A tax would raise the cost of AI experimentation, forcing firms to prioritize safe, efficiency-focused projects over speculative R&D. This 'known ROI bias' would hamper the discovery of transformative AI applications and entrench incumbents who can better absorb experimentation costs.

Instead of controversial wealth or broad income taxes, a more politically viable solution for AI-driven job displacement is to levy a higher corporate tax rate specifically on companies whose profit margins surge after replacing workers with AI.

The core argument for a token tax is not to penalize AI, but to ensure the tax system doesn't artificially favor automation. It shifts the tax base from human labor (payroll, income taxes) to AI's productive capacity, measured in tokens, to prevent tax-incentivized job displacement.

Taxing AI usage via a "token tax" is a flawed policy. It disproportionately harms the most ambitious and productive firms—those using AI to augment their human workforce and boost competitiveness. This creates a perverse incentive to avoid the very AI adoption that strengthens the economy.

Following Amazon's model, AI-native companies will reinvest all available cash into acquiring more compute power for a competitive edge. They will operate in a perpetual land-grab mode and never need to show a profit, making them impossible to tax effectively and rendering corporate taxation an obsolete funding mechanism for the state.

The push for an AI token tax isn't limited to politicians. Tech leaders, including Mark Cuban, DuckDuckGo's CEO, and Anthropic's CEO Dario Amadei, have publicly supported or floated the idea, signaling a surprising openness within the industry to novel policy solutions for AI's societal impact.

A flat per-token tax is fundamentally flawed because token consumption doesn't correlate with economic value creation. The same number of tokens can be used for low-value tasks like generating spam or high-value tasks like legal analysis, making it an inequitable and inefficient tax mechanism.

Mark Cuban suggests a federal tax on AI tokens to curb usage and raise funds. Critics argue this is a form of central planning that penalizes a specific business model, making foreign and open-source alternatives more attractive and hurting US competitiveness.

Policies capturing wealth from a few AI labs are too narrow. Long-term economic benefits will likely accrue to a wide range of companies that successfully integrate AI to boost productivity. This suggests a broad-based corporate tax would be a more effective tool for wealth redistribution than targeting a few supposed "winners."