In this, the first of a three-part series, Danny Bradbury explores the impact the bitcoin network is having on the environment.
Just how much carbon dioxide do we produce when we mine a bitcoin? It's becoming an increasingly important question. After all, it's great to disrupt an inefficient and sometimes corrupt incumbent economic system, but most of us would rather not do it at the expense of the planet.
The bitcoin network is stuck in a circle that drives up its power usage. People tend to put more computing power on the network so that they can make more more bitcoins. The software underpinning the network reacts by changing a parameter that makes it more difficult to solve the mathematical problem needed to solve a bitcoin block.
Because it's harder to solve the problem, people add even more computing power, and so on. As this cycle increases, it takes more electricity to mine a bitcoin. The hashing power of the network surpassed the world's top 500 supercomputers almost a year ago, and things have moved along quite a bit since then.
"The Red Queen is originally from Alice in Wonderland. In the Queen's race everyone runs faster, but you never get ahead," he says. "The same happens in hashing. All of the participants are co-adapting. You have to keep adapting to keep up."
Gogerty has been trying to put together a model for calculating the carbon cost of a bitcoin, but he admits that it needs work, and he is asking for volunteer help to improve it.
Other attempts have been made to nail down the cost of the bitcoin network in terms of carbon emissions and/or energy used, but it's a tricky business, says Guy Lane, founder of sustainability advisory service Sea O2.
Based in Brisbane, Australia, Lane is also the founder of Bitcarbon.org. That site contains his method for tracking bitcoin-based carbon emissions.
There are caveats. "Of course, bitcoin is mined everywhere from data centres, distributed locations, working pools, rigs set up in garages and even on PCs that have been hijacked by bots," Lane says. This makes it very difficult to ascertain the true carbon cost, because there are so many different types of equipment running the mining software.
Still, it doesn't stop him trying. His method is based on the premise that miners will spend up to 90% of the cost of a bitcoin on the electricity used to mine it. That electricity cost will naturally vary with the price of a bitcoin.
“In the Queen's race everyone runs faster, but you never get ahead.”
The method assumes that 50% of all the mining takes place in China or the US. It uses the most recent estimates from the International Energy Agency for carbon emissions per kilowatt hour (kW·h) of mains power in either country, and averages them (a kilowatt hour equates to outputting one kilowatt - or 1,000 watts - of power for a period of one hour). The result is that for every megawatt (MW) of electricity spent mining bitcoins, 0.65 tons (1300lbs) of CO2 are released into the atmosphere, it says.
The method maps these figures against average electricity prices, to produce an average carbon intensity of 6.98 kg (15.38lbs) of CO2 for every dollar that is spent on electricity used for mining Bitcoin.
Lane last ran this model in December, when bitcoin was priced at $1000, and calculated that the entire bitcoin network was putting out about as much carbon as Cyprus.
Lane believes that the estimates could be lower than the reality. For one thing, it doesn't account for the fact that miners might be willing to spend more electricity mining a bitcoin than the current value of that coin, in the hope that it may increase.
Mining for answers
Another way to look at carbon output for bitcoin mining is to go and ask a professional miner. If anyone should know how many hashes and therefore how much power it takes to compute a bitcoin, it's an institutional miner that makes a business of turning electricity into cryptocurrency.
Dave Carlson, founder of Megabigpower, runs a massive bitcoin mining datacentre in Washington state.
10 TH/sec (10,000 GH/sec) make 1 bitcoin per day at the current difficulty, he says. His hardware uses one watt per GH/sec, meaning that it takes 10,000 watts (10kW) to run 10TH of equipment.
He runs that 10kW of equipment for a whole day to mine a bitcoin, which means that he spends 240 kW·h. That's 24% of a megawatt hour (MW·h).
Remember that according to the IEA data, 1 MW·h of mains electricity produces 1300lb of carbon. Based on Carlson's figures, that means that the energy he's using would release 24% of that, or 312lbs, of carbon dioxide into the air per coin.
According to the EIA, that's about the same as burning 15.9 gallons of gasoline, without ethanol.
That would be a lot of carbon for Carlson to churn out, if his electricity was produced by fossil fuels – but it isn't.
"We are 100% hydroelectric," he says, adding that he hopes to announce the largest solar/wind powered mine later this year. "I am also looking at re-investment in wind power generation (mostly as a hedge against power prices rising). We are very aware of our carbon footprint and the likelihood that it will increase."
He isn't the only one. Over in Sweden, ASIC mining manufacturer KnCMiner uses a co-hosting facility. The electrons it runs on also have a distinctly green hue.
"What I can say real quick is that our data centre is run on hydropower. So we are about as green as they get," says co-founder Sam Cole.
So, a lot of bitcoins are being produced with green energy. But if you are burning fossil fuels for your bitcoins, then using just over a sixth of a ton of carbon for a single bitcoin isn't good, given that the network is churning out 150 of them per hour.
It's important to put this in perspective, though, by understanding what this is relative to, and what we're getting for this carbon throughput. That's what we'll be looking at tomorrow, in the second article of this three-part series.
Power plant image via Shutterstock
Updated: 5:30 GMT April 8. Kw/h and Mw/h symbols updated to reflect correct symbols for kilowatt-hours and megawatt-hours.