The year is 2018. After a rough Greek exit from the eurozone, economic malaise has spread to Italy, Portugal, Spain, and France. Nervous citizens across Europe look for a way to get their money out as currency traders hammer the weakening euro, banks impose withdrawal limits, and their purchasing power plummets.
Compared to the euro, the peer-to-peer decentralized electronic currency has now become a relatively stable digital asset. Fiendish buyers trade their euros en masse online for Bitcoin, and soon, depositors worldwide join them. The price of Bitcoin rises, prompting more user adoption by spenders and speculators, and recognition from governments and populations alike.
The above scenario sounds like a nice piece of prepper-bait from conspiracy site infowars.com. But could (or should) Bitcoin actually take over? Some of the more enthusiastic Bitcoin advocates argue that the currency is ready for prime time—in other words, ready to replace national currencies, or perhaps replace global banking’s creaking clearinghouses. Would this be good for the world?
From an environmental point of view, it certainly wouldn’t be good news. Unfortunately for Bitcoin advocates, the currency uses too much electricity right now—way too much: According to my calculation, a single Bitcoin transaction uses roughly enough electricity to power 1.57 American households for a day.
All that energy expenditure has an important purpose: it secures Bitcoin from attacks by speculators, criminals, and other evil-doers by raising the price of the computer power needed to gain control of all transactions on the network. The computers that make up the Bitcoin economy’s backbone are constantly ensuring security and verifiability for the network by solving cryptographic puzzles. This process is called “mining.” Those who participate in this network maintenance are rewarded in Bitcoin, incentivizing them to bulk up their machines so they can mine more efficiently.
There is potential for Bitcoin to become more efficient by stuffing more transactions into the mining process. But at the end of the day, if Bitcoin sees increased adoption and price and many more useful transactions, power consumption is almost guaranteed to grow.
Motherboard has previously covered how big Bitcoin mining operations can get. So how much electricity are we talking about?
Let’s take this Bitcoin mine in China as an example of the scale of today’s operations. It is supposedly running at 6 PH (quadrillion hashes) per second, according to a Chinese Bitcoin company CEO posting in a Bitcoin forum, with the aim to scale up to 12 PH per second. That would give it about 3.3 percent of the total power on the Bitcoin network. Because the Bitcoin network is set up to dole out around 3,600 BTC per day to miners, this mine would rake in about 118.8 BTC per day, or more than $30,000USD at the time of writing. That’s not a bad haul when your electricity costs are among the lowest in the world at 3 to 6 cents per kw/h, about a third of US prices.
Bitcoin’s power usage per transaction isn’t remotely sustainable as a wholesale replacement for the conventional financial system
Computer cooling firm Allied Control estimates the total power consumption of the Bitcoin network at 250 to 500 Megawatts. Looking at the total hashrate, which is the number of calculations the network can perform per second, and applying a generous miner efficiency of 0.6 watts per gigahash, we can estimate our own back-of-the-envelope Bitcoin network constant power draw at just under 215 MW, although this figure is always in flux (it’s important to note that many of the variables in my calculation are constantly changing slightly). That’s around enough zap to power 173,000 average American households’ daily electricity usage.
With about 110,000 transactions per day, that works out to 1.57 households daily usage of electricity per Bitcoin transaction. Yes, every time you buy something in Bitcoin, you could be using as much electricity as 1.57 American families do in a day.
“The actual figure is likely worse, given that a large number of transactions are exchanges and miners moving bitcoins around and other low-value ‘dust’ transactions,” said Matthew Green, a cryptography expert at Johns Hopkins University. “So each transaction where there’s an exchange of goods or services happening is really representing even more electricity.”
As climate change becomes a more pressing concern for humanity every day, this huge level of energy use is difficult to justify for a currency wanting to improve on the current arrangement.
“It appears there are significant challenges to ensuring that Bitcoin’s growth minimizes environmental impacts,” offered Jeremy McDaniels, a financial system sustainability expert with the UNEP. “Energy footprints could be an issue of major scale-up is achieved.”
There is hope that Bitcoin may be able to reduce its footprint, however.
One important thing to understand is that the electricity demands of Bitcoin mining won’t scale up linearly with increased usage or transactions. Bitcoin miners use special hardware to guess over and over at solutions to computational problems for each “block,” which records transactions into a permanent ledger. The first problem-solver “wins” the block and the reward: brand new bitcoins.
Bitcoin can currently handle up to 360,000 transactions per day given current limitations built into the technology, according to Jorge Stolfi, a computer science professor from Campinas University in Brazil, so there’s some headroom left before things bog down.
It would be possible to bring down the average power cost of each transaction by modifying the underlying Bitcoin protocol, but that’s no easy feat. The Bitcoin community is currently debating a big change that would mean the network could theoretically handle about 7.2 million transactions a day on a comparable level of electricity consumption, according to Stolfi. That would require a majority of the people mining Bitcoin to agree to the change, however.
