Ethereum Classic (ETC) sustainability report

Ethereum Classic is present on the following networks: Ethereum Classic.

Ethereum Classic reaches consensus through Proof of Work and continues to do so. It is the original chain that carried on unaltered after the 2016 contentious fork, and it did not follow the network it split from into Proof of Stake; there is no validator set, no staking deposit and no beacon chain here, and blocks are still produced by miners expending computation.

The hash function is Etchash, adopted in a 2020 upgrade as a modification of the Ethash algorithm the sibling chain used before its own transition. Like Ethash it is memory-hard: miners must hold a large pseudorandom dataset, the DAG, in memory and read from it unpredictably, which limits how far raw arithmetic throughput alone can be pushed. The modification doubled the interval over which that dataset is regenerated and resized, which slowed its growth so that graphics cards with smaller memories could keep participating and, at the same time, made the dataset incompatible with the sibling chain's, so hash power could no longer be redirected between the two without reconfiguration. Dedicated Etchash machines have since entered the market and now contribute a large share of the work alongside graphics hardware.

Blocks arrive roughly every thirteen seconds. Miners hash a header until the digest falls beneath the current difficulty target, and the difficulty is adjusted continuously to hold that interval. Because blocks are frequent relative to how fast they propagate, two valid blocks are often found at similar heights; the protocol lets a later block reference such orphaned siblings, so the work they represent is acknowledged rather than discarded and the branch with the greatest accumulated work prevails. Confirmation is probabilistic and grows firmer with depth.

The execution environment is the Ethereum Virtual Machine, and the chain tracks upstream virtual machine changes through periodic compatibility upgrades so that contract code and developer tooling remain portable. What it declines to adopt are changes that would reverse application-level outcomes or conflict with its fixed issuance rules, a distinction that defines its approach to protocol change.

Ethereum Classic is present on the following networks: Ethereum Classic.

Miners receive a block subsidy in newly issued units of the network's native asset plus the fees paid by the transactions they include. Issuance follows a fixed rule adopted in 2017: the subsidy falls by one fifth every five million blocks, roughly every two years, converging on a hard ceiling of about 210.7 million units in total. A block that references a valid block found at a similar height but left off the main branch also pays a reduced reward for that referenced block, so work that narrowly lost a race is not entirely wasted. There is no staking, no delegation and no bonded collateral, and consequently no slashing, no jailing and no ejection mechanism; an invalid block is simply rejected by every node and the electricity behind it earns nothing.

Users pay for computation rather than for bytes. Every operation the virtual machine performs has a fixed gas cost, storage writes cost considerably more than arithmetic, and a transaction specifies a gas limit bounding what it may consume. The fee is the gas actually used multiplied by the gas price the sender offers.

That price is set by a first-price auction: senders name a price, miners order transactions by what they are willing to pay, and the whole amount goes to the miner. The network has not adopted the base-fee-and-tip model used elsewhere, so no part of a fee is burned and there is no protocol-computed base fee that adjusts with congestion. A proposed upgrade would introduce a transaction type of that shape, but with the base fee routed to a protocol treasury rather than destroyed, in order to preserve the fixed issuance rule; it has not been activated on the main network, and legacy transactions would continue to work alongside it.

Contract storage is paid for once at the moment of writing, with a partial refund when a slot is cleared. There is no recurring rent and no periodic charge for keeping state alive, which means the cost of persistent storage is borne entirely at write time and the burden of retaining it falls on node operators thereafter.

Ethereum Classic is present on the following networks: Ethereum Classic.

The consumption figure for this network is a modeled estimate of what its mining hardware draws, not a metered total. The estimation approach used here is top down and driven by mining economics, on the premise that mining equipment accounts for effectively all of the network's electricity and that the machines actually powered on are those whose earnings exceed their running costs. This is deliberately not the node-counting method appropriate to stake-based networks; although the chain runs the same virtual machine as the network it separated from, that network's consumption is governed by a validator population while this one's is governed by a hash rate, and the two call for different models.

Candidate hardware is restricted to devices that can compute Etchash. Because the algorithm requires a large dataset held in memory, devices whose memory falls short of the current dataset size are excluded automatically, and that threshold rises over time, progressively retiring older graphics cards from the eligible set. The remainder, comprising graphics hardware and purpose-built Etchash machines, is characterized by published hash rates and power draws. A break-even point is derived by comparing the revenue a unit of hashing earns, taken from subsidies and fees observed on the chain, against a representative electricity price and other operating expenditure, and equipment below that point is assumed idle. Surviving models are allocated shares of the observed network hash rate, and total consumption is the sum of each model's share multiplied by its power characteristics.

