Supporting case
A fiscal-policy brief explains why taxing AI spending may miss the surplus—and the displacement
What happened
Aaron Till and Shai Akabas’s Bipartisan Policy Center brief, published August 24 and distinguishes taxes on AI companies, AI adoption, computing inputs and broader capital income. It is a conceptual analysis, not a new estimate of revenue or employment effects and not an enacted policy.
Its central design problem is concrete: if AI becomes cheaper, use and possible labor displacement can grow while spending on AI falls. A tax on purchase price could therefore collect less just as the concern motivating it grows. Revenue taxes can reach firms that have no profits; profit taxes cannot. Taxes on energy or computation use a different base again, and the ultimate burden can shift to customers. Broad capital-tax reform avoids some disputes about what counts as AI, while creating its own incidence and design choices.
This adds a useful comparison framework to the existing token-tax and surplus-capture debate. It helps distinguish a taxable activity from the economic rent or fiscal loss policymakers actually want to address; it does not establish that one instrument is best.
Dates and implementation stage
Development: 2026-08-24
