A new House bill would impose an excise tax on major AI companies, with rates that automatically rise if unemployment climbs. Introduced by Reps. Sara Jacobs, Greg Casar and Valerie Foushee, the proposal offers a bifurcated structure: taxing either the value of AI 'tokens' — the data units AI models use — or revenue from AI services and certain related-party transactions, whichever produces the higher sum.
The rates start at 2% for tokens and 3% for revenue when unemployment sits at 5% or below, and increase as joblessness rises. The proceeds would fund job creation in housing construction, infrastructure, and child and elder care.
'If Congress does nothing, the rise of AI could create the biggest wealth transfer in history from the bottom to the top,' Jacobs said in a joint press release. 'If AI profits off human work, workers deserve job security and a share of those profits.'
The measure is the latest in a string of Capitol Hill efforts to tax the AI sector. Sen. Ron Wyden has proposed a new excise tax on AI data centers, with revenue directed to displaced workers. Sen. Elizabeth Warren wants to tax AI companies partly based on the energy their data centers consume. Most aggressively, Sen. Bernie Sanders's American AI Sovereign Wealth Fund Act would levy a one-time 50% tax on OpenAI, Anthropic and xAI, converting the proceeds into public shares of those companies.
Not every proposal reaches for the tax code. The bipartisan AI Workforce PREPARE Act, from Sens. Jim Banks, Maggie Hassan, John Hickenlooper and Jon Husted, would improve federal tracking of AI-linked layoffs and study retraining programs — without new levies. Separately, Reps. Josh Gottheimer and Mike Lawler have proposed a tax credit covering 30% of qualified AI-training expenses, up to $2,500 per employee annually, for firms that retrain workers.
Even the industry itself has floated concessions. Bill Gates has called for a tax on AI tokens and robots to offset a tax system he says 'nudges you toward replacing people with machines.' DuckDuckGo founder Gabriel Weinberg said his company would accept a 10% AI token tax to match the payroll taxes employers already pay. Anthropic CEO Dario Amodei suggested the federal government could levy a 3% tax on AI revenue for redistribution, calling it a 'reasonable solution' even though it wasn't in his company's economic interest. OpenAI's Sam Altman met with Sanders in June to discuss giving the public a stake in his company.
What these bills share is a premise: that the government, not the market, should decide how AI's gains are distributed before the technology has even finished reshaping the labor market. An escalating excise tax pegged to the unemployment rate is not a neutral mechanism — it is a penalty on capital investment in computing power, priced to rise automatically the moment hiring softens, regardless of why.
The incentive-based alternatives on the table, like the Gottheimer-Lawler training credit, at least reward firms for retraining workers rather than punishing them for building the technology in the first place. Capital rewards clear rules; a tax bill that scales with the jobless rate does the opposite, injecting uncertainty into exactly the sector Washington claims it wants to keep leading globally.
