Headlines over the past several months addressing energy use and Bitcoin mining would have one believe the world’s supply of power is about to be devoted solely to large computers trying to solve complex algorithms in the global pursuit of profit. Is this really the case?
One head-spinning title from Newsweek reads: “Bitcoin mining on track to consume all of the world’s energy by 2020.” The reasoning behind this piece is that if the explosive growth that occurred during the late 2017 frenzy were to continue at the same rate, we’d be out of juice in three years. By that logic, if a 6-month old baby keeps growing at the same rate as has occurred from birth, he or she will be bigger than a brontosaurus within a few short years! But the world just doesn’t work that way. There is, however, some concern that Bitcoin mining energy use is increasing, and that it has the potential to pop-up quickly and in large increments in low-cost electricity markets.
What is Bitcoin mining?
Mining is the process of confirming blockchain transactions that have occurred in the prior ten minutes, sweeping all of the unique transaction identifiers into a list and creating a new and unique number (hash) that seals the block in the chain.
Every newly created hash is placed at the end of the most recent block. It includes the hash from the prior block so that if somebody tries to tamper anywhere within the chain, subsequent hashes would change and users would be alerted. The process of identifying that unique numerical sequence (programmed to become increasingly more difficult over time, thus increasing the energy intensity of the process) is based on consensus. Computers all around the world race to find that unique number and then agree on its creation. The winner is currently awarded 12.5 Bitcoins.
Energy use calculations built on an erroneous foundation
The articles comparing blockchain energy use to the consumption of Denmark, Ireland, Nigeria (pick a country!) are all based on the initial calculations used by digiconomist, which set up a Bitcoin Energy Consumption Index. The problem is in the methodology employed, which starts with estimating the revenue to be gained by the global Bitcoin mining endeavour.
The model then estimates the percentage of revenues dedicated to the power bill. It subsequently looks at how much the miners pay per kilowatt-hour to estimate the total of kilowatt-hours. By that logic, if the price of Bitcoin doubles from one month to the next, the energy used in mining is assumed to double as well. Of course, there will be more machines dedicated to the pursuit and the ones on the margin will likely be less efficient. But this macro linear ratio approach – while it generates amazing headlines - doesn’t provide much in the way of true insight, and has its skeptics.
There are also simple constraining factors that would keep those types of price responsive dynamics from kicking in. First, we wouldn’t be able to build data centers fast enough. Second, since the price has collapsed from $20,000 to less than $7,000 between December of 2017 to April 2018 - did energy use also collapse, so that two thirds of Denmark, Ireland, or Nigeria’s power is no longer used? Put another way, the Newsweek headline could now read “At the current rates of price decline, Bitcoin mining on pace to consume none of the world’s power by 2019.”
That having been said, clearly the industry is consuming a good deal of electricity, though nobody really knows how much, and the estimates vary widely, from 3,400 MW (Digiconomist) to perhaps 540 MW (an estimate in Bitcoin Magazine) as of late 2017. Without doubt, the industry has been growing rapidly, in part because it has been so lucrative. In 2017, it was estimated that Chinese company Bitmain – the largest player – made $3-4 billion in profit.
China’s crackdown may ultimately give us a better view of the global industry
In early 2018, however, Chinese authorities signaled they intended to orchestrate ‘an orderly exit’ from cryptocurrency mining, which helped prices collapse and forced Chinese miners to investigate other host sites.
Chinese miners began to scout other locations with low-priced power, including the Upstate New York, the U.S. Pacific Northwest, Iceland, the Nordics, and the province of Quebec (all of these markets enjoy cheap hydro). Hydro-Quebec (HQ) initially began luring miners, but was quickly overrun by interest. A spokesperson for HQ indicated that “Of the world’s top five largest blockchain players, we have at least three or four.” In a matter of months, some 100 mining companies had inquired about possibly setting up shop, with demand for new facilities said to range from 20 to 300 MW. The level of stated interest suggested that perhaps the industry was indeed quite large after all. Bloomberg noted that HQ currently hosts 40 data centers using 350 MW of electricity, and HQ had received hundreds of requests totalling over 9,000 MW. The critical question to be asked is how many of these were viable operations and how many were kicking tires? (*See footnote)
The bottom line: How much energy is being used?
Given the dizzying price swings, global markets and migrations, and changing technologies, nobody has a really good handle on how much electricity is being used by the Bitcoin mining industry. It’s probably not equivalent to the energy consumption of the countries listed above. But if the 9,000 MW of requests to HQ were real, rather than largely speculative. Governments and power planners are going to have to do more to understand this fast-moving, rapidly-growing industry, as it is by far the single fastest potential area of electricity load growth in economies powered by low-cost electrons.
- The Quebec government has since signalled that it is no longer interested in providing cheap electricity to Bitcoin miners if no other value is added to the economy. Across the border in upstate New York, Bitcoin miners were rebuffed by the state regulator, after they set up shop in areas supplied by small municipal utilities with limited supplies of cheap hydro-electric power, thus driving up costs to ratepayers.