Supporting case
A fiscal brief compares five ways to tax AI—and explains where each breaks down
What happened
Elena Spatoulas Patel and Aparna Mathur's October 27 Tax Policy Center brief examines taxes on training compute, electricity, AI use, windfall profits and automation that displaces work. It compares mechanisms rather than estimating a new tax's revenue or reporting an enacted measure. Compute taxes face definitional and reporting problems; electricity is metered but cannot cleanly distinguish AI from other demand. Tokens vary in the work they represent. Windfall taxes need a defensible definition of ordinary profits, while displacement taxes must separate AI's effects from other technological and organisational changes.
The brief also considers broader tax-system alternatives. Its contribution is a usable account of the gap between the activity policymakers wish to tax and the base a government can observe, administer and collect. It supplies neither a demonstrated revenue yield nor a settled recommendation for a particular AI tax.
The comparison explains the administrative gap between AI activity and observable tax bases. That bears on governments’ ability to capture surplus, without establishing revenue yield or an adopted tax.
Dates and implementation stage
Development: 2026-10-27
