Bitcoin is emerging as a valuable way to send and receive money, but this so-called cryptocurrency has some deep-rooted environmental concerns attached to it. The community of computer-based miners that create bitcoins uses vast quantities of electrical power in the process. The electricity-heavy process has led some experts to suggest that bitcoin is not an environmentally friendly endeavour.
Estimates vary about the amount of electricity a bitcoin consumes. In his testimony to the US Senate Committee on Energy and Natural Resources in August 2018, Arvind Narayanan, associate professor at Princeton University claimed that bitcoin mining accounts for about one per cent of the world’s energy consumption. “On today’s most prominent public blockchains, mining involves the computation of a large number of mathematical calculations, called hashes, in parallel. For example, as of this writing, miners of bitcoin, the original blockchain-based cryptocurrency, collectively compute about 50 billion billion hashes, or 50 billion giga hashes, every second.
“Most mining is carried out in large-scale commercial operations using purpose-built computing devices specialised to the task of repeatedly computing these hashes — and nothing else. These devices are housed in warehouses dedicated to mining, data centres. Substantial energy is required to operate the computing devices as well as to cool them to keep them within their operating temperature limits.
“An accepted method for deriving an estimate of the energy consumption of mining is to assume that all miners use the most energy efficient mining device available on the market. Commercial devices are accompanied by published specifications listing the number of hashes that can be computed per second using the device, as well as the power consumption of the device in watts. It is then straightforward to calculate how much power is required to compute 50 billion billion hashes per second using the most energy efficient devices available. I performed such a calculation and obtained an estimate of around 5 GW for bitcoin mining alone. This is slightly under one per cent of world electricity consumption, or slightly more than the electricity consumption of the state of New York. Other public blockchains also consume a substantial, albeit much lower, amount of energy.”
What is bitcoin
Bitcoin is a software protocol and peer-to-peer (P2P) network that enables the digital transfer of value across borders without relying on trusted intermediaries. Bitcoin is an open and permissionless system: anyone can participate in the network, as well as send, store, and receive payments without having to ask anyone for permission.
It has its own, native cryptocurrency called bitcoin (BTC) which acts as the universal unit of value within the bitcoin network. New bitcoins are issued, according to a transparent and predictable schedule, on average every ten minutes through a process called mining. While the Bitcoin protocol specifies that a maximum of 21 million bitcoins will ever be created, it is worth mentioning that one bitcoin can be divided out to eight decimal places. This means that one bitcoin corresponds to 100 million satoshi, the smallest base unit.
All bitcoin transactions are recorded in a public ledger that every network participant (node) stores locally. The ledger is represented using a particular data structure: transactions are bundled into data blocks, which are then cryptographically linked to each other. This process results in a growing chain of blocks –called the blockchain. The use of this specific data structure ensures that tampering with the transaction history (e.g., modifying past transactions) will be detected immediately by other network participants. The blockchain grows larger every day as new blocks of transactions keep getting added by special participants called miners.
Proof of Work (PoW) mining
At the heart of this energy consumption is a concept know as proof of work (PoW). Imagine a scenario where person one has ten bitcoins that he wants to send to person two. However, he changes his mind and suddenly wants to send the same ten bitcoins to person three instead. While both transactions are valid, only one can get processed as there are only ten bitcoins, not 20. So, in a decentralised system like bitcoin that has no central authority, who gets to decide which of the two are valid, but conflicting transactions will get processed?
“PoW mining is a mechanism for resolving this dilemma in a decentralised manner: instead of simply letting participants vote (and thus making the vote vulnerable to potential manipulation by attackers who can create multiple fake identities), the idea is to attach a financial cost to the vote,” Michel Rauchs, research affiliate at the Cambridge Centre for Alternative Finance at Cambridge Judge Business School, explains. “Anyone who wants to participate in the vote (miners) needs to prove that they performed some work – hence the term proof of work.
“This work consists of finding the solution to a cryptographic puzzle, which in simplistic terms can be thought of as guessing a random number. The only way of finding the random number — called a nonce in technical jargon — is to brute force all possible options. This way, the work cannot be faked but is trivial to verify by other network participants once the winning nonce has been revealed.
“Solving the PoW requires substantial computing power depending on the difficulty level: the more miners join the race, the more difficult the puzzle becomes. Rather than guessing numbers manually, miners operate specialised mining equipment (ASICs) that has been specifically designed to be particularly good at only one single task — solving the PoW.
“Miners need to incur financial costs in the form of capital expenditures (acquiring mining hardware) as well as operational expenditures (spending electricity to run and cool the machines). Miners are in constant competition between themselves: whoever finds the solution to the puzzle first obtains the right to add his block to the global ledger. In return, the successful miner gets rewarded for his efforts with newly minted bitcoin.
“In the event that two valid but competing blocks are found at roughly the same time by different miners, the chain will split into two branches. Network participants will always follow the longest chain, that is the branch that was most difficult to produce (i.e., required more computing power, and was thus more expensive to generate). The idea is that miners should have a financial incentive to play by the rules as they stand to gain more from being honest than if they were to cheat.”
Calculating energy cost of mining
PoW mining rests on the premise that a financial cost needs to be attached to the vote. It turns out that this cost primarily comes in the form of electricity that needs to be expended to run mining machines. And mining hardware consumes quite a lot of electricity. The more machines a miner operates, the more likely he is to find the solution to the puzzle. However, more machines also mean that more electricity is needed to run and cool the equipment, which in turn results in higher costs for the miner in question. Miners are thus always searching for abundant electricity sources at the lowest possible price.
While newer ASIC models are substantially more energy efficient than previous generations, they still consume a significant amount of electricity. Rising bitcoin prices make mining more attractive, as the potential reward increases in value. As a result, new mining hardware will get added to the network and lead to increasing electricity consumption overall.
The electricity consumption is thus intricately linked to total mining revenues (block subsidy and transaction fees), which are a function of the bitcoin price. It is not possible to exactly determine how much electricity bitcoin uses for a variety of reasons.
For instance, miners can choose between several ASIC models that can have different energy efficiencies: there is little data available on the exact market share of mining hardware, and miners often reconnect old, less efficient machines when block rewards go up in value. It is also difficult to determine what hardware miners are using as profitability may vary significantly from one region to another because of different electricity costs.
Moreover, mining operations require additional electricity to cool the machines to prevent them from overheating, and eventually breaking. Data centres can vary significantly in terms of how efficiently they use electricity: while cooling and overheads for the most efficient facilities account for less than two per cent of the electricity used to run the mining equipment, less efficient data centres can have significantly higher figures. It is not possible to exactly determine how efficiently mining facilities use electricity as they are in different regions and have different configurations and settings.
The Cambridge Bitcoin Electricity Consumption Index (CBECI) provides a real-time estimate of the total electricity load and consumption of the bitcoin network. The model is based on a bottom-up approach initially developed by Marc Bevand in 2017 that takes different types of available mining hardware as the starting point.
Given that the exact electricity consumption cannot be determined, the CBECI provides a range of possibilities consisting of a lower bound (floor) and an upper bound (ceiling) estimate. Within the boundaries of this range, a best-guess estimate is calculated to provide a more realistic figure that we believe comes closest to bitcoin’s real annual electricity consumption.
The lower bound estimate corresponds to the absolute minimum total electricity expenditure based on the best-case assumption that all miners always use the most energy-efficient equipment available on the market. The upper bound estimate specifies the absolute maximum total electricity expenditure based on the worst-case assumption that all miners always use the least energy-efficient hardware available on the market if running the equipment is still profitable in electricity terms. The best-guess estimate assumes that miners use a basket of profitable hardware rather than a single model.