What records verify Bitcoin mining energy claims?

Define the claim before collecting records

Before reviewing a claim, identify its subject and unit. A statement about a facility's electricity use needs an energy quantity over a defined interval; a statement about power describes a rate or load. Carbon dioxide-equivalent emissions require an emissions factor and a stated boundary. A carbon credit or avoided-emissions claim is a further claim with its own baseline, ownership and registry evidence. These quantities answer different questions and should not be collapsed into a single figure labelled 'energy impact'.

A useful review separates the observed record from the calculation made from it. A meter export can show electricity measured at a particular point, while an invoice shows what a utility billed for a period. A model may estimate a network-wide total from equipment and economic assumptions. In each case, record the source, unit, date range, geography and measurement or estimation method before comparing numbers. Without those fields, two figures that look similar may describe different boundaries or periods.

Start with the facility meter and boundary

Start with the site boundary: which buildings, mining containers, cooling equipment and auxiliary systems are included? Match that boundary to a single-line diagram, meter schedule, facility plan or other operating record. A meter attached to miners alone may exclude cooling, ventilation, transformers or common-area loads. A utility meter may include unrelated tenants. The person preparing the claim should show how included and excluded loads were assigned, rather than applying a general multiplier without a traceable basis.

For measured consumption, retain the raw interval data and the meter identifier, units, timestamps, time zone and coverage period. Reconcile totals with invoices and note missing or estimated readings. If a facility used behind-the-meter generation, storage or a separate supply, document those flows and avoid counting the same electricity twice. An invoice can support billed consumption but may not show when power was used; interval readings can show timing but still need a documented meter boundary.

A site-specific emissions estimate needs a factor that matches the location, period and accounting question. Preserve the factor publisher, version or vintage, geography, units and whether it describes average generation, delivered electricity or another basis. Convert units explicitly before multiplying consumption by the factor. If a report compares facilities in different regions, align the periods and factor methodology first. A national annual average and a real-time local estimate may answer different questions and should not be presented as interchangeable measurements.

Choose a grid factor that matches the site

The location-based and market-based approaches concern different evidence. A location-based calculation applies a grid factor for a stated region and period. A market-based calculation uses qualifying contractual energy instruments under the relevant accounting rules. The GHG Protocol's existing Scope 2 Guidance describes these approaches; its corporate-standard revision process is still under development as of 11 November 2026. State which method is used and check the current standard before a new reporting cycle.

Real-time grid data can help explain how electricity-system intensity changes by location and time, but its coverage and production-versus-consumption basis matter. The International Energy Agency notes that some real-time datasets may omit small or distributed generation and that its current emissions-factor database supplies annual national averages for location-based reporting. A time-matched estimate may be informative for operations, but it should not be called an audited corporate inventory unless it follows the applicable reporting method and assurance scope.

Keep network estimates separate from site readings

Network-wide Bitcoin energy indices are useful for questions about the network as a whole, not a particular mine. Cambridge's CBECI methodology estimates demand using a basket of real-world hardware and profitability assumptions because direct network-wide measurement is unavailable. Its electricity and greenhouse-gas indices use different inputs and have different limitations. A facility operator should not assign a global network estimate to its own site or claim that a share of global hash rate is a direct meter reading.

Read modeled estimates alongside their assumptions and update history. Cambridge presents lower-bound, upper-bound and best-guess electricity scenarios; its emissions methodology also relies on geographic mining-distribution data and electricity-mix assumptions. The predicted portion can use the most recent geographic map as a proxy until new data arrives, and earlier estimates can be revised. When citing an index, state its publisher, version or access date, metric, scope and estimate status; do not relabel a model output as measured energy use.

Treat electricity contracts as a distinct claim

A power-purchase agreement, supplier tariff or energy-attribute certificate supports a contractual claim only to the extent its terms, geography, delivery period and accounting criteria allow. Keep the contract, certificate serial or registry record, cancellation or retirement event and relevant load account together. A contract may allocate an attribute without proving that the same generator physically supplied each unit consumed at the mine. Avoid translating a renewable purchase into an unsupported statement that every mining hour used renewable electricity.

The GHG Protocol is revising its corporate standards following a public consultation in 2025-2026; its October 2026 update describes the process as ongoing and sets out proposed next steps. A reviewer should cite the current published guidance used for the reporting period and describe later proposals as proposals. This keeps an operational inventory distinct from potential claims about avoided emissions, system impacts or future accounting rules, which may use different methods and evidence.

Understand what a mining pool can document

Mining-pool records answer a narrower operational question. Depending on the service and reward method, an account report may show submitted shares, credited rewards, fees, worker identifiers and payouts. Those records can help reconcile participation and income, but they do not show the miner's electricity source, facility load, cooling demand or emissions factor. A pool's network status or reward statement is not a carbon certificate, power contract or independent audit of a mining facility.

For a financing or carbon-market review, build a file that follows each claim to its source: facility identity and boundary, meter data, utility invoices, equipment inventory, grid factor, procurement contract, registry record and calculation workbook. Note who supplied each item, the period it covers and whether it is measured, billed, modeled or contractual. Record missing intervals and shared loads as unresolved evidence rather than filling gaps with a convenient estimate. This makes the conclusion easier to update when the data changes.

Keep operational emissions separate from lifecycle and offset claims. A site electricity calculation does not by itself cover equipment manufacture, transport, refrigerants, construction or end-of-life treatment. A carbon-credit purchase also does not change the meter reading. If a report claims a reduction or compensation, identify the project baseline, ownership, issuance and retirement records, accounting boundary and any double-counting controls. Do not subtract an offset from a facility's gross electricity emissions unless the reporting framework and claim explicitly support that treatment.

Build an evidence file that can be reviewed

A concise evidence statement can say what was measured, where, over which period, using which meter and factor, and what the record excludes. A separate sentence can describe contractual electricity attributes or offsets. If only a network model is available, label the result as a network estimate and link to its methodology. If the source does not disclose a site boundary or factor vintage, say that the available record does not establish those fields. Precision about scope is more useful than a single headline number without a reproducible trail.

Reviewers can then test whether the claim follows from the documents: reproduce the unit conversion, compare the meter interval with billed totals, check the emissions factor vintage and confirm that contractual attributes cover the stated load and period. Preserve the original files and the calculation version so another reviewer can repeat the check. Where records conflict, report the difference and its likely cause instead of selecting the smaller figure. The purpose is to understand the evidence, not to infer a carbon outcome from mining revenue or hash rate.