Treasury Secretary Scott Bessent and Federal Reserve Chair Kevin Warsh have argued that artificial intelligence will soon unleash a productivity boom so substantial it could ease concerns about America's $40 trillion national debt. But economists are increasingly skeptical that ordinary workers will see the benefits. Labor's share of U.S. income has plummeted to 52.8% of GDP, the lowest level recorded since 1947, while corporate profit margins have expanded to a record 14.9% of output. This divergence predates the current AI wave: it reflects a decade of automation, post-pandemic cost discipline (including reduced hiring), and sustained capital investment.
The concern is structural. Gregory Daco, chief economist at EY-Parthenon, notes that productivity growth has historically protected corporate margins rather than boosted worker compensation. In the second quarter, economic output expanded 1.7% while labor hours rose only 0.3%—yet compensation grew 2.6%, effectively flat or contracting in real (inflation-adjusted) terms. Daco warns that labor's current 52.8% share may not represent a floor; as long as concentrated capital gains flow to a narrow set of large firms, the slide could continue. Historical precedent—the railroad and dotcom booms—shows that technological revolutions typically see large, vertically integrated firms capture initial gains, with smaller competitors facing persistent cost pressure and wage growth lagging for years.
A critical wrinkle complicates the AI narrative: the capital intensity is historically unprecedented. Data center investment could reach $31 trillion by 2050, nearly equal to current U.S. GDP. Yet imports of specialized AI servers (classified as "large computers" in Census data) have exploded to $450 billion at an annualized pace last month, up from roughly $50 billion annually through 2023. Since imported equipment adds to investment statistics but simultaneously subtracts as an import, the net GDP contribution is zero—explaining why strong capital spending and productivity gains have coincided with weak hiring and declining worker income share.