Keeping power consumption high in general also makes the network more secure by making it harder for any one entity to gain control. “The right way to think about this is that the energy expenditure provides a level of protection against attacks—it establishes a price floor, currently in the many millions, to launch a 34 percent or 51 percent attack [where an attacker can block transactions and double spend bitcoins as they please],” Emin Gun Sirer, a Cornell professor and blogger at Hacking Distributed, explained in an email.
However, that same level of security could be maintained while allowing for more transactions, he said, shrinking the cost per transaction.
All that needs to happen, then, is to expand the userbase so we have more transactions, right?
Unfortunately for Bitcoin, if user adoption spikes, so will price—and so must power consumption. Bitcoin mining leads to an arms race among miners to grab a slice of the fixed rewards doled out by the network, Stolfi said. The higher the financial rewards, the more miners will invest in powerful equipment to keep up with the competition. The Bitcoin protocol will continue to increase the difficulty of the cryptopuzzles to keep rewards constant, continuing the arms race until the last block is mined.
That makes Bitcoin about 5,033 times more energy intensive, per transaction, than VISA
The bottom line? Price = energy. “The total revenue of the mining industry is Bitcoin price times BTC revenue in USD/day, independently of anything else; and the electricity consumption, also in USD/day, is some large fraction of that,” concludes Stolfi.
Green agrees: “Almost everything in Bitcoin is flexible, but that dynamic isn’t. Miners always have the incentive to throw as many hashes [requiring power] at the job as the price dictates.”
Of course, it wouldn’t be fair to knock Bitcoin’s electricity consumption without comparing it to payment systems most people use today. Let’s take VISA as an example.
According to Network Computing, the VISA network can process more than 80 billion transactions per year or 2,537 transactions per second, using two mirrored data centers, each capable of running the entire network. The larger data center is currently pulling enough power for 25,000 households’ daily electricity, so we’ll double that to account for VISA’s total draw. In 2013, VISA’s investor reports say the company processed 58.5 billion transactions.
Working off these (admittedly imperfect) figures, each VISA transaction consumes around 0.0003 household’s daily electricity use. That makes Bitcoin about 5,033 times more energy intensive, per transaction, than VISA, at current usage levels.
Both networks use a lot of houses worth of daily juice, but one of them processes millions more transactions. Image: Motherboard
Of course, VISA runs call centers, offices, and a whole lot else on electricity as well, which isn’t counted in this comparison. But those hardly matter due to the extreme difference between the two figures.
In a rosy 2014 Bitcoin sustainability study, Bitcoin analyst Hass McCook concluded that “Bitcoin has 99.8% fewer [carbon] emissions than the banking system,” which we can treat as a rough proxy for energy use. The study neglects to account for the vast size difference between the Bitcoin economy and the conventional money system, however—the world banking system’s market capitalization in 2010 alone was over 1,989 times bigger than today’s total Bitcoin valuation.
In light of the above analysis, Bitcoin’s power usage per transaction isn’t remotely sustainable as a wholesale replacement for the conventional financial system. In the future, Bitcoin could massively gain popularity, pile on millions more transactions, and still be unsustainable due to the arms race between miners.
In an email, Bitcoin expert Piotr Piasecki added some context to the comparison: “With the increase of the block size, there will be more transactions included in the block, so the cost per transaction should go down. While it might not reach such low levels as Visa, we are talking about two somewhat different systems. One is a database entry in a single system, another one is an immutable record of history in a decentralized ledger.”
Watch more from Motherboard: Life Inside a Chinese Bitcoin Mine
Piasecki makes a good point, but Bitcoin will need some adjustments to increase capacity before the energy price of that immutable record is worthwhile. He also noted that numerous alternative digital currencies or altcoins, although currently worth a lot less than Bitcoin, could handle transactions in a less energy-intensive way.
Bitcoin mining will no longer reward new coins once 21 million coins have been mined, so the network’s power draw could drop considerably once there is less incentive for the mining arms race. But since the last Bitcoin block is projected to be mined around the year 2140, adopting Bitcoin as a major (or world) currency anytime in the next few decades would just exacerbate anthropogenic climate change by needlessly increasing electricity consumption until it’s too late.
McDaniels did offer a few ways for Bitcoin to improve its sustainability rating in light of the miners’ energy consumption: by offering low-barrier financial inclusion to the unbanked, and by leapfrogging current systems around physical bank branches. But it’s notable that both of these are already taking place without Bitcoin through popular mobile payment systems like M-Pesa.
The upshot of all this is pretty clear. Modified blockchain tech may well find a use case behind the scenes in the financial sector. But for now, energy-conscious people looking to use an alternative currency for the masses should probably look somewhere other than Bitcoin.
Corrections: An earlier version of this story suggested that mining will stop after 21 million bitcoins are created; in fact, mining will continue, but the block reward, and therefore much of the incentive, will stop. An earlier version of this story suggested a 34 percent attack or 51 percent attack could allow an attacker to control transactions; in fact it will only allow an attacker to block transactions and double spend bitcoins.