The electricity consumption and environmental footprint of blockchains

The topic of electricity consumption and environmental footprint of blockchains like Bitcoin and Ethereum has been recently vehemently debated. However, these discussions are not free from clichés, biases, misconceptions, and pitfalls. My goal here is to try to reduce this entropy, an apparent violation of the second law of thermodynamics.

How blockchains work

To understand the energy consumption of blockchain technology it is important to understand how a blockchain works.

A blockchain is a distributed ledger, using cryptography to secure an evolving consensus about tokens (digital assets) with economic value.

The atomic elements of a blockchain are the blocks. A block is a container of data. In its simplest form, it contains an identification number, a timestamp of creation, and a list of transactions moving tokens from a sender to a receiver. Each transaction has a fee (an amount in cryptocurrency) that must be payed by the sender, who digitally signs the transaction. The cost of the transaction is proportional to the computational complexity of the task that the transaction requires (for instance, minting an NFT costs more than transferring a coin). Each block is identified by a fingerprint called hash — created using a cryptographic hash function — that is used to certify the information content of the block. As soon as even a single bit of the block alters, its hash totally changes. Blocks are chronologically concatenated into a chain by adding to each block a field with the hash of the previous (parent) block in the chain. The genesis block is the first block of the chain and is the only one with no parent block.

The blockchain ledger is not stored on a single computer but instead it is distributed over a peer-to-peer network, so that each node of the network has an up-to-date copy of the entire ledger. To mine (create) a new block a node of the network needs to solve a computational problem that is hard to solve and easy to verify. This is a cryptographic puzzle that can be attacked only with a brute-force approach (i.e., by trying all possibilities), so that only the sheer computational power of miners counts. The node that solves first the puzzle becomes the miner of the block. This is called proof-of-work consensus method.

The miner broadcasts the newly minted block to all nodes. Nodes accept the block only if all transactions in it are authentic. In this case, they execute the transactions contained in the block and change the state of the ledger (who owns what). Nodes express their acceptance of the block by working on creating the next block in the chain, using the hash of the accepted block as the parent hash. Notice that uncompleted proof-of-work of nodes that did not mint the block is lost and all of them need to restart a new proof-of-work process with the next block. Finally, the miner, and only the miner, is rewarded with the fees of all blocked transactions plus a fixed, newly minted amount of cryptocurrency (this is how new coins are introduced in the cryptocurrency economy).

Why blockchains consume energy

The energy consumption of a proof-of-work blockchain is essentially due to the work of miners in solving the proof-of-work puzzle. The energy consumed to execute transactions when the block is minted is instead negligible compared to the proof-of-work use. In other terms, mining an empty block consumes the same amount of energy of mining a full block.

You can think of Ethereum kind of like a train engine throttled to the same speed all day, kept running by miners securing the network in exchange for ETH. In this analogy, transactions submitted to the network would be seats on the train. Due to the design of Ethereum, the train will keep running at the same speed and with the same energy consumption whether or not there are any seats filled [1].

Miners incur two types of financial costs. Capital expenditures, which are one-time fixed costs such as the purchase of specialised hardware to solve the the proof-of-work. Miners use application-specific integrated circuits (ASICs), which are purpose-built hardware optimised explicitly for proof-of-work algorithms. These ASICs have little to no use value outside of cryptocurrency mining or for a different mining process. Capital expenditures represent on average 45% of miners’ total costs, according to the 3rd Global Cryptoasset Benchmarking Study (2020) [2]. The remaining 55% of costs of miners are operational expenditures, ongoing variable costs dominated (75%) by the cost of electricity to run the application-specific hardware [2].

It is crucial at this point to understand the importance of the proof-of-work mechanism in the blockchain technology. Miners voluntarily incur these costs upfront in the expectation of a potential future reward (i.e. newly minted coins and block transaction fees). This is a clever way to guarantee the consistency and security of the blockchain, since the reward is distributed only if the miner honestly plays by the rules of the blockchain protocol.

