Data centre taxes, AI token levies and capital tax reforms emerge as policymakers confront the fiscal consequences of automation
Artificial intelligence is rapidly becoming one of the biggest tax policy challenges in decades.
The question is no longer simply how to tax AI companies. It is how governments continue funding public services and redistributing wealth when AI increasingly replaces taxable human labour with machines.
A growing number of proposals in the United States illustrates the dilemma. Rather than converging on a single solution, policymakers are experimenting with taxes on data centres, electricity consumption, AI processing (“tokens”) and capital investment. Others argue the answer lies not in new AI taxes at all, but in fundamentally rebalancing tax systems towards capital and consumption. The winner? Probably VAT on AI as it works already globally!
AI threatens the traditional tax model
Modern tax systems were largely built around employment. Income taxes, payroll taxes and social security contributions remain the largest source of government revenue in most developed economies.
If AI shifts a growing share of economic value from wages towards corporate profits, intellectual property and capital ownership, governments face two linked challenges:
- replacing lost labour tax revenues; and
- ensuring AI-driven wealth is shared more broadly across society.
This is why the debate has rapidly evolved beyond simply taxing technology companies. It has become a debate about wealth redistribution in an increasingly automated economy.
Data centres become the first target
Rather than taxing AI itself, several US lawmakers have targeted the infrastructure behind it.
Senator Ron Wyden has proposed removing generous federal tax incentives for data centres while introducing a new gross receipts tax on operators. The revenues could help fund support for workers displaced by AI.
Representative Andrea Salinas has separately proposed a federal excise tax on electricity consumed by large data centres, arguing the AI boom should contribute towards housing, infrastructure and environmental projects.
Taxing AI tokens
Representative Greg Casar AI Tax and Work Protection Act would impose a tax on AI “tokens” – the units of data processed by AI models. The proceeds would fund retraining, healthcare, education and other employment programmes aimed directly at workers affected by automation.
The proposal echoes Senator Bernie Sanders’ earlier suggestion that AI-generated wealth should ultimately be redistributed through a national sovereign wealth fund.
Economists disagree on the solution
While politicians increasingly favour targeted AI taxes, economists remain divided.
Brookings argues AI taxes may discourage innovation without solving the underlying fiscal challenge. Instead, it advocates broader taxation of capital income.
The Institute on Taxation and Economic Policy similarly argues governments should ensure AI-generated wealth is taxed through stronger capital gains and corporate tax rules.
The Tax Foundation reaches almost the opposite conclusion. It warns that removing investment incentives for AI infrastructure could simply push data centres overseas, arguing tax systems should remain neutral and instead tax profits once investments generate returns.
VAT may prove the winner
The debate highlights a more fundamental reality. Governments still do not know exactly what should be taxed: the models, the infrastructure, the electricity, the profits or the companies using AI.
As AI spreads through every industry, highly targeted AI taxes may prove difficult to define and enforce.
That leaves VAT and other broad-based consumption taxes looking increasingly resilient. Consumption continues regardless of whether income comes from salaries or AI-generated capital returns. In an economy where machines perform more work but people continue to consume, VAT may become one of the most reliable ways of preserving government revenues while broader debates over AI wealth redistribution continue.
How VATCalc has been using AI for years for tax determination
See how else VATCalc is applying Artificial Intelligence since 2022 to enhance VAT determination and reporting:
- AI and item classification in VAT determination;
- Tax advice generated by AI from our VAT expert-maintained tax engine
- Tax code mapping with AI to reduce significantly tax engine implementations and testing
