramatically in the past decade. In 2021, de Vries said energy usage had roughly doubled since 2017 to between 78 terawatt hours and 101 terawatt hours (Bambrough 2021), and his current estimate means that the network’s energy usage doubled again in little over a year.

Climate activists see Greenidge as a canary in a coal plant; if one retired fossil fuel plant can be brought back to life to mine bitcoin, what is to prevent the same thing from happening at other retired or idle plants? Other companies have already made moves to follow Greenidge’s lead—including mining companies making moves to buy power plants, and energy companies looking to get into bitcoin.

Bitcoin is giving new life to fossil fuels in even more insidious ways. In some places, miners are burning the dirtiest of dirty fuels—waste coal—to mine bitcoin, and they don’t only have the state’s blessing: They’re getting subsidies for it. Some of the biggest companies have started using the gas that escapes during the oil drilling process to mine bitcoin, claiming it mitigates their climate impact. Even more alarming, some companies are trucking in mobile generators and miners directly to stranded gas wells and burning fossil fuels that otherwise would have stayed in the ground to power miners on site.

If the first decade of bitcoin’s existence was characterized by the mad rush for cheap power, the second looks to be dominated by fossil fuel companies whose assets are quickly depreciating—in terms of both financial and cultural capital—desperate to wring out a bit more profit while they still can.

A gold rush

For over a decade, bitcoin miners have been chasing cheap energy to turn into cold, hard digital currency. When the network first launched in 2009, virtually anyone with a personal computer could mine bitcoins (Balan 2021). It required no special equipment nor extraordinarily large amounts of energy. By 2013, that was no longer the case (O’Grady 2013). The first wave of the bitcoin rush had begun.

Washington state became one of the foremost hotspots because of its cheap and ample hydropower. In 2014, bitcoin miners requested a total of 220 megawatts from the public utility district in Chelan County, fully 20 megawatts more than Chelan’s 70,000 residents used in a year (Roberts 2018). And those were just the miners going through the correct channels. “Rogue” miners set up operations in basements, garages, or empty apartments, wreaking havoc on the residential grid, which isn’t designed to handle such heavy loads. In at least one case, a mine operator overloaded a transformer and started a brush fire (Roberts 2018). Demand grew exponentially; nearby Grant County received requests for 2,000 megawatts in 2017, three times the amount of energy needed to power the county in a year (Rane 2021). In 2018, Chelan placed a temporary moratorium on new bitcoin mining operations, as did the public utility district in Franklin County and in Mason County.

A similar drama played out in upstate New York. The small city of Plattsburgh’s low-cost electricity attracted several large mining operations, but the city has only a limited supply of cheap hydropower each month. Once they use that up, the city must buy power from the grid, which is a lot more expensive (and more likely to come from fossil fuels). When that inevitably happened after the miners moved in, individuals and businesses saw their electricity bills soar (Lotemplio 2018). A working paper by Matteo Benetton, an assistant professor at the University of California at Berkeley, estimates that bitcoin mining in New York increased annual energy bills by $165 million for small businesses and $79 million for individuals across the state (Benetton 2021).

In 2018, Plattsburgh also put a moratorium on new cryptocurrency operations, which was only lifted in 2019 after a new regulation was created to put the burden of overage costs on the miners (Delisle 2019). Chelan County’s moratorium was lifted after the utility set higher rates for cryptocurrency operations and made them cover additional capital costs necessary to maintain the grid (McDaniels 2018).

These examples show how bitcoin operations—where the primary expense is electricity—are vulnerable to fluctuations in energy markets, but also to the shifting attitudes of fellow electricity customers or utilities toward bitcoin operations. Once cheap and plentiful electricity sources may not always remain cheap and plentiful, either because of increased demand driven by mining itself, or due to increased regulations targeting cryptocurrency mining.

The next logical step was for bitcoiners to seize the means of electricity production.

“It’s an obvious incentive, if you if you have a dormant power plant that you can turn profitable,” said de Vries. “That’s a perfectly valid incentive to do it. I mean, it’s not desirable from an environmental perspective, but it is certainly a valid incentive to just do it.”

The second wave

Greenidge Generation bills itself as the “first publicly traded, vertically integrated bitcoin mining company,” meaning that they can generate their own electricity.

Other companies have already made moves to follow suit. In March 2021, the bitcoin mining company Digihost announced a deal to purchase a 60-megawatt gas-fired power plant in North Tonawanda, New York, near Buffalo, which could power approximately 17,000 new miners. Digihost has secured permits from the city, and as of May, was only waiting for approval from the New York Public Service Commission to finalize the deal.

