The only thing more over-hyped than Bitcoin is the question about the fraction of world energy now consumed by the trendy virtual currency — and how much it will use in the near future. A flawed new analysis is only adding fuel to the fire.
“Bitcoin mining on track to consume all of the world’s energy by 2020,” a Newsweek headline claimed in December. In January, the New York Times wrote “The total network of computers plugged into the Bitcoin network consumes as much energy each day as some medium-size countries — which country depends on whose estimates you believe.”
The New York Times’ statement is simply false. It confuses “electricity” with “energy,” which is a mistake much of the media continues to make. A nation’s total energy consumption includes oil used for transportation, as well as fuels used for heating, like natural gas, coal, and wood — electricity is just one part of the whole energy system. So, Bitcoin may soon consume as much electricity as some modest size countries, but electricity is well under half of total energy consumption.
The Times statement is also based on an analytical approach that uses an “invalid methodology,” as Dr. Jon Koomey, a leading expert on information technology energy use, explained to ThinkProgress.
Predictions of Bitcoin’s future energy use ignore widespread efforts aimed at making virtual currencies far more efficient.
Creating new Bitcoins and verifying transactions using so-called blockchain “proof of work” technology is certainly electricity intensive. Anyone and indeed everyone who can solve very complicated computational problems can validate transactions, and in turn earn Bitcoins.
Here’s one of the most popular online video explanations of this complex scheme:
Because the central process behind Bitcoin transactions is computationally intensive, it is electricity intensive. That drives Bitcoin miners toward ever greater computing power and electricity consumption, especially as Bitcoins soared in value in 2017.
Calculating how much total electricity is being used around the world for every Bitcoin transaction is a very complicated task. “There is so little credible data on these [Bitcoin mining] facilities that there will be uncertainty on Bitcoin electricity use for some time to come,” explained Koomey, who led the End-Use Forecasting group at Lawrence Berkeley National Laboratory (LBNL) for many years. So “people should stop jumping to conclusions based on poorly estimated guesses.”
One of the estimates that has received a lot of attention is by Alex De Vries, who blogs at Digiconomist.net and created the Bitcoin Energy Consumption Index. On Wednesday, De Vries published a new Commentary in Joule, “Bitcoin’s Growing Energy Problem.”
An accompanying press release asserts, “Bitcoin estimated to use half a percent of the world’s electric energy by end of 2018.” Since total global electricity consumption is about 25,000 TeraWatt-hours), 0.5 percent would be 125 TWh (125,000 GWh).
Koomey explains the problem with the analysis:
De Vries uses an invalid methodology to do his estimates, making guesses about electricity prices paid by bitcoin miners as well as the percentage of bitcoin price that is [tied to] energy to calculate electricity consumption. No credible analyst would ever do this. The correct way to calculate these numbers is to build from the bottom up, using knowledge about actual bitcoin technologies and operations to make sure the parameters are in the ballpark. That’s what Marc Bevand does, with the caveat that there is a lot of uncertainty (a fact that Bevand explicitly acknowledges).
The top down approach is like trying to estimate U.S. energy consumption starting with U.S. GDP and making some back of the envelope calculations (about how much of GDP goes to energy and then what the cost of energy is). Bevand’s bottom-up approach would look at actual hardware — cars, houses, airplanes, and steel mills — and their actual energy consumption and then try to add it all up. It’s harder, but more accurate.
Note that Bevand’s best estimate is that as of January 2018, total global Bitcoin electricity consumption is 18.4 TWh, which is under 0.1 percent of global electricity consumption.
“If the price of Bitcoin continues to increase the way some experts have predicted,” the news release for the article warns, “de Vries believes the network could someday consume 5 percent of the world’s electricity.”
Yet as Koomey explained to ThinkProgress, “people need to see the larger context: Data center electricity use as a whole (of which Bitcoin is only a small part) has been flat since 2008 or so.” And it comprises only about 1.8 percent of all U.S. electricity.”
Koomey was a coauthor of LBNL’s 2016 “U.S. Data Center Energy Usage Report.” For the past two decades, he has been debunking claims that the Internet will engulf an ever-growing fraction of the global electricity. He and LBNL have also been working with IT companies to help them become more efficient.
Indeed, a major reason why data center electricity consumption did not achieve the kind of worst-case projections that had been made over a decade ago — the kind that are now being made for Bitcoin — is that the industry adopted more efficient hardware and more efficient software, to minimize energy costs.
That is precisely what is happening now with virtual currency. Some of the top universities and companies are working to make blockchain more efficient, as CNBC reported in February. Some are pursuing a different validation process that would be vastly less computationally and energy intensive.
Finally, Intel “recently filed a patent for a system-on-a-chip (SOC) that includes hardware to accelerate bitcoin mining and reduce its power consumption,” as Investopedia reported earlier this month. “The hardware is optimized to reduce space and energy utilized during the process.”
The bottom line is that the bitcoin energy crisis has been overhyped and to the extent that there is wasteful energy use, top engineers and professors are working to solve the problem.