There's little debate that mining cryptocurrency uses a lot of energy. Mainstream news outlets seem extremely concerned about this, while crypto-enthusiasts seem extremely dismissive. The concerned argue that the carbon footprint (now bigger than that of many small countries, and growing) is too dangerous to ignore, while proponents argue that it will offset even bigger emissions from the bloated financial industry.
Writing from 2021, it's clear that these concerns have been wildly overblown in the past (See, from 2017: Bitcoin Mining on Track to Consume All of the World's Energy by 2020), and Elon Musk's decision to suspend Tesla purchasing with bitcoin has re-invigorated the debate. My intent here is to establish a more balanced understanding of the situation to help myself (and maybe others) navigate the contentious discourse.
Why does cryptocurrency use so much energy?
Cryptocurrency transactions are logged on blockchains- public ledgers and sources of truth. To maintain accuracy and trust in a decentralized system, most cryptocurrencies (including Bitcoin) use an energy intensive process called proof of work.
Adding blocks to a proof-of-work blockchain requires miners to generate huge quantities of guesses (hashes), competing against each other find the right answer first. Better analysis will not expedite this process, but more computational power will give you a higher probability of guessing correctly. This ensures that blocks cannot be added too quickly and that modifying existing blocks is (for all practical purposes) impossible.
Basically, we trust that these blockchains are accurate precisely because they required such a huge amount of work to construct and would therefore require a huge amount of work to change.
Miners are willing to spend money on energy because "correct answers" are rewarded with transaction fees and newly minted digital coins. When the price increases, the rewards become more valuable. This incentivizes miners to spend more money on electricity to generate more hashes. The software then adjusts the difficulty of the problems so that the time between blocks is relatively constant (about 10 minutes for Bitcoin), regardless of computational input.
Currently, new coins make up the vast majority of miner rewards, so energy usage scales primarily with the price of the currency, not the number of transactions. Miners will use as much energy as is economically viable.
How much energy is actually being used?
While Bitcoin is of course not the only cryptocurrency, I'll use it as the primary example here because it is the largest and best-documented.
Although Bitcoin's energy demands primarily scale with its price, which is well known, estimating energy use and emissions is not as simple as many articles might lead you to believe. Because there's no regulating body to provide concrete, verifiable numbers, calculating energy use and emissions requires many estimations. This introduces uncertainty and fuels a lot of internet debate.
That consumption yields an estimated 55 million metric tons of annualized CO2 emissions, roughly 0.2% of global emissions from fossil fuel use.
Bitcoin gets a lot of attention for its carbon footprint, but high energy use is inherent in any proof-of-work blockchain. Ethereum's energy demands have recently skyrocketed to 46.95 TWh, sending associated emissions to an estimated 22.3 million metric tons of CO2. Combined, the two could account for nearly 1% of total global electricity consumption.
There's no escaping that these estimates rely on a lot of assumptions and come with some uncertainty. But given what we know about the profitability of Bitcoin mining and its intrinsic energy requirements, it makes sense that these numbers are big. Including the demands of other cryptocurrencies will only increase the estimates further.
Could it be greener?
Some have claimed that continued hardware improvements will solve the problem by making mining operations increasingly efficient. But mining does not require a fixed number of computations. If hardware becomes twice as efficient, then the same energy input will yield twice as many computations, and the problems will become twice as difficult, leaving energy demands unchanged.
Likewise, increasing the number of transactions per block could drive down energy costs per transaction, but miners would still be incentivized to add blocks as quickly as possible, using as much energy as is economically viable.
Proof-of-work cryptocurrencies inherently use a lot of energy, but that doesn't necessarily imply a large carbon footprint. If clean electricity is affordable and reliable, miners will use it.
Unfortunately, clean electricity is notoriously inconsistent. There's promising evidence that many miners use excess hydroelectric power near the Three Gorges Dam, but seasonality in hydropower means that this demand is often met by burning coal during the dry season, negating any environmental benefit. There are also discouraging reports of fossil-fueled power plants reopening or expanding to mine cryptocurrency. Currently, only about 39% of Bitcoin mining is powered by renewables, and claims to the contrary are likely overly optimistic.
There are less energy-intensive alternatives to proof of work, but even the most prominent face lingering concerns about viability. Ethereum is planning a switch to proof of stake, but it seems unlikely that Bitcoin would follow. For now, Bitcoin and most of the crypto-economy seem tied to high energy demands and high emissions.
What about other big costs?
Bitcoin might use a lot of energy and release a lot of CO2, but what about the rest of the financial sector?
This defense is a good reminder that everything costs energy to produce, even the things we're used to seeing. Mining gold, printing fiat currency, and processing credit card transactions all use energy- more energy than Bitcoin does currently. But we should be careful to compare things fairly (apples to apples), and to make reasonable assumptions about the future of finance.
Cryptocurrency defenders argue that the centralized financial industry yields much higher emissions than Bitcoin mining. But that comparison is highly speculative and a little unfair. For one, even if cryptocurrency offsets large parts of the current system, mining will not be the only cost, and Bitcoin wont be the only currency. People will still want convenient interfaces, insurance, and financial guidance, all of which would require energy. Even if cryptocurrency's operational costs are lower than those it replaces, we can't do much more than speculate about what that world would look like or how big the offsets might be.
Gold mining, which releases roughly 81 million tons of CO2 per year, is perhaps a better comparison. In 2019, almost $200 billion worth of gold was mined, compared to roughly $16.5 billion worth of bitcoin (144 blocks per day * 6.25 bitcoin per block * $50,000 * 365 days). While its total emissions are higher, gold mining's marginal emissions (0.4kgCO2/USD) are significantly lower than bitcoin's (3.3kgCO2/USD). Of course, as a physical commodity, gold will consume additional energy for transportation, security, and storage. These figures are hard to quantify, but they could help close the efficiency gap.
And printing money? Conservatively, if $1 trillion worth of fiat currency is minted globally each year, yielding an estimated 6.7 million metric tons of CO2, it would have a marginal emissions rate of just 0.007kgCO2/USD: far lower than Bitcoin.
It's worth considering how Bitcoin compares to other systems, but it's also worth acknowledging the absolute scale of the problem. Regardless of what it offsets, most cryptocurrencies inherently rely on high energy consumption, which will yield high emissions for the foreseeable future.
Conclusion: Is this a legitimate concern?
This question is so contentious in part because it pits two internet darlings- climate change and cryptocurrency- against each other. Those defending their cause are prone to exaggeration and willful misinterpretation.
Despite some uncertainty about specifics or potential solutions, the costs of mining cryptocurrency are reasonably clear. Even if advancements in clean electricity eventually solve this problem, our emissions over the next few decades will have a big impact on whether we inflict irreversible damage on the environment. While cryptocurrency isn't the biggest emitter, it's not a trivial expense.
Whether you think that expense is worth incurring will depend on how you value decentralized finance. In the best case, where we avoid catastrophic inflation and corruption, the energy use seems like a bargain. In the worst case, this will have been an extremely expensive experiment. Speculating on the expected value of cryptocurrency is outside the scope of this article, but I think that's the crux of the question. Instead of frittering about details and pointing fingers at other big emitters, proponents of cryptocurrency should concede the costs and focus on making a strong case for the benefits.