Generative AI spurs a wave of layoffs and corporate profits, but erodes public finances that rely on payroll taxes. The United States and other countries face the dilemma of how to impose new taxes on capital.
(Background: Bloomberg: Coinbase to launch prediction markets and tokenized US stocks next week, moving towards the “everything exchange”)
(Additional context: a16z predicts four major trends to be announced by 2026)

Table of Contents

• The Invisible Erosion of the Tax Base
• The Definition Dilemma of Robot Taxes
• Structural Imbalance Between Capital and Labor
• Political Battles and Future Choices

Wall Street’s stock prices have been climbing steadily this year, but by the end of 2025, the tens of thousands of layoffs announced by giants like Amazon, Meta, and UPS highlight another grim reality: the productivity boom brought by AI is coming at the expense of the national tax base.

Let’s imagine a scenario: white-collar and blue-collar workers who are laid off stop paying income and social security taxes, while algorithms, which replace them, have no tax obligations. The fiscal gap begins to widen—will this become a headache for finance ministers worldwide? The Spanish newspaper El País even directly raises a point: If AI replaces workers, should it also pay taxes?

The Invisible Erosion of the Tax Base

We know that modern welfare states mainly rely on labor taxation to sustain operations. In the US, personal income tax and social security taxes account for about 85% of federal revenue.

However, with the proliferation of generative AI, companies are replacing human labor with software. If future salary expenses continue to decline, every lost paycheck means a further erosion of government tax revenue. Reports indicate that Amazon’s profits jumped 38% this year, while it also cut 14,000 jobs worldwide.

The Definition Dilemma of Robot Taxes

Faced with revenue shortfalls, the most straightforward political proposal is to impose a “robot tax” on automation equipment. U.S. Senator Bernie Sanders advocates this approach to slow down layoffs and fund retraining programs. But academia warns that practically defining what constitutes a “robot” is difficult.

Daniel Waldenström, an economist at the Stockholm School of Economics, states:

We will never be able to precisely define it. What is automation? Is it a chip, an application, or a line of code?

If tax policies focus only on visibly mechanical arms, they might overlook the truly profitable cloud algorithms. The International Federation of Robotics also warns that taxing could penalize productivity-enhancing tools, reducing national competitiveness. No major economy has yet successfully implemented a direct robot tax.

Structural Imbalance Between Capital and Labor

Another fundamental reason is tax system distortion. Over the past decade, corporate tax rates in OECD countries have fallen sharply from 33% in 2000 to around 25% today; meanwhile, the tax burden on workers (including income taxes and social insurance contributions) has only decreased by 1.3 percentage points, from 36.2% to 34.9%.

The long-term “reward capital, burden labor” approach results in companies preferring to invest in depreciable, tax-deductible hardware and software rather than hiring people who need to pay social security.

Based on this phenomenon, IMF researchers suggest that instead of taxing machines, it might be better to increase taxes on AI-driven excess profits or capital gains to rebalance the burden while still maintaining incentives for innovation.

Political Battles and Future Choices

Ultimately, the issue boils down to politics. Nearly a year into Trump’s presidency, he must balance the expectations of Silicon Valley capital for deregulation with the anxieties of voters in the Rust Belt (Trump’s stronghold) over job losses.

MIT economists Acemoğlu and Johnson warn that, although automation has increased overall output over the past forty years, it has not brought “shared prosperity” but instead widened income inequality. If the tax system cannot redistribute AI’s gains, social divisions may fully surface in the 2026 elections.

As algorithms concentrate wealth at an unimaginable speed, the welfare structures supported by payrolls are becoming fragile. The 2026 decision will test each government’s imagination and ability to implement a new social contract for the next generation.

!Dongqu Official Website TG Banner-1116 | Dongqu Trends - The Most Influential Blockchain News Media