U.S. Energy Agency Estimated Bitcoin Mining at Up to 2.3% of Power Demand

The U.S. Energy Information Administration estimated on February 1, 2024 that domestic Bitcoin mining consumed between 25 and 91 terawatt-hours of electricity during 2023. That range represented approximately 0.6% to 2.3% of total U.S. electricity demand, which the agency placed at 3,900 TWh for the year.

The estimate mattered because it gave federal policymakers and grid planners an official—though deliberately preliminary—measure of an industry whose facilities could add substantial, mobile loads to regional power systems. The EIA simultaneously explained why it had obtained emergency approval for a mandatory survey of identified commercial cryptocurrency miners.

The agency did not establish that miners definitively consumed 2.3% of U.S. electricity. Its unusually broad range reflected unresolved questions about where mining occurred, how intensively facilities operated and whether older geographic data still described the network in 2023.

How the agency constructed the range

For its top-down calculation, the EIA used the Cambridge Bitcoin Electricity Consumption Index’s estimate that global Bitcoin mining consumed between 67 and 240 TWh during 2023. It then assumed the United States retained approximately 38% of global mining activity, based on Cambridge’s 37.8% estimate for January 2022—the latest published national distribution available to the agency.

Applying that share produced the domestic range of 25 to 91 TWh. The limitation is material: the calculation assumed that a geographic distribution measured nearly two years earlier remained approximately valid throughout 2023. It also covered Bitcoin, not every proof-of-work cryptocurrency, despite the EIA article’s broader discussion of cryptocurrency mining.

The EIA developed a separate bottom-up check by identifying 137 U.S. mining facilities. It found maximum electricity-use information for 101 of them, totaling 10,275 megawatts. Assuming 80% utilization would translate that capacity into roughly 70 TWh annually, close to the upper portion of the top-down range.

That 80% figure was an approximation based on information from only a few facilities. Maximum capacity was not the same as actual consumption, and the agency said it lacked enough operational data for a well-sourced utilization estimate. The facility inventory was also not presented as a complete census.

A mandatory survey was meant to narrow the uncertainty

The Office of Management and Budget approved the emergency information request on January 26, 2024. The EIA said on January 31 that identified commercial miners would be required to report energy-use details beginning the following week, while its February 1 analysis described planned monthly collection from February through July 2024.

OMB characterized the collection as experimental and provisional. Its approval required the EIA to test survey content, solicit public comment and avoid using the results for population-wide inference. The collected responses would represent only the identified companies in the market-study sample, not every U.S. miner.

The intended questions extended beyond a national consumption total. The agency wanted to measure changes in mining demand, locate areas of concentrated growth and identify the electricity sources serving mining facilities. Those details could help planners distinguish a flexible industrial load from demand that might aggravate peak-period constraints.

The EIA also cited 41 gigawatts of proposed cryptocurrency-mining capacity in the Electric Reliability Council of Texas interconnection queue, including 9 GW for which planning studies had been approved. Queue requests were potential projects, not operating demand, and could not be treated as evidence that 41 GW would be built or consumed.

What the February 1 record established

The defensible event-day conclusion was limited: federal analysts believed U.S. Bitcoin mining had become a potentially significant electricity consumer, but existing information could support only a wide range. The estimate did not measure carbon emissions, electricity prices, miner profitability or Bitcoin’s market value, and it did not prove that mining caused any specific grid disruption.

Later context must remain separate. The EIA subsequently discontinued the emergency collection after litigation, an outcome not knowable from the February 1 record. That later reversal did not retroactively improve or invalidate the published estimate; it underscored why the estimate’s assumptions and data limitations must remain explicit.