Remember when Bitcoin’s energy consumption was going to destroy the world? It turns out that the planet’s demise at the hands of crypto miners may have been greatly exaggerated. Bitmain’s recent financial filings and a new change in in how a much-touted number, the Bitcoin Energy Consumption Index, is calculated should allay fears that crypto will singlehandedly destroy our planet.
Long ago—December 5, 2017, to be exact—panic appeared to set in among the chattering class. The prices of cryptocurrencies were skyrocketing and journalists acted like they needed to justify missing out on the tether-fueled rally. On that day, Bitcoin changed hands at around $11,900, nearly 10 times where it was just six months prior, according to data from CryptoCompare. And it was also on that day environmental tech site Grist published a piece that gave the type of publications that publish clickbait-y “hot takes” a reason to publish clickbait-y hot takes.
“Bitcoin could cost us our clean-energy future,” wept the Gist headline in an article by meteorologist Eric Holthaus. “By July 2019, the bitcoin network will require more electricity than the entire United States currently uses,” he wrote in a passage that got the world’s attention. “By February 2020, it will use as much electricity as the entire world does today.”
A month later, we described the reaction:
Newsweek jumped on story. Wired wired in. CNN sent chills down globally warming spines with “Bitcoin boom may be a disaster for the environment”. Forbes, which has sold out its brand equity and lets anyone publish anything they want as if it were LiveJournal with a premium name, published “Here’s Fiat Money’s Advantage Over Bitcoin That Many Seem To Forget,” capped with a headline meant to troll free market libertarians like, say, Steve Forbes. Public Radio International said they read the Grist story and checked the math. The Verge published a question in their headline, “Can renewable power offset bitcoin’s massive energy demands?” And if that weren’t enough, it also had another question in its subheadline: “Some miners are going green, but will that be enough?” (Rule of thumb: the answer to any question in a headline is generally “no” but we’ll get to that).
Then there were the bloggers. Thinkers thought thoughts on the topic. So many news outlets and people rewrote the same story over and over again and it usually centered on how terrible humanity is and bitcoin is bad because people are bad. You know, the usual.
[Even on our own pages, we’ve published pieces that advocate other types of blockchains in part because of Bitcoin’s environmental costs.]
As we also explained in subsequent paragraphs, Holthaus’ piece hinged on the big assumptions of Digiconomist. That’s a blog covering crypto energy use and is a side project run by Alex de Vries, a PwC consultant. At the end of 2017, his “Bitcoin Energy Consumption Index” graphed an exponentially upward-sloping chart.
The methodology behind the index was to use sales of new crypto mining machines—the computers that keep the Bitcoin blockchain running—to calculate the number. Here’s how de Vries justified his methodology and it’s where Bitmain comes in:
“The best support ended up coming from the biggest manufacturer of Bitcoin mining machines itself. Bitmain decided to file for an Initial Public Offering (IPO), and disclosed that it had sold 2.56 million mining machines in the first half of 2018 alone. Assuming that each machine consumes about 1,500 watts each (the consumption of an Antminer S9), all of these machines combined would be responsible for an annual energy consumption of 34 terawatt hour (TWh). Over this time period, the Bitcoin Energy Consumption Index showed an equal size increase in Bitcoin’s energy consumption from 37 to 71 TWh per year. Additionally, documents showed that Bitmain had sold 1.62 machines in the full year of 2017. These machines could easily be responsible for a total energy consumption of 21 TWh per year. This way, the combined production of 2017 and the first half of 2018 could represent a total consumption of 55 TWh per year. It should be noted that Bitmain is estimated to have a market share of 67%. Hence, if 67% represents 55 TWh, 100% could represent about 82 TWh per year. The latter number even exceeds the energy consumption estimate provided by the Bitcoin Energy Consumption Index on June 30 2018 (71 TWh per year).”
Unfortunately for Bitmain—though maybe good news for the environment—2018’s third-quarter sales were miserable. In the first half of the past year, Bitmain reported revenue of $2.8 billion, according to their IPO filings. But when total numbers came out for the first nine months of 2018, it showed revenue of $3 billion. In other words, sales were just $200 million from July to September of 2018. That puts monthly average sales at merely $66 million, or 14 percent of what they were from January to the end of June.
Let’s assume—as de Vries used to—that all Bitmain sold were its now-obsolete Antminer S9s (an awful assumption because the company’s newer machines are twice as efficient as the S9) at a price of about $1,093.75. That means just 183,000 or so units were sold in Q3. According to de Vries’ old math, if we kept hashrates (simply explained, the rate for a machine to conduct a Bitcoin operation) the same, we would also get an increase of annual energy consumption of roughly 2.4 TWh. Now if we improve on one assumption and say that Q3 units were roughly twice as efficient as the S9—which are what Bitmain’s newer models are—average energy consumption would instead increase by 1.2 TWh.
Sure enough, a small footnote on de Vries’ Bitcoin Energy Consumption Index page announced a change last week:
“The minimum is calculated from the total network hashrate, assuming the only machine used in the network is Bitmain’s Antminer S9 (drawing 1,500 watts each). On February 13, 2019, the minimum benchmark was changed to Bitmain’s Antminer S15 (with a rolling average of 180 days).”
What’s more, we don’t know the monthly breakdown of Bitmain’s unit sales. If the majority of its transactions were in the earlier weeks of Q3 (our guess is that it is), then we’re witnessing a painful collapse of the crypto mining business and a lot of the environmental impact it had.
Also not factored is that hashrates plummeted from the end of Q3 to now. From a seven-day average peak of 54.8 million tera (trillion) hashes per second on October 1, 2018, it plummeted to as low as 34.8 million in mid-December. As of this publication, the average has inched up to 43.6 million. The Bitcoin Energy Consumption Index has gone sunk to 45.8 from a high of 73.1 in November.
Thus more efficient machines dealing with falling hashrates means projections of a Bitcoin-generated energy crisis weren’t just far off, they were irresponsible. According to de Vries’ own data, Bitcoin now uses the equivalent amount of energy as the island city-state of Singapore and just 1 percent of the total energy consumption of the United States—so about 100 times less than what was forecasted for this year.
That isn’t to say Bitcoin isn’t without problems. Market manipulation and mania led to the run up in prices that, in turn, caused billions of dollars in resources to be spent on mining. And, yes, those are resources—money and energy—that could have been put to more productive use. Yet the market, when informed, sometimes has a funny way of severely punishing those on the wrong side of things as it does lavishing reward on those who get it right. Just ask Elizabeth Holmes and Theranos.
As for Eric Holthaus, author of the story that got everyone so riled up, his latest piece advocates adoption of the “Green New Deal.”