A clever economic incentive design that promotes honesty over cheating underpins Bitcoin’s consensus process. Miners voluntarily incur financial costs ex ante in the expectation of a potential future reward. The threat of sunk costs (i.e. not receiving the block reward because of dishonest behaviour but having already paid for the performed work) — creates the financial incentive for miners to play by the rules. Assuming miners are profit-maximising economic agents, honesty is the most rational strategy. As a result, Bitcoin may be considered less a technical innovation and more a carefully calibrated socio-economic system that relies on a complex combination of economic incentives, game theory, and a solid technical foundation. [3]

Hence, essentially, proof-of-work is a method to reach a consensus over the state of the blockchain system (who owns what on the blockchain) in a trustless environment (peer nodes might be malicious or not) and the energy spent during for the proof-of-work method is functional to maintain the integrity of the blockchain system.

How much energy blockchains consume

This said, the energy consumption of proof-of-work blockchain is not negligible. Since all proof-of-work blockchains operate in a similar way, we focus on the main one, that is Bitcoin. We’ll investigate the energy consumption of Bitcoin with the help of the Cambridge Bitcoin Electricity Consumption Index (CBECI) [3]. The CBECI is an ongoing project created and maintained by the Digital Assets Programme (DAP) Team at the Cambridge Centre for Alternative Finance, an independent research institute based at The University of Cambridge, Judge Business School.

It is worth stressing that Bitcoin’s electricity consumption can only be estimated for a plurality of reasons. Moreover, these approximations are based on theoretical models that rely on specific assumptions [3].

The main driver of Bitcoin’s electricity consumption is expected mining profitability (i.e. forecasted revenues minus costs). This determines whether machines are running or sitting idle. [3]

Mining revenues are highly volatile and mainly depend on the bitcoin price (which is, essentially, unpredictable). Operational costs are more predictable and are primarily determined by electricity rates. Rising bitcoin price or decreasing electricity costs generally lead to increased electricity consumption as profitability is higher and hence more hardware (including less efficient hardware) will be employed.

According to CBECI [3], today (27th January 2022), Bitcoin blockchain uses 14.99 gigawatts (GW) of electrical power, which corresponds to a total yearly electricity consumption of 131.38 terawatt-hours (TWh). This figure is an annualised measure that assumes continuous power usage at the aforementioned rate over the period of one year. This corresponds to 0.59% of world’s total yearly electricity consumption. To put these figures into perspective, CBECI offers some intriguing comparisons, some of them are outlined below:

1. Bitcoin’s closest and most referenced real-world analogue is gold. Interestingly enough, gold mining consumes 131 TWh, essentially the same as its digital counterpart;
2. global air conditioning consumes 2199 TWh (16x more than Bitcoin), global data transmission networks consume 250 TWh, and global data centers consume 200 TWh;
3. TVs and lighting in US each use 60 TWh, while fridges in US use 104 TWh;
4. comparing to electricity consumption of entire countries, Bitcoin consumes a bit more than Ukraine (124.5 TWh) and a bit less the Egypt (149.1 TWh). US uses 3843.83 TWh, more than 29 times the use of Bitcoin;
5. as a fun fact, Bitcoin’s electricity consumption could power all tea kettles used to boil water in the UK for 29 years.

Electricity consumption of the mining of physical gold and digital gold (Bitcoin). Source: CBECI

Environmental implications

However, electricity consumption and environmental footprint are not necessarily correlated. As explained on CBECI’s website:

It is essential to distinguish between electricity consumption and environmental footprint. The first concerns the total amount of electricity used by the Bitcoin mining process. The latter concerns the environmental implications of Bitcoin mining. What ultimately matters for the environment is not the level of electricity consumption per se, but the carbon intensity of the energy sources used to generate that electricity.

For instance, one kilowatt-hour (kWh) of electricity generated by a coal-fired power station has a substantially worse environmental footprint than one kWh of electricity produced by a wind farm. As a result, rising (or falling) power demand does not automatically lead to a proportional increase (or decrease) in carbon dioxide and other greenhouse gas emissions. [3]

So, what kind of energy mix do miners use? According to the 3rd Global Cryptoasset Benchmarking Study (2020) [2], the most reported power sources used by miners are hydroelectricity (62% share of miners), coal (38%), and natural gas (36%). Other renewables besides hydro (wind, solar, and geothermal) are also part of their energy mix.

