This article was compiled and organized by BlockWeeks
Is the electricity consumption of the Bitcoin network an acceptable use of energy? Since Bitcoin's inception, this question has been debated repeatedly. Despite the large number of published reports and analyses, the debate has never ceased—and it often resurfaces with every bull market cycle.
To answer this question, one must first clarify a premise: Bitcoin is an entirely new technology, and it does not precisely replace any existing system. It is not merely a settlement layer, not merely a store of value tool, and not merely a medium of exchange. The Bitcoin network does consume a large amount of energy, but it is precisely this energy consumption that forms the foundation of its robustness and security.
Because Bitcoin's ledger is transparent, estimating its energy consumption is relatively easy—which also makes it the easiest target for attack. By contrast, few people ask the same standards of traditional industries. Bitcoin is most often compared with two mature systems: the traditional banking system (payments, savings, and settlement) and gold (a non-sovereign store of value). But the energy consumption of these two industries is not publicly disclosed, and outsiders have almost no way to audit it. If there is to be an honest conversation about Bitcoin's energy consumption, it should be measured alongside these industries that are most often used for comparison.
The Composition of Bitcoin's Energy Consumption: 99.8% Comes from Mining Machines
Bitcoin's direct energy consumption comes from three places: the nodes that validate and relay transactions, the mining pools that coordinate global miner activity, and the mining machines. The vast majority of this, about 99.8%, comes from the operation of mining machines. These specialized integrated circuit machines, known as ASICs, have only one task: to perform proof-of-work (PoW) to produce new blocks, settle transactions, and maintain network security. PoW cryptographically binds a virtual asset to the physical world through the simple requirement of "must consume energy," which is one of Bitcoin's most revolutionary designs.
As of the time of the research, the total energy consumption of the Bitcoin network is estimated at approximately 113.89 TWh/year. A few comparative figures can make this number easier to understand:
- Global annual energy supply is about 166,071 TWh, which is 1,458.2 times that of the Bitcoin network;
- Global annual electricity generation is about 26,730 TWh, which is 234.7 times that of Bitcoin;
- The electricity lost each year in transmission and distribution is about 2,205 TWh, which is 19.4 times that of Bitcoin (based on World Bank and IEA estimates);
- The energy consumption of appliances in U.S. households that are "constantly on standby" is about 1,375 TWh/year, which is 12.1 times that of Bitcoin.
These comparisons show that, under basic assumptions, Bitcoin's energy consumption can be estimated in real time. But the accusation that it "consumes a lot of electricity" must first answer one question—relative to what?
Gold and Banking: "Best-Effort Estimates" for Opaque Industries
Since public data is scarce, the researchers chose to make their methodology fully open source for others to review, and made best-effort energy consumption estimates for the gold industry and the banking system respectively. The gold portion covers mining, refining, vault storage, and other stages; the banking portion is more complex—no bank is required to disclose its own electricity usage, so arriving at a comprehensive figure can only rely on voluntary disclosure by individual banks.
On the payments side, measurement can start from public data: using 8,760 hours per year as the calculation basis, the research estimates that the VISA network's electricity consumption is about 0.84 TWh/year; in 2019, VISA processed 185.5 billion transactions, part of the global total of 441 billion card transactions. Adding up the operational electricity consumption of all card network organizations worldwide, the estimate is about 2.00 TWh/year—a figure that does not even include the electricity used by card organizations' office buildings.
The banking industry as a whole, based on available public information, is estimated to consume about 263.72 TWh per year. But it must be emphasized that this is only a partial view: central banks, clearing institutions such as the DTCC, and other parts of the traditional financial system are excluded. The true energy consumption of the entire banking system remains unknown to this day, and its externalities have not been accounted for.
Comparing by Transaction Count Is Itself Wrong
Even if these figures are added together, the comparison between Bitcoin and the various layers of the banking system remains inherently unfair. Traditional finance is a complex hierarchical structure stacked with counterparties and intermediate settlement layers: credit card networks excel at fast payments and the circulation of IOUs, but rely on the banking system to complete settlement; the banking system in turn relies on central banks to complete final settlement; and central banks rely on the entire dollar system.
More importantly, the two scale in different ways. The banking system expands as the number of transactions grows—the more transactions, the more infrastructure is needed; Bitcoin's energy consumption does not expand in sync with the number of transactions. Using "how much electricity is consumed per transaction" as a verdict forces a system that scales by volume and a system that scales by security budget to be measured against the same standard.
Attack Costs and Miners' Economic Rationality
The security purchased by energy consumption can be quantified. To launch an attack on the Bitcoin network, an attacker would need to accumulate computing power equivalent to adding 16 GW of mining machines and bear the enormous electricity costs of running those machines—an extremely massive undertaking. Moreover, the network is not static: by the time an attacker accumulates a meaningful share of computing power, the network's hash rate will have already continued to climb.
Miners are first and foremost participants driven by economic interests, actively seeking the lowest-cost electricity to maximize profits. Because miners across the network compete with one another, everyone has an incentive to keep operating costs down and position themselves at the lowest end of the cost curve—even if that means migrating from one energy-rich region to another. Even if miners are bullish on Bitcoin, they will only invest in mining farms they expect to be profitable; once profitability disappears, they will shut down and exit. After the most recent halving in May 2020, the network's hash rate fell by 23% due to lower revenue, while highly operationally efficient miners remained profitable and benefited from reduced competition.
Conclusion: Energy Use Itself Is Not a Sin
This research uses quantitative methods to place Bitcoin's energy consumption on the same table as other industries, but it also acknowledges that Bitcoin's value is not only quantitative.
Subjective views on the importance of the Bitcoin network vary from person to person, but Bitcoin's properties do not change from person to person: anyone can use Bitcoin, anyone can hold Bitcoin themselves, and any Bitcoin transaction can achieve probabilistic final settlement in about one hour—24/7, all year round. For those who cannot enjoy stable and accessible financial infrastructure, these characteristics mean financial freedom. For the Bitcoin network, intermittent energy and surplus energy have found a perfect use, which in turn can also benefit the energy sector. And as long as adoption requires it, the network will continue to scale.
Energy use itself is not necessarily a bad thing. As Vaclav Smil wrote, energy is the only universal currency. Humanity will continue to invent new technologies that require more energy to challenge the existing order, and Bitcoin is just another example.
So back to the original question: Is the electricity consumption of the Bitcoin network an acceptable use of energy? The answer is clear—yes.
