AI Productivity Gains May Not Fix U.S. Deficit, Yale Study Finds

An artificial intelligence-fueled productivity surge could boost U.S. economic growth substantially over the next decade, but new research suggests those gains would do surprisingly little to address the nation's mounting fiscal crisis under existing tax policies.

The Budget Lab at Yale has released modeling showing that even an optimistic AI growth scenario would generate only modest increases in federal tax revenue—far short of what would be needed to meaningfully reduce deficits that are projected to reach $2.2 trillion by 2030.

Why it matters

With U.S. public debt climbing and political appetite for spending cuts or tax increases virtually nonexistent, policymakers have hoped that economic growth might provide a path out of fiscal trouble. This analysis suggests that AI-driven expansion alone won't be enough—at least not without significant changes to how the tax code treats different forms of income.

The capital-labor tax gap

The core issue lies in how the federal government taxes different types of income. Labor income faces a top marginal rate of 37%, while capital income receives substantially more favorable treatment. The corporate tax rate stands at 21%, long-term capital gains are taxed at a maximum of 23.8%, and significant capital ownership flows through tax-advantaged vehicles like retirement accounts and charitable endowments.

If AI enables companies to generate higher profits while reducing their human workforce, that shift from labor to capital would mean less taxable income at higher rates—even as overall economic output grows.

Projected revenue impact

The Yale team modeled two scenarios. In a slow-growth case where AI provides only modest economic benefits, federal revenue in 2030 would remain essentially flat. In a rapid-growth scenario with 3.3% annual GDP expansion and a declining labor share of national income, federal revenue would increase by $216 billion in 2030.

That sounds substantial until compared against the Congressional Budget Office's baseline projection of a $2.2 trillion deficit that year—making the AI revenue boost roughly one-tenth of the shortfall.

Uncertainty and policy implications

"On the one hand, faster productivity growth would generate more tax revenue, all else equal," wrote Yale researchers John Iselin and Ryan Nunn. "On the other hand, our current tax system may not be structured to efficiently collect revenue from the economic activity produced by AI."

Iselin noted that "without substantial changes to how the U.S. taxes capital income, the federal government will be leaving a lot of revenue on the table."

The analysis comes with significant caveats. Predicting AI's economic impact involves massive uncertainty around how much labor's share of income will decline, how inequality among workers will shift, and whether displaced workers will require expanded social safety net programs beyond current law.

Tax policy itself could change. If AI displaces large numbers of workers while the fiscal situation deteriorates, Congress may decide to increase taxes on capital income to rebalance the burden.

The research was first reported by Axios and underscores that the relationship between AI-driven growth and federal revenues is far more complex than simple extrapolation might suggest.