A share of 76% of the miners claim to use renewable energies as part of their mix, and 39% of mining’s total energy consumption comes from renewables.

In the worst case scenario, assuming that all the energy used by miners comes exclusively from coal (the most-polluting fossil fuel), the Bitcoin network would be responsible for roughly 0.35% of the world’s total yearly carbon dioxide emissions [3].

Furthermore, miners are economic actors that want to maximise their profits. Unlike almost all of the energy used worldwide, that must be produced relatively close to its end users, mining can happen anywhere.

Miners are energy nomads, attracted by renewable and waste energy that cannot be distributed or used in a cost-effective manner.

Hydro is the most well-known example of this, but another promising example is flared natural gas [4].

Alternatives to proof-of-work

Is there an alternative consensus method that is as secure as proof-of-work and consumes less energy? A popular candidate is proof-of-stake.

In proof-of-stake, electricity is replaced as the proving resource with staked capital in the form of locked-up cryptocurrency. Hence, instead of proving we have done some work (spending in hardware and electricity, therefore investing an amount of money), we simply prove we have staked that amount of money in the blockchain protocol, without computing anything.

More precisely, the proof-of-stake model is based on the idea that the more stake a user has invested into the system, the more likely they will want the system to succeed, and the less likely they will want to subvert it. Stake is an amount of cryptocurrency that once staked is no longer available to be spent. The likelihood of a user validating (mining) a new block (and hence receiving a reward) is tied to the ratio of their stake to the overall staked cryptocurrency. With this consensus model, there is no need to perform resource intensive computations.

Many blockchains are being created with this alternative consensus mechanism, for instance the Tezos blockchain. The use of proof-of-stake method reduces dramatically the energy consumption/environment footprint of blockchain, as confirmed in this careful report prepared by PwC in partnership with Nomadic Labs. Ethereum, the second most important blockchain after Bitcoin by market capitalization of its coin Ether, is indeed currently switching to proof-of-stake. The long process of upgrading started in December 2020 with the Beacon Chain, a proof-of-stake chain that is working in parallel with the main network Ethereum that is still operating under proof-of-work. The merge of the two chains, hence the upgrade of the main net to proof-of-stake, is estimated to happen in late 2022. The third and last phase, scheduled in 2023, is the introduction of shard chains, that will improve Ethereum’s capacity to process transactions and store data, and will possibly decrease costly transaction fees currently paid by users.

Furthermore, because of the prohibitive fees paid on Ethereum blockchain, many layer 2 scaling solutions are more and more popular on top of Ethereum. These solutions are designed to help scale applications by handling transactions off the main Ethereum chain (layer 1). This in turn might decrease gas fees on Ethereum, and hence also the profitability and energy consumption of miners.

However, it is unclear to date whether alternative consensus algorithms like proof-of-stake can replicate the same security assurances as proof-of-work without engaging in substantial trade-offs.

At the moment the proof-of-work method has been empirically proved the most secure: even if one owns 100% of hash power they cannot rewrite blockchain history without a proof of work, that is, without spending a conspicuous amount of energy and resources to rewrite the ledger.

It is a fact that the major proof-of-work blockchains (Bitcoin and Ethereum) have never been successfully attacked so far (Bitcoin started in January 2009, Ethereum in July 2015).

Moreover, proof-of-stake raises ethical issues as wealth directly determines voting power. Users who can only stake smaller amounts of native tokens are unlikely to be chosen as validator nodes, leaving them with no other option than delegating their tokens to a pool. This pools in turn collect more voting power and get more rewards, which can be staked to further increase their position in the system. In other terms, the proof-of-stake mechanism without any limitation might ignite a cumulative advantage (or rich-get-richer) process.

