This story is part of BREAKER’s week-long celebration of the bitcoin white paper, which is ten years old this week.

It was in 2013, midway through the decade inaugurated by Satoshi Nakamoto’s paper, that Guy Lane began to suspect that bitcoin was eating the world. Lane, an environmental scientist in Australia, came upon two articles that described how the energy required to mine bitcoin was ballooning. The unit for measuring electrical power is the kilowatt-hour (kWh), the energy used by a 1000-watt appliance running for an hour. The data Lane found suggested that bitcoin’s energy consumption was increasing by 70,000 kWh every week. “That’s equivalent of a medium-size town,” he says. “As a carbon auditor, I was intrigued. I wanted to understand what was going on.”

Based on the assumption that a bitcoin miner’s primary expense was electricity, Lane ran some numbers. The greater the value of a bitcoin, the more electricity miners were willing to consume. Bitcoin’s price had hovered at about $100 through 2013, before spiking to nearly $1000 by the end of the year. By Lane’s reckoning, that value meant 8.2 million metric tons of carbon dioxide emissions every year, the same amount as a small nation. If the price were to rise to $10,000, the carbon output would increase tenfold.

Lane was probably the first person to attempt a serious assessment of how the energy consumed by bitcoin was linked with its price. “His numbers have mostly held up over the years,” says John Quiggin, an economist at the University of Queensland in Brisbane, Australia.

His initial calculations may have skewed too high. A study by scientists at the University of Hawaiʻi at Mānoa, published in late-October, estimates bitcoin’s annual carbon output is 69 million tons, an amount the researchers claim could, on its own, lead to catastrophic global warming by 2033. Although this alarming conclusion has already received significant pushback from other scientists, it arrives in the wake of a report released by the United Nations last month that warns the catastrophic effects of climate change could arrive by 2040.

This year, bitcoin will probably burn through between 55 billion and 73 billion kWh, or about the same amount as the entire country of Switzerland.

Even if bitcoin’s carbon footprint is difficult to calculate with precision, it is undeniable that mining requires a lot of energy. A lot. This year, bitcoin will probably burn through between 55 billion and 73 billion kWh, or about the same amount as the entire country of Switzerland. This energy guzzling is built into bitcoin’s proof-of-work model. Miners race to be the first to solve complex processing-heavy math problems. The victor adds a block to the blockchain, and is rewarded with bitcoin. Because the system is designed to ensure that this happens only every ten minutes, as the mining tools have gotten faster and more powerful, the complexity of these math problems has increased—which in turn increases the power required to solve them. The rise in value of bitcoin creates an incentive to make the necessary energy investment.

Harald Vranken, a computer scientist at the Open University of the Netherlands, sees it as a classic arms race. “If people use better hardware, they can compute a block faster, but to keep it at one block every ten minutes, the [math] problems become more difficult,” he says. “So people throw in more hardware. In the end, nobody really gains. You don’t get more bitcoin, you just have to do more work for it.”

Today’s miners process 50 quintillion calculations per second, ten times the rate in 2017. (The size of that number is almost impossible to fathom, but consider this: it’s larger than the number of seconds that have elapsed since the Big Bang.) The specially designed hardware that miners use has grown more efficient, somewhat mitigating the ravenous appetite for electricity. According to Vranken, if miners still used the normal CPUs that were commonplace in the early days of bitcoin, their calculations would require the entire electricity output of planet Earth. He believes some of the higher estimates of bitcoin’s power consumption do not adequately consider this increasing efficiency.

The changing nature of electricity generation is another complicating factor. One common critique of alarmist calculations of bitcoin’s power requirements is that the amount of electricity used is less important than how it is created. Although more than half of all bitcoin mining originates in China, likely using energy from coal-burning plants, regions where renewable energy has lowered electricity costs—such as in the Pacific Northwest, where hydropower is commonplace—are enticing miners to relocate their rigs.

“That’s a common misconception about the electricity industry,” says Catherine Wolfram, a business administration professor at U.C. Berkeley. “The electrical grid that covers the western United States is essentially a big bathtub. If you’re [a miner] located there, you’re using the coal plants in Utah as well as the hydro plants in eastern Washington—even if you’re located next to a hydro plant in eastern Washington.”

Wolfram believes an often neglected factor in discussions about bitcoin’s power consumption is the inflexible nature of the utilities market. “Because it’s regulated, electricity is just not set up to deal with the kind of of onslaught of consumption we’re seeing from bitcoin mining,” she says. “If demand for iPhones exploded, Apple might raise the price. But utility prices are regulated. The basic laws of supply and demand don’t work in a regulated market like electricity.” Miners benefit from an undervalued commodity, while generating billions of dollars in “external” costs that we all pay for, from the pollution, environmental damage, and the often adverse health effects of generating electricity.

The power surge that Lane noticed five years ago shows no sign of abating.

The power surge that Lane noticed five years ago shows no sign of abating. One bitcoin transaction now requires the same amount of power used by 1.5 typical American households in one day. With the current number of daily transactions now averaging just below 300,000, that’s more than the daily requirement of the state of Montana. By 2020, according to one estimate, mining a single bitcoin could require power equal to what a household uses in six months—and generate 4,000 tons of carbon dioxide.

“Even if not all the electricity is being generated by polluting energy sources, another perspective is that the computations used for bitcoin mining are in fact meaningless,” Vranken says. “Of course they’re required to secure the blockchain, but if you look at it from a distance, the computations are just wasted energy. Should you waste electricity for this kind of work? That’s the question.”

If the answer is no, one antidote would be to eliminate the major waste factor. “It’s the proof-of-work algorithm that’s responsible for the energy consumption, so [bitcoin] could switch to another algorithm, which is what’s happening at Ethereum right now,” Vranken says. “They’re using proof-of-stake, which doesn’t rely on showing you did work. It’s more about probability.”

If the arms race continues, could there come a day when the necessary electricity investment outstrips the value of bitcoin? Lane, who has spent five years watching both variables enact an ongoing hold-my-beer scenario, doesn’t think so. “With oil, the last barrel is never extracted, as the effort needed to get the hard-to-get oil increases to the point where you need a barrel of oil to recover a barrel of oil,” he says. “As for the bitcoin endgame, I think they’ll be mined right up to the last one. Whether that’s apocalyptic for humanity depends on how much [bitcoin] is worth.”

Photo courtesy of Giga Watt Mining.