In 2020, the North Tonawanda plant generated a total of 19,875 megawatt hours of electricity, a slight increase from the 17,791 megawatt hours generated in 2019. That’s less than 20 percent of its capacity. In 2020, Greenidge generated 246,926 megawatt hours, up from just 55,675 in 2019. The big difference? Greenidge had started mining bitcoin. Increased power generation resulted in a corresponding jump in emissions at Greenidge, from 37,627 tons of carbon dioxide in 2019 to 246,060 tons of carbon dioxide in 2020. If Digihost secures the go-ahead from the New York Public Service Commission, it only follows that the emissions at North Tonawanda will similarly soar.

This isn’t only happening in New York. In 2021, Stronghold Digital Mining acquired two power plants in Pennsylvania, and intend to buy a third. The first, Scrubgrass, in Venango County, Pennsylvania, has a capacity of 85 megawatts; the second, Panther Creek, has a capacity of 80 megawatts. Scrubgrass generated a mere 5,893 megawatt hours in 2020, and Panther Creek generated 60,143 megawatt hours. Even with that relatively low generation, Scrubgrass emitted 16,359 tons of carbon dioxide and Panther Creek emitted 84,444 tons of carbon dioxide. That’s roughly 2.77 and 1.4 times more carbon dioxide, respectively, emitted per megawatt-hour of electricity than Greenidge Generation in 2020. Why so high? Because the two facilities burn waste coal, which has even greater greenhouse gas emissions than gas.

These are examples of bitcoin miners that have acquired power plants, but the reverse may occur as well. In April of 2021, the largest utility in Missouri installed a data center to mine bitcoin at the coal-fired Sioux Energy Center in West Alton, just north of St. Louis, but kept it a secret until the fall, when the utility tried to pass on $8,000 in associated energy costs to its customers, revealing the scheme (Gray 2021, Moss 2021). The utility’s rationale is that the extra demand from the miners can help justify running the coal plant during periods of low energy demand, although for now the half megawatt the miners require is a drop in the bucket of the plant’s 972-megawatt capacity. Still, the company sounds eager to significantly scale up the project. Ameren’s head of research and development said he could see installing bitcoin miners that use as much as 20 to 80 megawatts per site (Tomich 2021).

Nor is Ameren likely to be the last energy company to pursue bitcoin mining to pad out their coffers. In May, Rainbow Energy Marketing Corp. purchased the largest power plant in North Dakota—the 1,151-megawatt Coal Creek Station. The Washington Post reported that company officials told local officials of their plans to mine bitcoin; although Rainbow didn’t admit to the bitcoin mining plans to the Post, the company did say they were planning on building data centers (Partlow 2022).

Bitcoin mining could be the only thing to make the plant profitable; it lost $170 million in 2019. It was expected to close in 2022. If indeed Rainbow does begin bitcoin mining there in order to keep the struggling coal plant (and the adjacent open-pit coal mine that feeds its furnaces) afloat, the impact on the country’s greenhouse gas emissions would dwarf that of Greenidge Generation’s.

Twisted environmental incentives

When I first wrote about Greenidge Generation in early May 2021, the company was describing itself as a “clean” and “environmentally-sound” plant with a “unique commitment to environmental stewardship” (McKenzie 2021). A week later, Elon Musk said that Tesla would no longer accept bitcoin over concerns about its climate impact, and two days later Greenidge announced that it would become “carbon neutral” by purchasing voluntary carbon offsets (Greenidge 2021b). The effectiveness, or lack thereof, of carbon markets to address the problem of greenhouse gas emissions is beyond the scope of this article, but bitcoin miners are eager to portray themselves as green.

Perhaps the most egregious example comes from Stronghold Digital Mining, which claims to be “the environmentally beneficial” bitcoin miner, because they mine using waste coal. But critics say that it’s a bait-and-switch, taking one very real environmental problem and replacing it with another.

Waste coal, also known as slag, culm, boney, or gob, are the undesirable leftovers that can’t be burned in traditional coal plants. At the turn of the 20th century, it was sorted out by young boys, and later by machines, and then deposited in massive heaps across coal country. These refuse piles are toxic, barren of plant-life, and prone to leaching toxic substances into groundwater. Sometimes they catch on fire, emitting noxious fumes and greenhouse gases. They are an environmental catastrophe.

