Renewable-only dispatch contracts achieve the same zero-carbon computing as a carbon tax exceeding every operating U.S. carbon price: evidence from the Cipher–Luminant power purchase agreement

Authors

  • Shreyas Kondapuram Department of Finance, George Mason University, Fairfax, VA
  • Jiasun Li

DOI:

https://doi.org/10.13021/jssr2026.5667

Abstract

Bitcoin mining is the archetypal flexible computing load: an interruptible electricity user that can absorb renewable surplus or shed demand within seconds. Whether hosting such loads lowers grid carbon depends on the contract governing their dispatch, yet the two leading instruments — a Pigouvian tax on fossil generation and a renewable-only dispatch cap — had not been compared inside a single equilibrium model calibrated to a real U.S. mining contract. That gap is closed here by extending the Li (2026) flexible-computing-load framework to nest both instruments and calibrating it to the Cipher Mining–Luminant power purchase agreement (SEC Exhibit 10.35, 2022), a 200–210 MW fixed-price contract inside ERCOT. Analytical solution of the deregulated spot-market variant yields a closed-form fully greening tax at which fossil-served computing exits; the renewable-only cap attains identical zero-carbon computing at the same entry threshold, and the two instruments differ only by a renewable pass-through wedge that is strictly positive under the tax and identically zero under the contract. Numerical solution of the joint capacity–entry–dispatch equilibrium across eight residual-demand distributions confirms all seven comparative-static signs, sign-invariance of the carbon gap over a Wasserstein ambiguity set of radius 1.2, and monotonically heavier worst-case tails (Pareto index 2.50 to 2.10). At contract-calibrated values (R ≈ $69/MWh, c ≈ $45/MWh, e ≈ 0.43 tCO2/MWh) the fully greening tax is ≈ $56/tCO2, exceeding RGGI and California allowance prices by 40–180%. Renewable-only dispatch contracts are therefore the feasible near-term decarbonization instrument for flexible computing loads.

Published

2026-09-24

Issue

Section

Costello College of Business: Department of Finance