Hashprice modulates the electricity demand response of Bitcoin miners
For power system planners and market designers, this paper shows that Bitcoin miners' demand response is state-dependent on crypto revenue, challenging assumptions of stable flexibility.
This paper examines how Bitcoin mining load in Texas responds to electricity costs and finds that the response is modulated by hashprice (expected mining revenue). When hashprice is high, load is less responsive to electricity price increases, implying that treating miners as stable demand-response resources may overstate grid flexibility.
Large, fast-controllable loads such as Bitcoin mining facilities are increasingly viewed as potential sources of flexibility in modern power systems, yet the conditions under which this flexibility is realized remain incompletely understood. Using the Texas power market as an empirical setting, we examine how Bitcoin-mining load responds to two distinct electricity-sector cost channels: contemporaneous wholesale electricity prices and incentives created by coincident-peak-based transmission charges. We find that mining load responds to both cost channels in a manner consistent with miners operating around a breakeven point. At the aggregate level, we observe that mining load decreases as electricity-sector costs rise, but the strength of this response depends on hashprice, a measure of expected mining revenue from the crypto-financial sector. When hashprice is higher, aggregate load responsiveness is weaker. This mechanism is especially evident in the wholesale-price response. Mining load remains largely online at low prices and begins to decline only when electricity costs become large relative to expected mining revenue, with higher hashprice shifting the implied curtailment threshold toward higher wholesale prices. These findings indicate that Bitcoin-mining demand response to electricity-sector costs is economically state-dependent and shaped by revenue conditions in the crypto-financial sector. Treating such loads as stable demand-response resources may therefore overstate available grid flexibility, with implications for power-system planning, market design, and reliability assessment.