AI-driven productivity gains could markedly reduce the U.S. budget deficit, but even in optimistic scenarios they would not resolve fiscal sustainability. That is the conclusion of a Brookings Institution working paper.
These factors claw back around half of the potential gains. See the paper here: https://t.co/zVcRA4Yr4F
— Ben Harris (@econ_harris) July 1, 2026
According to a projection by the Congressional Budget Office, under current law the U.S. federal debt could reach 175% of GDP by 2056.
What the model shows
Authors Ben Harris, Neil R. Mehrotra, and William Overcash examined four 10-year scenarios:
- a traditional productivity shock;
- a labor-displacement shock;
- a scenario with lower mortality and changes in health-care spending;
- a combined disruptive scenario.
In the baseline “traditional” scenario, the picture looks almost salvific: the primary deficit turns into a surplus, the annual deficit shrinks by more than $2 trillion by 2036, and the deficit-to-GDP ratio falls by nearly 5 percentage points. The logic is straightforward: faster growth broadens the tax base and lifts revenues.
But subsequent channels erode the gains. The authors estimate that AI-specific effects could claw back more than half of the potential improvement.
Why the effect diminishes
The authors identified five channels that pressure the budget alongside higher productivity:
- Lower mortality and longer life expectancy increase the number of older citizens and spending on age-based social programs.
- Labor displacement raises demand for transfers and income-support programs.
- A potential AI arms race could accelerate defense spending.
- A shift of the tax base from labor to capital lowers the average tax rate because capital income is taxed differently from labor income.
- A higher neutral rate makes borrowing more expensive and increases interest costs.
As a result, the researchers do not view artificial intelligence as a standalone solution to the U.S. debt problem: output gains may be strong, but the net fiscal effect is much weaker.
Fortune noted that Elon Musk previously called broad adoption of AI and robotics almost the only way to solve the U.S. debt crisis. The Brookings paper says faster productivity alone is insufficient if accompanying obligations rise as well.
More cautious assessments of AI’s impact on the economy also persist. According to CEPR, AI-related labor productivity gains were about 0.6% in 2025 and could reach 1.8% in 2026. In finance and high-skilled services, the figure is projected above 2%.
Why it matters for markets
If AI speeds up growth but also lifts the neutral rate, the cost of money in the U.S. could remain higher for longer than investors expect. That affects Treasury yields and debt-service costs: the more expensive borrowing is, the greater the pressure on the budget and the higher markets’ sensitivity to Treasury auctions, inflation, and signals from the Fed.
For bitcoin and other risk assets, this implies dependence not only on AI optimism and capital inflows into technology, but also on the price of liquidity. As yields rise, investors more often cut positions in volatile assets, and dollar fixed-income instruments become more competitive.
A separate channel relates to stablecoins. Their issuers remain large holders of short-term U.S. Treasuries, so the path of public debt, rates, and demand for Treasuries matters not only for traditional markets but also for crypto infrastructure.
In June, Bank for International Settlements analysts said that the investment boom around artificial intelligence, which supported the global economy in 2025, is itself becoming a source of macrofinancial risks.