However, according to Vitalik Buterin (notably, the inventor and co-founder of Ethereum), proof-of-stake is superior to proof-of-work for the following three reasons [6]: (i) it offers more security for the same cost; (ii) attacks are much easier to recover; (iii) it is more decentralized. Buterin admits, nevertheless, that proof-of-stake has two limitations compared to proof-of-work: it is more like a closed system, leading to higher wealth concentration over the long term, and it requires a form of weak subjectivity, since new nodes must trust someone else in the system when they enter.

Good practices

As outlined above, the main driver of blockchain electricity consumption is expected mining profitability, that is forecasted revenues minus costs. Revenues are determined by the amount of fees paid by senders of transactions as well as by the cryptocurrency rewards of mining which is directly influenced by the market price of the mined coin, while costs include the expenses for the mining hardware and that of electricity to run it. We have no control on most of these variables but, partially, the fees we agree to pay when we sign a transaction. As also suggested in [5], as individuals we can adopt some good practices:

1. perform transactions using a low gas price. This can be done by tolerating a longer confirmation time or executing transactions during times of the day with lower gas demand;
2. reduce gas usage. For example, lazy minting, adopted for instance on OpenSea for NFT minting, minimizes gas requirements by treating the creation and sale of an NFT as a single transaction. Moreover, blockchain programmers should strive to write efficient code when implementing smart contracts such that the execution of the code has lower computational complexity and hence smaller gas fees are required;
3. when possible, use layer 2 solutions on Ethereum or alternative blockchains that already adopt proof-of-stake;
4. stake money to Ethereum to become a validator or take part to a staking pool. This improves the security of the forthcoming to proof-of-stake version of Ethereum;
5. purchase carbon offsets to fund activities that have a negative greenhouse gas balance, such as planting trees or increasing the commercial viability of renewable energy.

Key takeaways

We can sum up the take-home message as follows:

1. the energy consumption of a proof-of-work blockchain (such as Bitcoin or, at the moment, Ethereum) is not negligible;
2. blockchain energy footprint is linked to block production (mining) via proof-of-work consensus method and not to transaction processing;
3. proof-of-work is fundamental to maintain the security of the blockchain system; it is a fact that major proof-of-work blockchains have not been successfully hacked so far;
4. the main driver of blockchain energy consumption is expected mining profitability. This is mainly determined by the market price of the mined coin, the amount of fees paid by transaction senders, and the price of electricity used for the mining process;
5. it is essential to distinguish between electricity consumption and environmental footprint. What ultimately matters for the environment is not the level of electricity consumption per se, but the carbon intensity of the energy sources used to generate that electricity;
6. miners are energy nomads, attracted by renewable and waste energy that cannot be distributed or used in a cost-effective manner;
7. alternative consensus methods, like proof-of-stake, can dramatically reduce the use of computation and hence energy for block validation; however, it is unclear to date whether they can replicate the same security assurances as proof-of-work.

Conclusions

But how much energy should a blockchain consume? How you answer that likely depends on how you feel about blockchain and how much value you think it creates for society. If you believe that blockchains offer no utility beyond serving as a technology to create financial Ponzi schemes, to launder money or commit other crypto crimes, and to support the diffusion of awful digital art and pesky collectibles and profile pictures, then it would only be logical to conclude that consuming any amount of energy is wasteful. If, instead, you believe that blockchain will build a new decentralized Web, where users fully own and control their own data, identity, and money, you most likely think that the consumed energy is extremely well spent.

References

[1] SuperRare Labs Team. No, CryptoArtists Aren’t Harming the Planet, 2021.

[2] Apolline Blandin, Dr. Gina Pieters, Yue Wu, Thomas Eisermann, Anton Dek, Sean Taylor and Damaris Njoki. 3rd Global Cryptoasset Benchmarking Study, 2020.

[3] Michel Rauchs, Anton Dek, and Apolline Blandin. Cambridge Bitcoin Electricity Consumption Index (CBECI).

[4] Nic Carter. How Much Energy Does Bitcoin Actually Consume? Harward Business Review, 2021.

[5] Samuele Marro and Luca Donno. Green NFTs: A Study on the Environmental Impact of Cryptoart Technologies,
arXiv:2202.00003, 2022.

[6] Vitalik Buterin. Why Proof of Stake, 2020.