Viability of Bitcoin Mining as a Carbon-Reducing Flexible Load: Calibration and Empirical Test Across U.S. Electricity Markets

Abstract

Bitcoin mining has been proposed as a flexible computing load capable of reducing grid carbon emissions by absorbing renewable surplus and curtailing during scarcity, yet the conditions under which this mechanism holds have not been empirically tested. We evaluate the viability condition C̅_M = (R − c) × (phi) across five U.S. ISO markets using 2024 data. We calibrate mining revenue R from live Bitcoin prices and network hashrate against current ASIC specifications, derive marginal fossil cost c from EIA-923 fuel receipts and heat rates (natural gas only, generation-weighted, by ISO), and measure grid parameters phi and renewable-surplus share q from 2024 hourly ERCOT and CAISO price data. Fossil-capacity share by ISO is drawn from EIA-860. At the July 2026 Bitcoin price of approximately $64,000 — 31% below the paper's assumed $95,000 — the viability condition fails in all five ISOs at standard three-year hardware lifetimes. Viability emerges only at four-to-five year horizons and fails entirely in CAISO and PJM, where marginal gas costs eliminate the mining margin regardless of amortization. The measured grid parameters further diverge from model assumptions: ϕ exceeds 0.99 in both ERCOT and CAISO versus the assumed 0.935, and renewable-surplus hours q range from 2.3% to 4.2% — one-quarter to one-half the assumed rate. These findings indicate that the FCL carbon-reduction mechanism is theoretically coherent but its empirical preconditions are not uniformly met in current U.S. electricity markets.