No merged-mining adjustment applies. This chain's work is not shared with another chain, so the whole of the modeled consumption is attributed to it, unlike networks whose hashing is submitted to two chains at once.

The caveats are substantial and belong with the figure. The fleet's composition is inferred from what is commercially available and plausibly profitable rather than observed; a single assumed electricity price and overhead factor represents wide real variation; the mix between graphics hardware and dedicated machines is uncertain and materially affects the efficiency assumed; and the profitability test reduces multi-year investment decisions to a snapshot. Where evidence is thin, the assumption chosen raises rather than lowers the estimate, so the published value is more likely to overstate than understate, and it is revised as hardware and chain data improve.

Ethereum Classic is present on the following networks: Ethereum Classic.

The renewable share reported for this network comes from identifying where its electricity is consumed and applying the generation mix of those locations. Consumption on a Proof of Work chain sits overwhelmingly in mining equipment rather than in ordinary relay nodes, so the geographic question is where that equipment is installed and which grid it is connected to. Locations are inferred from publicly observable network data: the addresses and metadata advertised by reachable peers, whatever the pools coordinating mining work publish, and other public sources, each resolved to a country or region. Observation is necessarily partial, since mining machines do not disclose their whereabouts and pools are not required to report where their customers operate. Where the footprint cannot be established with confidence, the observed distribution of a structurally comparable network stands in for it, chosen because a similar consensus family and reward structure tend to draw the same sort of operator toward the same sort of electricity. Comparability here is judged on the mining economics, not on the shared virtual machine, so the appropriate stand-in is another network secured by hashing rather than the stake-based chain this one shares an execution environment with.

The country weights are matched to national statistics on the renewable proportion of each grid's generation, and averaging across the footprint by weight gives the share attributed to the network. The statistics come from Share of electricity generated by renewables, compiled from Ember's yearly electricity data together with the Energy Institute's Statistical Review of World Energy, with major processing by Our World in Data.

Energy intensity is a distinct measure from the annual total, defined as the marginal energy cost of one additional transaction. On this chain that relationship is weak, because hashing continues at the same rate regardless of how full blocks are, and because transactions differ greatly in the computation they demand of the virtual machine. The figure should be read as an allocation of a largely fixed consumption across observed activity rather than as the energy a particular transfer or contract call causes to be spent.

Both inputs carry uncertainty. National averages hide regional and seasonal variation in generation, and mining capacity concentrates where electricity is cheap rather than spreading evenly within a country, so the renewable share is an approximation.

Ethereum Classic is present on the following networks: Ethereum Classic.

Emissions attributed to this network are derived from its estimated electricity consumption combined with the carbon intensity of the grids that supply it, resting on the same geographic work as the renewable share. Mining equipment accounts for essentially all of the electricity, so the question is which grids that equipment draws from and how carbon-intensive each one is.

Locations are inferred from publicly observable network data, including peer addresses and metadata and any public disclosure by the pools that coordinate mining work, and are resolved to countries or regions. Where the footprint cannot be established directly, the distribution of a structurally comparable network is used instead, selected on the similarity of its mining economics rather than on any resemblance in execution environment.

Two scopes are reported separately. Scope 1 covers emissions from sources the operators of the infrastructure control directly, such as fuel burned on site at a mining facility. For participants who purchase electricity from a grid and run computing equipment on it, this is normally nil, and it is stated as nil rather than absorbed into the other scope. Scope 2 covers the indirect emissions embodied in that purchased electricity and accounts for the whole of the network's footprint. It is obtained by weighting each region's share of estimated consumption by that region's emissions per unit of electricity generated and summing across the footprint. The carbon intensities used are taken from Carbon intensity of electricity generation, compiled from Ember's yearly electricity data and the Energy Institute's Statistical Review of World Energy, with major processing by Our World in Data and made available under a CC BY 4.0 license.

Greenhouse gas intensity is expressed per transaction, as the marginal emission associated with settling one more. It carries forward every uncertainty in the consumption estimate and adds the grid data's own: intensities are annual national averages that conceal hourly and seasonal variation, and mining capacity is not spread evenly within a country. Transactions also vary widely in the computation they require, which makes a per-transaction figure a coarse unit of comparison. Where a parameter is unresolved the conservative choice is taken, which tends to overstate rather than understate the result, and the figures are restated as observation improves.