Klement: If AI Productivity Materializes Immediately, US Debt-to-GDP Could Fall from 185% to 113%

In an October 13, 2026 post, investment blogger Joachim Klement shifts the discussion from whether AI will boost productivity to a more concrete fiscal transmission question: if AI does raise productivity, how would it alter the US debt trajectory? He draws on research by Stephan Bales at Germany's development bank KfW to translate abstract technological optimism into traceable fiscal arithmetic.

The KfW mechanism chain is relatively direct. AI productivity gains drive GDP growth; higher nominal GDP expands the government's tax base while economic improvement reduces social welfare spending needs. Together these forces compress the primary deficit (before interest expense), slowing debt accumulation and ultimately lowering the debt-to-GDP ratio. KfW estimates suggest that if AI adds 0.2 percentage points of annual productivity growth, the US projected primary deficit in 2056 could shrink from 2.8% to 1.5%; at 0.4 percentage points, the primary deficit might be eliminated entirely within thirty years.

A narrower deficit brings not only slower debt stock growth but also reduced interest burden. Lower primary deficits mean fewer Treasuries issued, and supply-demand shifts may compress yields, creating a dual improvement in debt cost and debt scale. Under this path, US debt-to-GDP could fall from the currently projected 185% in 2056 to 143% (0.2% scenario) or even 113% (0.4% scenario). Klement derives a practical implication for fixed-income investors: buying 30-year US Treasuries today is implicitly a large bet that AI will significantly increase productivity; otherwise, rising debt burdens may push long-term yields much higher.