Stronghold claims its operations removes 90 percent of nitrous oxide emissions, 99.9 percent of particulate emissions, 99.9 percent of mercury emissions, and are “capable” of removing 98 percent of sulfur dioxide emissions. But if you look back at the emissions at these plants in 2020—before bitcoin mining started at these sites—Scrubgrass emitted 335.5 tons of sulfur dioxide and 11.99 tons of nitrogen oxide in 2020 and Panther emitted 39.58 tons of sulfur dioxide and 35.01 tons of nitrogen oxide. Both nitrogen oxide and sulfur dioxide pollution have negative effects on respiratory health, both interact with other chemicals in the air that can result in particulate matter pollution, and both contribute to acid rain.

“In our view, that’s not getting rid of the pollution, that’s moving the pollution,” said Rob Altenburg, the director of the PennFuture Energy Center. “You’re moving it from the land to the air.”

But because Pennsylvania considers burning waste coal to be a form of environmental remediation, Stonghold is reaping rewards for polluting the air. Pennsylvania classifies energy from coal waste as a Tier II alternative energy resource, the equivalent of large-scale hydropower. Stronghold receives both Coal Refuse Energy and Reclamation Tax Credits and Pennsylvania Tier II Alternative Credits.

Without these incentives, according to documents Stronghold filed with the Securities and Exchange Commission, it would cost $37 to generate a megawatt hour of electricity; with them, it costs only $18.

Burning at the source

There are even more efficient ways to turn fossil fuels into bitcoin, as pilot projects from ExxonMobile and ConocoPhillips demonstrate.

In January 2021, ExxonMobile began a pilot project with Crusoe Energy Systems Inc. to use gas produced at oil wells in the Bakken shale basin in North Dakota to power generators to mine bitcoin, Bloomberg reported (Malik 2022). The pilot project was expanded last July and the company is considering expanding the project to other drilling sites in Alaska, Nigeria, Guyana, and Germany, anonymous company insiders told Bloomberg.

In February, ConocoPhillips confirmed to CNBC that the company is also selling gas produced at oil wells in the Bakken to third-party companies that use it to power generators and mine bitcoin instead of flaring it (Sigalos 2022). The project dovetails with the company’s goals to reduce flaring, or the burning of excess gas at oil wells that can’t be transported as fuel. If the gas is sold to a third-party, it will reduce ConocoPhillips’ direct emissions, although the net greenhouse gas emissions (technically referred to as Scope Three emissions by organizations such as the EPA) won’t significantly change.

“You could effectively see it as a subsidy for fossil fuel extraction, because you’re turning a byproduct that would otherwise not be profitable, profitable,” said de Vries.

It’s a win-win for the oil company—which gets to sell what was previously a waste product—and the bitcoin miner—who gets a cheap source of energy. But less so for the environment.

At least in these cases, greenhouse gases would be emitted one way or another, whether by flaring or by burning to power a generator. That isn’t always the case.

As previously reported in the Bulletin, in places with a lot of fracked gas wells—like Pennsylvania, Kentucky, and Alberta, Canada—people are installing generators at stranded wells and burning fossil fuels that otherwise would have stayed in the ground to mine bitcoin with the electricity produced by mobile generators (McKenzie 2022). This can be more profitable than selling gas at wholesale rates to utilities.

Going against the narrative

The bitcoin mining industry desperately wants to portray mining as environmentally friendly. These defenders will point to a few outliers, like Marathon Digital Holdings, which announced in April its plans to move mining operations from Montana, where they own and operate a coal plant to mine bitcoin, to Texas, pledging to become “100% carbon neutral by the end of 2022”—although like Greenidge, they could claim that by purchasing carbon offsets (Gkritsi 2022). Bitcoin miners are largely treating concerns about their climate impact as more of an image problem than a genuine problem.

But as long as burning fossil fuels to mine bitcoin is economical, people and companies will continue to do it, even if some miners seek out low-carbon energy sources. And as long as mining bitcoin remains profitable, fossil fuel companies will increasingly try to use it to prop up their dying industry.

Greenidge Generation is in many ways an outlier for receiving so much attention. This happened for two reasons: First, because there is a robust environmental activist community in the Finger Lakes who previously banded together to defeat a planned liquefied petroleum gas project slated for the region. And second, because New York State has some of the more ambitious climate targets in the nation, giving activists a law to cite when they point out the hazards of increasing fossil fuel use.

Meanwhile, states like Kentucky are actively encouraging the burning of fossil fuels to mine bitcoin, and Pennsylvania and others may follow suit. Like so much else in this country, the debate is skewing along sharply ideological lines, with the environment on one side and the blind pursuit of profit on the other.