Bitcoin (BTC) Energy Use Estimates Diverge by Measurement Method

Bitcoin (BTC) electricity estimates differ by tens of terawatt-hours because one relies on a mining-company survey and another annualizes an economic model, making the figures unsuitable as a direct comparison.

A survey-based estimate put annual electricity use at 138 terawatt-hours as of June 2024. A live economic estimate displayed 204.44 TWh on an annualized basis on Oct. 15, 2026. The first reflects survey data collected through June 2024; the second is a model estimate of an annual run rate. Neither number is a direct meter reading of electricity consumed across the network.

Bitcoin miners use specialized computers to compete in proof-of-work. They repeatedly hash candidate block headers until one meets the network’s target, helping secure the transaction history. A miner that produces a valid block can receive the block reward and transaction fees. The electricity estimate therefore describes the computing activity used to compete for blocks, not power assigned to a particular payment.

The 138 TWh estimate drew on a survey of 49 mining firms. Those firms represented 48% of the network’s hashrate, or 268 exahashes per second, during data collection. The estimate equaled about 0.54% of global electricity use. It also indicated that mining electricity consumption rose 17% from a year earlier, even as estimated hardware efficiency improved 24% to 28.2 joules per terahash.

That combination shows why an efficiency improvement does not by itself establish that total electricity use fell. The survey recorded a year-over-year efficiency gain alongside higher estimated consumption. Those figures describe the same survey period, but they do not identify a single cause for the increase.

The survey also estimated the sources of mining electricity. Sustainable sources accounted for 52.4%, including 42.6% from renewables and 9.8% from nuclear power. Natural gas supplied 38.2%, coal 8.9% and oil 0.5%. On the same survey basis, annual greenhouse-gas emissions were estimated at 39.8 million metric tons of carbon dioxide equivalent. An alternative model based on IP addresses put emissions at 69.6 million metric tons, illustrating that emissions estimates also depend on methodology.

The 204.44 TWh figure comes from a different kind of estimate. It is annualized, so it does not represent a tally of electricity used during the 12 months before Oct. 15, 2026. The economic model infers electricity use from miner income and costs, including electricity costs. Its output can change as those inputs change.

Mining revenue, electricity prices and hardware efficiency can affect power demand and model results. That means a live estimate is a snapshot of a modeled annual rate, while a survey records information gathered from participating firms for a particular period. The 138 TWh and 204.44 TWh values do not establish a measured high-low range for one period, nor do they alone show how much network consumption changed between 2024 and 2026.

Per-transaction energy figures require similar care. Dividing total network electricity use by the number of transactions creates an average allocation across those transactions. It does not measure the electricity needed to process an individual payment. For readers comparing Bitcoin energy estimates, the measurement period, method and energy mix all shape what a headline number can show.